WLTP - GRPE 72 02 Rev.1

Submitted by the Chair of the IWG on WLTP Informal document GRPE-72-02-Rev.
1
72nd GRPE, 11-15 January 2016,
agenda item 3(b)
Technical Report on the development of a

Worldwide Harmonised Light duty vehicle Test
Procedure (WLTP)
Informal document no. GRPE-72-02-Rev.1
UN/ECE/WP.29/GRPE/WLTP-IG
Final version
January 2016
Author:
Iddo Riemersma - Sidekick Project Support
(sponsored by the European Commission)

Contents
1 Introduction .............................................................................................................................. 5
2 Objective of WLTP.................................................................................................................... 6
3 Organisation, structure of the project and contributions of the different subgroups to

the UN GTR ............................................................................................................................... 7
3.1 WLTP Informal Group (WLTP-IG) ....................................................................................... 7
3.2 DHC group ......................................................................................................................... 11
3.3 DTP group and subgroups in phase 1a ............................................................................. 11
3.3.1 Laboratory procedures for electrified vehicles (LabProcEV) ..................................... 13
3.3.2 Particulate mass/Particulate number (PM/PN) .......................................................... 14
3.3.3 Additional pollutants (AP) .......................................................................................... 15
3.3.4 Reference fuel (RF) ................................................................................................... 15
3.4 WLTP phase 1b ................................................................................................................ 17
3.4.1 Drafting GTR ............................................................................................................. 18
3.4.2 EV subgroup .............................................................................................................. 19
3.4.3 AP Taskforce ............................................................................................................. 19
3.4.4 Round Robin testing .................................................................................................. 20
3.4.5 Taskforces on open issues ........................................................................................ 22
3.4.5.1 Reference Fuels ................................................................................................. 24
3.4.5.2 Definitions .......................................................................................................... 25
3.4.5.3 Normalization ..................................................................................................... 27
3.4.5.4 Number of tests .................................................................................................. 30
3.4.5.5 Review of coastdown tolerances ........................................................................ 32
3.4.5.6 Fuel consumption calculation ............................................................................. 32
3.4.5.7 Speed trace tolerance / drive trace index ........................................................... 33
3.4.5.8 Utility Factors ..................................................................................................... 36
3.4.5.9 Additional pollutants ........................................................................................... 37
3.4.5.10 Mode selection and predominant mode ............................................................. 39
3.4.5.11 Other taskforces ................................................................................................. 41
2
4 Test procedure development ................................................................................................ 42
4.1 General Purpose and Requirements ................................................................................ 42
4.2 Approach .......................................................................................................................... 43
4.3 Improvements in the GTR ................................................................................................ 43
4.4 New concepts of the GTR ................................................................................................ 46
4.4.1 Interpolation method .................................................................................................. 46
4.4.2 Vehicle selection ....................................................................................................... 48
4.4.3 Interpolation/extrapolation range ............................................................................... 49
4.4.4 Vehicle test mass ...................................................................................................... 49
4.4.5 Vehicle coastdown mode and dynamometer operation mode................................... 50
4.4.6 Tyres ......................................................................................................................... 51
4.4.7 On-board anemometry .............................................................................................. 51
4.4.8 Default road load factors ........................................................................................... 52
4.4.9 Road load matrix family ............................................................................................. 52
4.4.10 Torque meter method ................................................................................................ 54
4.4.11 Wind tunnel method .................................................................................................. 56
4.4.12 Alternative delta Cd.A determination.......................................................................... 62
4.4.13 Road load family ........................................................................................................ 63
4.4.14 Manufacturer’s responsibility on road load ................................................................ 66
4.4.15 Alternative vehicle warm-up procedure ..................................................................... 67
4.4.16 REESS charge balance (RCB) correction for ICE vehicles ....................................... 67
4.4.17 Electrified Vehicles .................................................................................................... 68
4.4.18 RCB correction for OVC-HEVs, NOVC-HEVs and NOVC-FCHVs ............................ 69
4.4.19 Shortened test procedure for PEV range test ........................................................... 73
4.4.20 Phase-specific values for EVs ................................................................................... 77
4.4.21 Interpolation method for electrified vehicles .............................................................. 80
4.4.22 End of PEV range criteria .......................................................................................... 82
4.4.23 FCV test procedure ................................................................................................... 84
4.4.24 WLTP post-processing .............................................................................................. 85
3
4.5 GTR structure ................................................................................................................... 87
4.5.1 Annex 3 – Reference fuels ........................................................................................ 87
4.5.2 Annex 4 - Road and dynamometer load .................................................................... 87
4.5.3 Annex 5 – Test equipment and calibrations .............................................................. 95
4.5.4 Annex 6 – Type 1 test procedure and test conditions ............................................... 96
4.5.5 Annex 7 – Calculations .............................................................................................. 98
4.5.6 Annex 8 - Pure electric, hybrid electric and fuel cell hybrid vehicles ......................... 99
4.5.7 Annex 9 – System equivalency ............................................................................... 104
5 Validation of the test procedure ......................................................................................... 105

5.1 Validation Tests ............................................................................................................... 105
5.1.1 Participants and vehicles, measured parameter ..................................................... 105
5.1.2 Evaluation issues .................................................................................................... 111
5.2 Results ............................................................................................................................. 112
5.2.1 Overnight soak temperatures .................................................................................. 112
5.2.2 Test cell temperatures ............................................................................................. 112
5.2.3 Test cell humidity ..................................................................................................... 114
5.2.4 Speed trace violations ............................................................................................. 116
5.2.5 Charge depleting tests for PEV and OVC HEV ....................................................... 118
Appendix 1 – Utility Factors ....................................................................................................... 129
Appendix 2 – Road Load Matrix Family .................................................................................... 140
Appendix 3 - Emission legislation ............................................................................................. 145
Appendix 4 - List of participants to WLTP ................................................................................ 147
4
1 Introduction
The development of the WLTP was carried out under a program launched by the World
Forum for the Harmonization of Vehicle Regulations (WP.29) of the United Nations Economic
Commission for Europe (UN ECE) through the working party on pollution and energy
(GRPE). The aim of this project was to develop, by 2014, a World-wide harmonised Light
duty driving Test Procedure (WLTP). A roadmap for the development of a UN Global
Technical Regulation (UN GTR) was first presented in August 2009.1
Most manufacturers produce vehicles for a global clientele or at least for several regions.
Albeit vehicles are not identical worldwide since vehicle types and models tend to cater to
local tastes and living conditions, the compliance with different emission standards in each
region creates high burdens from an administrative and vehicle design point of view. Vehicle
manufacturers therefore have a strong interest in harmonising vehicle emission test
procedures and performance requirements as much as possible on a global scale.
Regulators also have an interest in global harmonisation since it offers more efficient
development and adaptation to technical progress, potential collaboration at market
surveillance and facilitates the exchange of information between authorities.
Apart from the need for harmonisation, there was also a common understanding that the new
test procedure was expected to represent typical driving characteristics around the world.
Increasing evidence exists that the gap between the reported fuel consumption from type
approval tests and the fuel consumption during real-world driving conditions has grown over
the years. The main driver for this growing gap is the pressure put on manufacturers to
reduce CO2 emissions of the vehicles. As a result, this has led to exploiting the flexibilities
available in current test procedures, as well as the introduction of fuel reduction technologies
which show greater benefits during the test than on the road. Both issues are best managed
by a test procedure and cycle that match the conditions encountered during real-world
driving as close as possible.
Since the beginning of the WLTP process, the European Union had a strong political
objective set by its own legislation (Regulations (EC) 443/2009 and 510/2011) to develop a
new and more realistic test cycle by 2014. This very aspect has been a major political driving
factor for setting the time frame of the phase 1 of the WLTP development.
The development of the WLTP took place taking into account that two main elements form
the backbone of a procedure for vehicle emission legislation, namely:
a) the driving cycle used for the emissions test, and
b) the test procedure which sets the test conditions, requirements, tolerances, and other
parameters concerning the emission test
The development of the WLTP was structured accordingly, having two working groups in
parallel.
Within the roadmap of WLTP there are 3 phases distinguished, and the first phase is further
subdivided in a phase 1a and 1b (see paragraph 3.1).This document is the technical report
that describes the development of the test procedure, and explains the elements that are
new or improved with respect to existing emission testing procedures. This report was
published at the time that phase 1b was completed.
The technical report on the development process of the driving cycle is described in a
separate document2, which was published at the point where WLTP phase 1a had finished.
1
See document ECE/TRANS/WP.29/2009/131 at
http://www.unece.org/fileadmin/DAM/trans/doc/2009/wp29/ECE-TRANS-WP29-2009-131e.pdf
2
See document GRPE-68-03 at http://www.unece.org/trans/main/wp29/wp29wgs/wp29grpe/grpeinf68.html
5
2 Objective of WLTP
The primary objective of the global technical regulation (GTR) developed in the WLTP
process is to form the basis for the emission regulation of light-duty vehicles within regional
type approval and certification procedures, as well as an objective and comparable source of
information to consumers on the expected fuel/energy consumption and electric range, if
applicable. Each of the Contracting Parties to the 1998 Agreement could then transpose this
new standard into their own legislative framework.
As a result of this overarching objective, the work on WLTP aimed to develop a test
procedure that would fulfil the following basic demands:
a) The test procedure should be globally harmonised and applicable, and
b) The results should be representative for average real-world vehicle performance in terms
of emissions, fuel and/or energy consumption.
The work on the WLTP was chosen to be structured in such a way that the two main
elements that form the backbone of the procedure for vehicle emission legislation were
separately developed. These two elements are:
a) the test cycle, which should be representative for average real-world vehicle operation,
and
b) the test procedure, which should comprise a method to determine the levels of gaseous
and particulate emissions, fuel and/or electric energy consumption, CO2 emissions and
electric range –if applicable- in a repeatable and reproducible manner.
The underlying report highlights the work that took place during the course of the
development of the test procedure. The technical report on the development process of the
driving cycle is described in a separate document2.
6
3 Organisation, structure of the project and contributions
of the different subgroups to the UN GTR
3.1 WLTP Informal Group (WLTP-IG)

In its November 2007 session, WP.29 decided to set up an informal WLTP group under
GRPE to prepare a road map for the development of the WLTP3. After various meetings and
intense discussions, WLTP informal group presented a first road map in June 2009
consisting of 3 phases. This initial roadmap was subsequently revised a number of times,
and consists of the following main tasks:
a) Phase 1 (2009 - 2014): development of the worldwide harmonised light duty driving cycle
and associated test procedure for the common measurement of criteria compounds, CO2,
fuel and energy consumption;
b) Phase 2 (2014 - 2018): low temperature/high altitude test procedure, durability, in-service
conformity, technical requirements for on-board diagnostics (OBD), mobile air-
conditioning (MAC) system energy efficiency, off-cycle/real driving emissions;
c) Phase 3 (2018 - …): emission limit values and OBD threshold limits, definition of
reference fuels, comparison with regional requirements.
The first meeting of the WLTP IG group took place in Geneva, on 4 June 2008. After the 4th
meeting the WLTP-IG was disbanded and the steering group as shown in Figure 1 took the
lead over the development process.
Three technical working groups were established under WLTP, each with a specific
development task (see Figure 1):
a) the development of the worldwide harmonised test cycle (DHC) group, to develop the
worldwide-harmonised Light-duty vehicle Test Cycle (WLTC), including the validation test
phase 1, i.e. to analyse the test cycle and propose amendments where necessary;
b) the development of the test procedure (DTP) group, to develop the test procedure, and to
transpose this into a UN GTR;
c) the validation task force (VTF) group, to manage the validation test phase 2, i.e. to
analyse the test results and to propose amendments to the test procedure where
necessary.
Within the DTP subgroup, the following working groups were established that would deal with
specific technical areas of the test procedure:
 ICE-Laboratory Procedures (LabprocICE) for the development of the road load
determination methods and laboratory test procedures for conventional vehicles with an
internal combustion engine (ICE)
 E-Laboratory Processes (LabprocEV) for the development of all laboratory test
procedures related to electrified vehicles, including hybrids
 PM/PN for the development of a test procedure for the determination of Particulate Matter
and the Particle Number in the exhaust gas.
3
The UNECE World Forum for Harmonization of Vehicle Regulations (WP.29) is a worldwide regulatory
forum within the institutional framework of the UNECE Inland Transport Committee. For more
information refer to the UNECE website: http://www.unece.org/trans/main/wp29/introduction.html
7
 Add
ditional Pollutants (AP
P) for the test
t proced
dure of currently non--regulated gaseous
g
emisssion comp
ponents (NOO2, N2O, NH
H3, EtOH, aldehydes, ettc.)
 Refeerence Fue
el (RF) to define
d the specificatio
ons for refe
erence fuelss used in emission
e
testiing.
WLTP
P Informa
al Group
ICE-Lab. Process
P
Steeriing Group Develop the test proce
edure for Internal
Represenntative of the con
ntracting C
Combustion Engine (IC
CE) vehicle
party and
d/or Organizationns
E-Lab. Process
P
Develop the test proce
edure for Electrified
v
vehicle
DHC
C subgrou
up PM/P
PN
edure for PM/PN
DHC: The Development
D of the Harrmonized test Cycle m
measurements
AP
P
DTP subgrou
up edure for Additional
Pollutants (NO2, NH3,, N2O etc)
DTP: Develo
opment of Test Proced
dure
Referencce Fuel
S the fuel property fo
Set or Validation tests
VTF team
VTF: Validattion Task Force GTR text
t
Figure 1: Origina
al structure
e of WLTP--IG
Figure 2: Origina
al time schedule
s set
s by WLTP-IG
W or Cycle and Proc
fo cedure
develo
opment
8
Figure 2 shows the
e original ro
oad map for the develo
opment of WLTP.
W The
e developme
ent work
started in Septemb
ber 2009.
Since th he beginninng of the WLTP
W devellopment pro ocess, the European Union had a major
political objective set
s by its ow wn legislatio on to impleement a new w and more e realistic te
est cycle
by 2014 44. This waas a strong political driving facto or for setting the time frame of p phase 1.
Howeve er, during the work off the DTP group g it be
ecame clear that a nu umber of isssues, in
particulaar but not only
o in relation to electrric and hybrrid-electric vehicles,
v co
ould not be resolved
r
in time for an ado option of th he first verssion of the WLTP GT TR by WP.2 29 in Marcch 2014.
Therefo ore it was ag
greed that the work of Phase 1 wo ould be dividded into 2 sub-phases
s :
a) Phaase 1a (20099 - 2013): d developmen nt of the worrldwide harm monized light duty drivving
cycle and the basic
b test prrocedure. This led to th
he first versiion of this GTR,
G which was
published as offficial workinng documen nt ECE/TRA ANS/WP.29 9/GRPE/201 13/13 and a series
of amendments s published as informal documentt GRPE-67-04-Rev.1;
b) Phaase 1b (2013 3-2016): furrther develo
opment and
d refinementt of the test procedure,, while
inclu
uding additional items into the GTR.
Figure 3: Propos
sed structu
ure for WLT
TP phase 1b
4
Refer to
t Regulatio
ons (EC) 44
43/2009 and
d 510/2011
9
2013 2014 2015
Items
3. Qtl 4. Qtl
Q 1. Qtl 2. Qtl 3. Qtl 4. Qtl 1. Qtl 2. Qtl 3. Qttl 4. Qtl
Phase 1b Work
W
Report too 71st GRPE ◆ 167
7th WP.29 ◆
gtr drafting
g
(a) LabProc
cICE ;
(i) normaliza
ation methods,
speed tra
ace index
(ii) energy ecconomy rating and abbsolute speed
change ra
ating for speed trace violations
(iii) wind tunn
nel as alternative meth
hod for road load
determinaation
(iv) suppleme
ental test with represe
entative regional
temperatu
ure and soak period
(b) EV-HEV ;
(i) calculation method of each phase range for pure
ehicles (PEVs)
electric ve
(ii) shortened
d test procedure for PEV
P range test
(iii) combinedd CO2 (fuel consump ption) of each

phase forr off-vehicle charging hybrid electric
vehicles (OVC-HEVs)
(
(iv) hybrid Ele
ectric Vehicle (HEV)//PEV power and
maximum m speed
(v) combined test approach for OVC-HEVs
O and
PEVs
(vi) fuel cell vehicles
(vii) utility facctors
(viii) precond
ditioning
(ix) predomiinant mode
(c) APM ;
measuremennt method for ammon
nia, ethanol and
aldehydes
(d) DHC ;
(i) speed vio

olation criteria
(ii) further do
ownscaling in wide op
pen throttle (WOT)
operationn
(iii) sailing an
nd gear shifting
(e) all ;
others (furth
her improvement of gttr)
3. Qtl 4. Qtl
Q 1. Qtl 2. Qtl 3. Qtl 4. Qtl 1. Qtl 2. Qtl 3. Qttl 4. Qtl
2013 2014 2015
Figure 4: Changeover of the

t WLTP organizatio
on from ph
hase 1a to
o phase 1b
b, and
ng of phas
plannin se 1b
10
The work for phase 1b was structured and organised according to the following expert
groups under WLTP informal working group (see Figure 4 and 4):
 GTR drafting: coordination over all groups, to ensure that the GTR is robust, coherent,
and consistent. This is a continuation of the GTR drafting work under phase 1a;
 E-lab: specific test conditions and measurement procedures for electric and hybrid-
electric vehicles. This is a continuation of the EV-HEV group under phase 1a;
 Additional Pollutants (AP) for the test procedure of currently non-regulated emission
components (NO2, N2O, NH3, EtOH, aldehydes, etc.). This is a continuation of the AP
group under phase 1a;
 Taskforces: for each specific topic that had to be amended or be added in phase 1b, the
informal working group would designate a taskforce leader, who would work in a group
with interested stakeholders on developing a testing methodology and a GTR text
proposal. This could be any issue related to the former DHC, LabProcICE, PM/PN or RF
working groups;
 Round Robin testing, i.e. to analyse the test results and to propose amendments to the
test procedure where necessary.
 Drafting: a subgroup has been established under the lead of the drafting coordinator and
with members from WLTP leading team, Annex coordinators, Contracting Parties and
NGO experts. The main tasks were a „peer review“ of the GTR, check for
inconsistencies, editorial review of IWG and expert proposals.
3.2 DHC group

The structure and details of the DHC group are outside the scope of this report, and can be
found in the Technical Report of the DHC2.
3.3 DTP group and subgroups in phase 1a

The first meeting of the DTP subgroup took place at Ann Arbor (United States of America)
from 13 to 15 April 2010. The DTP group was first chaired by Michael Olechiw
(Environmental Protection Agency, United States of America). The chairmanship was later
taken over by Giovanni D’Urbano (Federal Office for the Environment, Switzerland). Initially
the secretary was Norbert Krause (International Organisation of Motor Vehicle Manufacturers
(OICA)), later followed-up by Jakob Seiler (German Association of the Automotive Industry
(VDA)).
DTP Chairs and secretaries

Chair Secretary
Michael Olechiw (Environmental Protection Norbert Krause (OICA)

Agency, United States of America)
Jakob Seiler, German Association of the
Giovanni D’Urbano, Federal Office for the Automotive Industry (VDA)
Environment (Switzerland)
11
As indicated in Figure 1, there were five working groups established within the DTP group to
promote an efficient development process by dealing with specific subjects of the test
procedure:
a) laboratory procedures for internal combustion engine vehicles (LabProcICE) to work
on the road-load determination and test procedures in the testing laboratory for
conventional vehicles;
b) laboratory procedures for electrified vehicles (LabProcEV) to work on all test
procedures that specifically address electrified vehicles;
c) particulate mass/particle number (PM/PN) to work on test procedures for the
determination of particulate mass and particulate numbers in the exhaust gas;
d) additional pollutants (AP) to work on test procedures for gaseous emission
compounds other than CO2, NOx, CO and HC;
e) reference fuel (RF) to work on specifications for reference fuels used in emission
testing.
The subgroup leaders were appointed at the second DTP meeting which was held in Geneva
in June 20105 (see WLTP-DTP-02-03). After this meeting, the subgroups started their work
and the following DTP meetings (14 in total until mid of 2013) were dedicated to discussions
about the reports from the subgroups. The structure of the work distribution and the
allocation of tasks are illustrated in Figure 5.
• Supervision of subgroups
• Assure working interfaces
DTP • Development of regulatory code
• Report to WLTP informal group
Conventional Reference
HEV/PHEV/EV
Engines Fuels
Minimum
PM/PN Additional Pollutants Lab Process Lab Process
Requirements
 PM  NO2  Road load  Electrical range

- Harmonize  Inertia classes  Utility factor
US/ECE/Japan  N2O  Shift mode  Emissions CO2
 Vehicle dyno mode  Correction factor
 PN -Methodology  Twin vs. single roller dyno  Reduced cycle length
-Completion of ECE and equipment – performance requirement  Coast down (unique to
regulation  2WD/4WD this technology.)
 Electrical accessories  Delineation of vehicles
-Methodology and  Tire pressure w/ electric powertrains
equipment  Preconditioning
 Vehicle soak
 Vehicle cooling during test
 Test cell operation
 Calculations  Sulfur level
 Emissions measurement  Metal additives
 Bench aging  Distillation curve
 Ki factors
 Vehicle option content
 Calibration gas
specifications
Figure 5: Structure of the DTP and its subgroups6

A more detailed overview for the scope of activities of these subgroups is presented in the
next paragraphs.
5
See document WLTP-DTP-02-03
http://www.unece.org/trans/main/wp29/wp29wgs/wp29grpe/wltp_dtp02.html
6
See document WLTP-DTP-01-14
12
The terms of reference were the same for all subgroups and are listed below:
1. The working language of the subgroup will be English.
2. All documents and/or proposals shall be submitted to the Chair (in a suitable electronic
format) in advance of scheduled meetings/web-conferences. Participants should aim to
submit documents at least 5 working days in advance of meetings/web-conferences.
3. An agenda and related documents will be circulated to all subgroup participants in
advance of all scheduled meetings/web-conferences.
4. Documents will also be uploaded by the Chair to the European Commission’s website
and a link provided from the UN-ECE website.
5. The progress of the subgroup will be reported to DTP group meetings by the Chair (or
other nominated person). Reporting will include a list of “Open Issues” on which
agreement has yet to be reached within the subgroup, which will be updated by the Co-
chair.
6. Following each meeting/web conference the Chair (or other nominated person) will
circulate a short status report, along with the list of “Open Issues” to chairs and co-chairs
of DHC, DTP and other DTP subgroups.
Another point which was common to all subgroups is the development approach. The
development of the measurement procedures was based on a review and comparison of
already existing regional regulations in the Contracting Parties of the 1998 Agreement.
The scope of activity was dedicated to the issues covered by the tasks of the different
subgroups and is further detailed in the following paragraphs.
3.3.1 Laboratory procedures for electrified vehicles (LabProcEV)

Chair Secretary
Per Öhlund – Swedish Transport Agency Yatuka Sawada, OICA

(Sweden)
Kazuki Kobayashi - NTSEL (Japan)
The first meeting of this subgroup took place at 21.09.2010. The LabProcEV subgroup was
tasked with developing a test procedure which includes vehicle preparation, vehicle
configuration, vehicle operation, measurement equipment and formulae for the measurement
of criteria pollutants, CO2, fuel consumption and electric energy consumption for electrified
vehicles.
The scope of activity was described as follows7:
a) identify content of Contracting Party legislation relevant to laboratory procedures for
Electrified vehicles excluding PM/PN and additional pollutants measurement procedures;
b) compare relevant content of Contracting Party legislation (US, UN ECE, Japanese);
c) decide upon which content to use for WLTP or, where appropriate, to specify alternative
requirements for WLTP;
7
See document WLTP-DTP-E-LabProc-001-ToR._V2, available at CIRCABC under WLTP-DTP section
13
d) identify additional performance metrics associated with electrified vehicles that may not
be covered by existing regulations. (i.e. battery charging times). Create harmonised test
procedures for the new performance metrics;
e) if necessary, conduct improvements on the basis of the following principles:
(i) narrow tolerances / flexibilities to improve reproducibility;
(ii) cost effectiveness;
(iii) physically reasonable results;
(iv) adapted to new cycle.
f) draft laboratory procedures for electrified light duty vehicles and specification text.
The LabProcEV subgroup was responsible for Annex 8 (pure and hybrid electric vehicles) of
the UN GTR. This is where measurement procedures and equipment dedicated to electrified
vehicles (and deviating from Annexes 5 and 6) are defined.
3.3.2 Particulate mass/Particulate number (PM/PN)

Chair Secretary
Chris Parkin, Department for Transport Caroline Hosier, OICA (after Chris Parkin
(United Kingdom) left WLTP she chaired this subgroup)
The PM/PN subgroup started its work by a web/phone conference at 07.07.2010. The scope
of activity included the following tasks8:
a) identify content of Contracting Party legislation relevant to PM and PN measurement
procedures;
b) compare relevant content of Contracting Party legislation (US, UN ECE, Japanese);
c) decide upon which content to use for WLTP or, where appropriate, to specify alternative
requirements for WLTP;
d) draft PM and PN measurement procedure and specification text.
The approach taken by the PM/PN group was to start from a detailed comparison of the
regulations from European Union, Japan and the United States of America. PM/PN
established a number of small expert teams to review and make recommendations back to
the wider team on measurement equipment specifications, particulate mass sampling,
weighing and all aspects of particle number measurement.
PM measurement is made by collecting the particulate on a filter membrane which is
weighted pre and post-test in highly controlled conditions. It was decided to update the
requirements as far as possible for technical progress and harmonisation, in such a way that
it would not require to replace the majority of existing particle mass measurement systems. A
major aspect of this decision is that particle number is also measured.
Regarding PN, only the UN Regulation No. 83 contains particle number measurement
requirements. Particle number measurement is an on-line measurement process to count
solid particles in the legislated size range in real time, where the total number of particles per
kilometre is reported for the test. The experts on particle number measurement reviewed the
procedure in detail to identify opportunities for tightening the tolerances to improve
8
See document WLTP-DTP-PMPN-01-02 Rev.2, available at CIRCABC under WLTP-DTP section
14
repeatability / reproducibility as well as improvements to the process and calibration material
specifications to adapt this method to recent technical progress.
The work of the PM/PN subgroup was incorporated in relevant parts of Annex 5, 6 and 7 of
the UN GTR.
3.3.3 Additional pollutants (AP)

Chair Secretary
Oliver Mörsch – OICA Covadonga Astorga, Joint Research

Centre (European Commission)
The first web/phone meeting of the AP subgroup took place at 20.07.2010.

The scope of activity for the AP subgroup (see WLTP-DTP-AP-01-01) included the following
tasks, building on procedures in existing legislation and expert knowledge within the group:
a) agree on additional pollutants to be addressed;
b) identify appropriate measurement methods for each of the pollutants;
c) describe measurement and calibration procedures and calculations based on existing
legislation and on output from lab procedure subgroup;
d) draft legislation text.
The following guidelines have been applied for the development of measurement methods
for the additional pollutants:
a) use or modify existing methods where reliable, cost effective and easy to apply
technologies are available;
b) reflect state of the art;
c) stipulate development of new measurement technologies;
d) replace cumbersome offline methods by online methods.
The work of the AP subgroup was incorporated in relevant parts of Annex 5, 6 and 7 of the
UN GTR.
3.3.4 Reference fuel (RF)

Chair Secretary
William (Bill) Coleman – OICA
No separate meetings were held for the RF subgroup. The scope of activity for the RF
subgroup was described as follows:
a) defining a set of validation fuels to support the development stages of the WLTP project
(stage 1), and;
b) defining a framework for reference fuels to be used by Contracting Parties when applying
the WLTP UN GTR (stage 2).
The scope of activity in phase 1a is restricted to stage 1. The subgroup had to undertake the
following tasks on the basis of a comparison of reference fuels in existing legislation and
expert knowledge within the group:
15
a) agree a limited number of fuel types and/or blends for which reference fuels are expected
to be required in the time frame of implementation of the WLTP project;
b) identify a list of fuel properties that will be significant to the validation of a future drive
cycle and/or test procedure for emissions and/or fuel consumption;
c) propose limits for the variation of these critical properties in order to specify a limited
number of candidate validation fuels to assess potential impact of the future drive cycle
on emissions and/or fuel consumption;
d) obtain approval from the WLTP project for the technical scope of the validation fuels
described in (c);
e) upon approval of the above mentioned parameter list, develop specifications for
candidate validation fuels to be used in the validation of the proposed drive cycles and
test procedures. These fuels should be limited in number, available at reasonable cost
and are not intended to restrict the decisions regarding reference fuels for the final
implementation of WLTP (Stage 2);
f) provide a forum of reference fuel experts who can at relatively short notice provide
coordinated advice and support on fuel related project issues to members of other sub-
groups of the WLTP Project.
These tasks required a fruitful cooperation with experts from the fuel production industry.
Since this cooperation could not be established, points (a) to (d) and (f) could not be fulfilled.
Already defined regional reference fuels were used for the validation tests of the proposed
drive cycles and test procedures.
As a consequence, Annex 3 of the UN GTR dedicated to reference fuels consists only of the
two paragraphs, requiring the recognition of regionally different reference fuels, proposing
examples of reference fuels for the calculation of hydrocarbon emissions and fuel
consumption, and recommending that Contracting Parties select their reference fuels from
the Annex. The text recommends to bring regionally agreed amendments or alternatives into
the UN GTR by amendments, without limiting the right of Contracting Parties to define
individual reference fuels to reflect local market fuel specifications.
In addition to that, tables with specifications for the following fuel types are included in the
UN GTR:
a) liquid fuels for positive ignition engines:
(i) gasoline/petrol (nominal 90 RON, E0);
(ii) gasoline/petrol (nominal 91 RON, E0);
(iii) gasoline/petrol (nominal 100 RON, E0);
(iv) gasoline/petrol (nominal 94 RON, E0);
(v) gasoline/petrol (nominal 95 RON, E5);
(vi) gasoline/petrol (nominal 95 RON, E10)
(vii) ethanol (nominal 95 RON, E85);
b) gaseous fuels for positive ignition engines:
(i) LPG (A and B);
(ii) natural gas (NG)/biomethane:
a. "G20" "High Gas" (nominal 100 % methane);
b. "K-Gas" (nominal 88 % methane);
c. "G25" "Low Gas" (nominal 86 % methane);
d. "J-Gas" (nominal 85 % methane)
16
c) liquid fuels for compression ignition engines:
(i) J-Diesel (nominal 53 Cetane, B0);
(ii) E-Diesel (nominal 52 Cetane, B5);
(iii) K-Diesel (nominal 52 Cetane, B5);
(iv) E-Diesel (nominal 52 Cetane, B7).
3.4 WLTP phase 1b

At the time that phase 1a was concluded, the main development on the test cycle and the
test procedure had finished. This resulted in the first version of the GTR, which was
published as official working document ECE/TRANS/WP.29/GRPE/2013/13 and a series of
amendments published as informal document GRPE-67-04-Rev.1. Even though the main
body of the GTR was now in place, still quite a number of open issues were yet to be
resolved. Especially on the area of electrified vehicles a considerable effort was needed to
finish the work on closing the open issues.
 With the changeover from phase 1a to phase 1b of WLTP, the structure of organization
was modified in such a way that the remaining open issues would be addressed by
dedicated taskforces. This new structure is shown in Figure 3 and Figure 4 of
paragraph 3.1.
 Drafting: a subgroup has been established under the lead of the drafting coordinator
and with members from WLTP leading team, Annex coordinators, Contracting Parties
and NGO experts. The main tasks were a „peer review“ of the GTR, check for
inconsistencies, editorial review of IWG and expert proposals.
The former subdivision into subgroups DHC and DTP was abandoned, and only a few
working groups. All activities in WLTP would from now on be managed by the WLTP-IG
leading team.
WLTP-IG leading team

Chair Technical Secretary
Stephan Redmann – BMVI (Germany) Noriyuki Ichikawa – OICA/Toyota
Co-Chair Co-Technical Secretary
Kazuki Kobayashi - NTSEL (Japan) Konrad Kolesa – OICA/Audi
The meetings of WLTP-IG were held in conjunction with the GRPE meetings that take place
in Geneva every January and June. They were supplemented by meetings every fall and
autumn to a total of 4 meetings per year. The first meeting was on 14 January of 2014 in
Geneva, the last one is on 11-12 January 2016 in Geneva. Over that period, a total of 9
WLTP-IG meetings took place. For the subgroups and taskforces the same basic terms of
reference as outlined in paragraph 3.3 were also applicable to the working groups in phase
1b.
The scope of activity was dedicated to the issues covered by the tasks of the different
subgroups and is further detailed in the next paragraphs. A separate activity is formed by the
Round Robin tests, which were conducted during the course of phase 1b by OICA. The
taskforces that were formed to deal with the open issues are listed and described in
paragraph 3.4.5.
17
All of the open issues addressed in phase 1b regarding the test cycle, gear shifting,
downscaling etc. are reported separately. At the time that this report was published the
update of the DHC report2 for phase 1b was not available. The main driving cycle related
issues that were discussed and agreed during phase 1b are the following:
a) Further downscaling in Wide Open Throttle (WOT) operation: the coefficients in the
calculation formulae were amended at the request of Contracting Party of India
b) Modifications to the gear shifting calculation tool: the 3s rule was replaced by 2 s rule,
crawler gear prescriptions were added as well as an additional safety margin for the WOT
power curve
3.4.1 Drafting GTR

Chair
Serge Dubuc – on behalf of the European Commission
The European Commission had offered to WLTP leading team to fund an expert as being the
Drafting Coordinator (DC) for the GTR. The main objective of the DC would be to coordinate
all drafting activities into a logically structured and technically, legislatively and grammatically
robust technical regulation.
To accomplish this objective, text, tables and figures resulting from decisions reached by
various technical task forces and the IG were incorporated by the DC into the GTR.
Technical gaps and inconsistencies were identified and either corrected or the responsible
person(s) made aware of these. The DC participated in those task force, subgroup and IG
meetings necessary for the processing of his task. Furthermore, the experts were
occasionally contacted directly for any necessary clarifications. To support all participants,
the GTR was uploaded on a regular basis to the UN ECE server in ‘tracked changes’ and
clean versions. The ultimate goal was to have a complete and homogeneous technical
regulation.
In addition to the above, a Drafting Subgroup was founded in January 2015 to be led by the
Drafting Coordinator in order to support him in developing the GTR. To achieve this, five
drafting sessions were held in 2015 in the form of face-to-face meetings (Brussels in March,
Stockholm in April, Brussels once again in June, and Tokyo in October) and an audio/web in
September. In all of these sessions, the Drafting Coordinator prepared a collection of
PowerPoint slides summarizing all open points and expert proposals which required
clarification. In most cases, the open points and expert proposals were clarified during these
meetings or were passed on to the IG for its consideration.
The final GTR version at the end of phase 1b was uploaded to the UN-ECE website as
formal document ECE/TRANS/WP.29/GRPE/2016/3, and is referred to as GTR no.15.9
9
Document ECE/TRANS/WP.29/GRPE/2016/3 with GTR no. 15 can be downloaded here:
http://www.unece.org/trans/main/wp29/wp29wgs/wp29grpe/grpedoc_2016.html
18
3.4.2 EV subgroup
Chair Secretary
Per Öhlund – Swedish Transport Agency Noriyuki Ichikawa, OICA-Toyota

(Sweden)
Matthias Naegeli, OICA-Volkswagen
Tetsuya Niikuni - NTSEL (Japan)
The first meeting of the WLTP EV subgroup (also referred to as E-lab subgroup) took place
at 25.03.2014. This subgroup was tasked with modifying, improving and complementing the
electrified vehicles’ test procedures which were developed in phase 1a of WLTP. In addition,
the development of compressed hydrogen fuel cell hybrid vehicles (NOVC-FCHV) test
procedure was newly added to the scope of the subgroup.
Annex 8 of the UN GTR describes the test procedures for pure electric, hybrid electric and
compressed hydrogen fuel cell hybrid vehicles. The WLTP EV subgroup was responsible for
the delivery of the GTR text on the test procedures in Annex 8 and the other parts in the GTR
related to electrified vehicles.
The scope of activities during phase 1b was described as follows:
a) Improving and complementing the test procedures for EVs which were developed in
phase 1a;
b) Developing test procedures for NOVC-FCHV;
c) Providing an additional test procedure for pure electric vehicles to allow long range
vehicles to be tested with low test burden;
d) Developing a method to obtain cycle phase specific values for electrified vehicles;
e) Implementing the interpolation approach which had been developed for conventional
vehicles during phase 1a of WLTP to electrified vehicles;
f) Improving the correction procedure for REESS energy imbalance, in particular
considering the phase specific values and NOVC-FCHV.
3.4.3 AP Taskforce
Chair Technical Secretary
Cova Astorga – EC-JRC Les Hill- Horiba
The former DTP- AP sub-group, active during phase 1a, ended its trajectory with the
validation phase (VP) for Ammonia (NH3). When the new structure for the WLTP-IG was
agreed for phase 1b (67th GRPE in November 2013), all pending commitments were
undertaken by a new AP Task Force integrated in a unique WLTP working group. From that
moment, the AP task force reported directly to WLTP chair.
The complete set of objectives envisaged by the AP subgroup at the beginning of Phase 1b
has been fulfilled:
a) To demonstrate the feasibility to measure ammonia at the vehicle exhaust with an online
measurement method;
b) To describe measurement and calibration procedures, as well as calculations, based on
existing legislation and on the output from laboratory procedures led by the AP subgroup,
in particular for the pollutant emissions of ethanol, formaldehyde and acetaldehyde.
c) Drafting GTR text protocols and procedures including new measurement, technologies
and proposing new on-line methods.
19
3.4.4 Round Robin testing
Chair
Bill Coleman, OICA – Volkswagen European Round Robin leg

Takashi Fujiwara, OICA - Honda Asian Round Robin leg
After the phase 1a version of the GTR 15 was published, a Round Robin testing activity was
planned to check the understanding and application of this GTR version in difference labs
and estimate the repeatability and reproducibility of the test procedure under type approval
conditions. The aim of this Round Robin was to deliver input based on which the GTR could
be improved during phase1b.
The original road map proposal for development of WLTP foresaw a concluding series of
tests with an open decision whether they would be confirmation tests or round robin tests or
both. At the time it was reported that traditionally the Informal Working Group would organise
and sponsor Confirmation testing where necessary and OICA would do the same for Round
Robin testing. The differences between Validation, Confirmation and Round Robin testing
are subtle, sometimes unclear and certainly overlapping. As a second phase of Validation
testing was deemed necessary it was agreed that this would also serve the purposes of
Confirmation testing, leaving OICA with the decision whether to initiate a Round Robin. They
considered that a Round Robin testing activity would be valuable and decided to support this.
There can be many reasons to perform Round Robin testing such as:
a) checking repeatability and/or reproducibility of the test results, and/or
b) focussing on the use of physical equipment (vehicles, labs or test equipment), and/or
c) focussing on how of procedures are interpreted and applied.
These reasons obviously affect the instructions for conducting the round robin tests, the
selection of the vehicles, fuels and tests themselves, the instructions to the accompanying
engineer and many other aspects. As some of these objectives and decisions are
contradictory it is impossible to cover everything in the round robin, hence some questions
remain unanswered.
It is difficult to plan the timing of a round robin as it generally involves a vehicle being
transported between laboratories, which is a time-consuming process that cannot be easily
shortened. At the same time the concept of round robin testing requires a level of stability in
the subject being studied and therefore cannot start before the legislative development is
very mature. Thirdly there is normally more political pressure towards the point where the
legislation needs to be completed in order to be able to implement it. These contradictory
constraints lead to the conclusion that the timing of the round robin is always a compromise.
The following decisions were taken towards the end of phase 1a:
 Round Robin testing is considered necessary and is desired by experts.
 A worldwide Round Robin would take so much time that the results could not be
considered within the development period of WLTP.
 Therefore an Asian and a European Round Robin leg would be performed with a level of
interaction between accompanying engineers and some vehicle overlap towards the end
of the regional testing.
 As little or no new measurement technology is prescribed by WLTP, the focus would be
on the operation of the tests (as recorded by the accompanying engineer), with aim to
reveal the test requirements that might be misinterpreted or are not complete.
ACEA took the role of coordinating and sponsoring a European Round Robin for which 2
vehicles were sourced, one with a petrol engine and automatic transmission and one with a
20
diesel engine and manual transmission. The French technical service UTAC was contracted
to supply the golden engineer and Celine Vallaude was allocated to the task. Labs from both
the automotive industry and from authorities participated in the testing.
European Round Robin Asian Round Robin

 Check the understanding and the application of the GTR15 (based on
Objectives
phase 1a text) in different labs
 Estimate the repeatability and reproducibility of the test procedure
under type approval conditions
Participants BMW, FIAT, UTAC, PSA, Japan: JARI, NTSEL, TOYOTA,
Daimler, Bosmal, Horiba, DEKRA, India: ARAI
VW, TÜV Nord, JRC,
Korea: NIER, KEMCO, KATRI
China: CATARC, CRAES
Test vehicles BMW 116i 1.6L Petrol 6MT Toyota WISH 1.8L CVT - Petrol
Alfa Romeo Giulietta 2.0L Diesel Mahindra & Mahindra XUV500 2.2L
6MT -Diesel
Nr. of tests at 3 (mostly) 3 (mostly)

each laboratory
Expected January 2016 January 2016

completion (additional testing after January (additional testing after January may
timing may be performed at India and be performed at India and Europe.
Europe.
Table 1: Round Robin overview (performed in two parallel and linked legs in
Europe and Asia)
As neither of the Round Robin legs was completed before the final Informal Working Group
meeting of phase 1 (WLTP IG meeting 12, Sept./Oct. 2015 in Tokyo), it is currently only
possible to deliver interim results.
The European golden engineer (Celine Vallaude, UTAC, France) reported instances of
participating laboratories where the facilities were not yet upgraded to a WLTP standard and
also inconsistencies of interpretation of the GTR text between participating laboratories.
The Japanese diamond engineer (Takahiro Haniu, JASIC/JARI, Japan) reported that the
Asian Round Robin tests would be completed by the January 2016 with the participation of
China, India, Japan and Korea. Two test vehicles are used for the testing (see Table 1).
Even though it is found that there are also several laboratories whose facilities are not yet
upgraded, no other urgent issues have come up yet that would lead to a change of the
current GTR text.
21
The following are examples of issues that were found during the European and Asian Round
Robin tests so far:
 The rotating inertia mass was not used appropriately at most of laboratories. This
issue has been taken care of by improving the GTR text to be more specific
description.
 Because the gear shift tool was under development during the round robin testing,
different versions of this tool have been used by the laboratories. The final version of
the gear shift tool is expected to be released soon.
 The measurement equipment for the RCB correction was not prepared by all
laboratories during the Round Robin testing and the necessity of such a stringency on
the equipment requirements was questioned. The required accuracy of the equipment
was reviewed and revised in the final GTR text.
 The vehicle warm-up just before the coastdown for road-load determination on the
chassis dyno had not been performed at one laboratory. It was corrected to what the
GTR described.
It is expected that more issues will be raised towards the completion of the Round Robin
tests, and they should be taken care during WLTP phase 2.
The full analyses of both legs of testing should be combined on completion of the testing and
reported during the informal working group meeting in early 2016. Recommendations should
be made for improvement of the GTR text during phase 2.
3.4.5 Taskforces on open issues

The remaining open issues from phase 1a were clustered, and then assigned to dedicated
taskforces. For each taskforce a suitable taskforce leader was appointed, and interested
stakeholders could join the group. The assignment for each taskforce was formulated as to
discuss the issues they were tasked with, work out possible solutions, and come forward with
an agreed proposal to the WLTP-IG. After approval by the IG, the proposal would then be
worked out into a draft text for the GTR.
For a complete overview of the open issues table (OIT) please refer to document WLTP-12-
03 at the UN-ECE website.10
An overview of the main topics that were addressed by the taskforces in phase 1b and were
added to the GTR is presented in Table 2; also a reference is included to the paragraph
where this issue is further detailed. Issues which have led to the introduction of a new
concept to the testprocedure of the GTR (with respect to the emission test procedures
currently in use) are described under paragraph 4.4: New concepts of the GTR. The
remaining issues are outlined in the following subparagraphs.
10
https://www2.unece.org/wiki/display/trans/WLTP+12th+session
22
Conventional ICE vehicles
Issue Paragraph Taskforce leader
Reference Fuels 3.4.5.1 William Coleman, OICA
Definitions 3.4.5.2 William Coleman, OICA
Normalisation 3.4.5.3 Nikolaus Steininger, EC
Number of tests 3.4.5.4 Takashi Fujiwara,

OICA-Honda
Review of coastdown tolerances 3.4.5.5 Rob Cuelenaere, TNO
Calculation and interpolation of fuel consumption 3.4.5.6 Konrad Kolesa, OICA-

Audi
Speed trace tolerance / speed trace index 3.4.5.7 Noriyuki Ichikawa,

OICA-Toyota
On-board anemometry and wind speed conditions 4.4.7 Rob Cuelenaere, TNO
Interpolation family and road load family concept 4.4.9 Rob Cuelenaere, TNO
Torque meter method 4.4.10 Rob Cuelenaere, TNO
Wind tunnel as alternative method for road load 4.4.11 Rob Cuelenaere, TNO
determination
Alternative Cd.A determination 4.4.12 Rob Cuelenaere, TNO
Road load matrix family 4.4.13 Rob Cuelenaere, TNO

Appendix 2
Manufacturer responsibility on road load 4.4.14 Rob Cuelenaere, TNO
Alternative vehicle warm-up procedure 4.4.15 Rob Cuelenaere, TNO
REESS charge balance (RCB) for ICE vehicles 4.4.16 Rob Cuelenaere, TNO
WLTP post-processing 4.4.24 Christoph Lüginger,

OICA - BMW
23
Electrified Vehicles (E-lab expert group)
Issue
Taskforce leader
Utility factors 3.4.5.8 Tetsuya Niikuni, NTSEL

Appendix 1 Japan
Mode selection and predominant mode 3.4.5.10 Tetsuya Niikuni, NTSEL

Japan
RCB correction for OVC-HEVs, NOVC-HEVs and 4.4.18 Tetsuya Niikuni, NTSEL
NOVC-FCHV’s Japan
Shortened test procedure for PEV range test 4.4.19 Tetsuya Niikuni, NTSEL
Japan
Phase-specific values for EVs 4.4.20 Tetsuya Niikuni, NTSEL

Japan
Interpolation approach for EVs 4.4.21 Tetsuya Niikuni, NTSEL

Japan
End of PEV range criteria 4.4.22 Tetsuya Niikuni, NTSEL

Japan
Fuel Cell Vehicle test procedure 4.4.23 Tetsuya Niikuni, NTSEL

Japan
WLTP post-processing 4.4.24 Nico Schütze, OICA -
BMW
Alternative Pollutants (AP)
Issue Paragraph Taskforce leader
Measurement method for ammonia, ethanol, 3.4.5.9 Cova Astorga, EC-JRC

formaldehyde and acetaldehyde
Table 2: Overview of taskforces to work on open issues and the responsible

taskforce leader; references to the respective paragraphs are included
The next paragraphs will describe the scope and the results of what was developed by the
taskforces on the open issues. Those issues that led to the introduction of a new concept to
the GTR will not be described here, but have been added as subparagraph to paragraph 4.4.
See Table 2 for an overview of the reference paragraphs.
3.4.5.1 Reference Fuels

In phase 1b no activity was anticipated other than drafting for correction of errors and
continuing an advisory role for the WLTP experts and the Round Robin participants
As indicated in paragraph 3.3.4 it was not possible to establish a cooperation with the fuel
production industry to fulfil the scope of the RF subgroup. Therefore it was not feasible to
obtain within WLTP an approval on the technical scope of the validation fuels and their
relevant properties.
24
In practice however the list of reference fuels included in the GTR now serves as a guideline,
albeit non-binding. Validation was performed on the local reference fuels of the participating
regions and the current disharmonization of drive cycles within GTR combined with a
foreseeable continuing disharmonization of reference fuels, particularly regarding the bio-fuel
content, renders a cross regional validation of the cycles and procedures somewhat
irrelevant. Thereby the RF scope of activity points (b) to (e) listed in paragraph 3.3.4 will not
be pursued unless the situation changes significantly.
The fuels experts from OICA will remain available to fulfil the role described in point (f)
3.4.5.2 Definitions
It was recognised at the conclusion of phase 1a that there was need for review and revision
of many of the definitions that were included in the first version of the GTR. The subject
areas where such actions were deemed necessary are as follows:
a) Definitions addressed by the Informal Working Group VPSD (Vehicle Propulsion System
Definitions)
b) Definitions of Masses
c) Definitions concerning the measurement of Particulate and Particles (PM/PN)
d) Definitions regarding road load
e) Definitions where the wording had light differences from those currently used in other
Regulations under the responsibility of GRPE
f) Definitions where improvement was possible regarding the language or structure of the
text
Finally, during the work of the IG-VPSD, advice was taken on better definitions from the UN-
ECE secretariat and from the EU Commission legal services. This included keeping to
defining terms without including prescriptive technical requirements, trying to keep where
possible to one sentence and avoiding the use of examples unless absolutely necessary for
clarity. These pieces of advice were applied to a number of definitions and the amendments
were subsequently adopted.
Further detailed information on items a) through d) is provided in the following section:
VPSD
The IG-VPSD agreed on a set of definitions which differentiate between fundamental
definitions of elements (e.g. energy storage system) and those elements which are used for
propulsion (e.g. propulsion energy storage system). This differentiation is justified and helpful
and the VPSD definitions were therefore largely adopted into the GTR. However, IG-VPSD
also agreed on some definitions that mixed the concepts of fundamental definitions and
propulsion systems (e.g. "Electric machine" means a propulsion energy converter
transforming between electrical and mechanical energy). The IG-WLTP found these not to be
helpful rather confusing and they were not adopted.
Masses
As the concept of the combined approach to determination of CO2 values (later renamed
interpolation approach) was finalised late in the development of phase 1a, the definitions of
vehicle masses did not necessarily reflect the whole concept that had been conceived. One
significant contributor to this discrepancy was a concept that vehicles ‘High’ and ‘Low’ should
be the absolute best and worst cases of the vehicle family. This concept contradicts the
25
decisionn in phase 1a that extrapolation of o CO2 valu ues should be allowed, within a to olerance
band. A further diffficulty was tthe decisionn that the mass
m repressentative of vehicle load, which
contribu
utes to the vehicle
v testt mass, sho
ould be a sinngle value forf the family derived from
f the
heaviesst vehicle. A solution was
w found byy adopting the t current European definitions
d o “mass
of
in runniing order”, “mass of th he optional equipment” and “tech hnically perrmissible maximum
m
laden mmass” as thet basis for
f develop ping test mass
m definittions. Durin
ng this discussion,
inconsisstencies weere identifieed in the Eu uropean deffinitions and the EU agreed
a to addopt the
changess to its deffinition of “mass of the optional equipment”” in the reg gional legisslation in
order too remove these
t inconnsistencies. This set of definition ns allowed description n of the
heaviesst and lighttest vehicle es covered by an app proval, while the Euro opean defin nition of
“actual mass of thee vehicle” was
w adopted d to permit definition
d off the test mass of an inndividual
vehicle.
ole set of ve
The who ehicle masss definitions
s and their interconnecctions is sho
own in Figurre 6:
Figure 6: Overview of the mass

m defin
nitions that build toge
ether the ve
ehicle test mass
PM/PN
A drafting review of GTR phase 1a revealed to t the non n-expert rea ader a number of
inconsisstencies in the use of terminology. The assiistance of the t IG-PMP P was nece essary to
solve these. Thiss expert and non-exxpert review w finally concluded
c s
some fund damental
problem ms such as potential different und derstandingss of the abbreviation PM P (e.g. pa articulate
matter oor particulatte mass). Potential
P solutions were e considered such as “mass
“ of pa articulate
matter”,, “PM-mass s” and otherrs. The iden ntification off a second fundamenta
f al problem, that two
massess were being g referred to
o in this conntext, the mass
m collectted on the sample
s filterr (in mg)
and the e distance specific
s parrticulate maass emissio ons (in mg//km), led to o the set off PM/PN
definitio
ons that werre adopted aas follows:
26
3.6.1. "Particle number emissions" (PN) means the total number of solid particles emitted
from the vehicle exhaust quantified according to the dilution, sampling and
measurement methods as specified in this GTR.
3.6.2. "Particulate matter emissions " (PM) means the mass of any particulate material from
the vehicle exhaust quantified according to the dilution, sampling and measurement
methods as specified in this GTR;
A further piece of explanatory text delivered by IG-PMP which clarifies the difference
between particles and particulate was also found to be very helpful and was included in the
definition section of the GTR. The following clarification is now included in paragraph 3.6 of
the GTR’s definition section:
“The term ‘particle’ is conventionally used for the matter being characterised (measured)
in the airborne phase (suspended matter), and the term ‘particulate’ for the deposited
matter.”
Road load
Some definitions of elements of the road load of vehicles were identified by industry experts
to be physically incorrect. These were corrected by the Annex 4 task force and adopted. See
also paragraph 3.4.5.5.
3.4.5.3 Normalization
Background
During phase1a, WLTP IG has already adopted many new elements to reduce testing
flexibilities and tolerances, such as the soaking and test room temperature, test mass
determination, the vehicle warm-up procedure, road-load calculation formula and so on.
Within the framework of a test procedure it is inevitable to allow tolerances in order to get to
a valid test result in a real test environment, because it is simply not possible to execute the
test procedure exactly according to what is prescribed. For example, the test driver will try to
follow the target speed trace as well as possible, but is unable to match this completely.
However, such tolerances may lead to test-to-test variations of the quantitative test cycle
results, in particular CO2 emissions. Even worse, if the tolerances are set too wide they offer
the possibility of being exploited systematically to obtain better test results. The repeatability
of the test procedure would be increased if the test results are corrected for any (systematic)
deviation from the target value. Correction methods for the used tolerances can therefore
improve the quality of quantitative predictions of the cycle results and also render the
systematic use of tolerances unattractive.
This issue was raised by the European Commission as an issue that needs to be addressed
in phase 1b. The EC assigned a contractor to develop a report on such correction methods11,
and these would serve as input for the discussions within the group. The report investigates
a series of corrections that could be applied to variations of test parameters within the
tolerance ranges allowed by the WLTP GTR provisions. During phase 1b the concept of
applying correction methods or algorithms was referred to as ‘normalization’.
11
See document WLTP-08-39e at https://www2.unece.org/wiki/display/trans/WLTP+8th+session
27
Correction algorithms
Table 3 gives an overview of the parameters for which the report has suggested correction
algorithms. It also provides a suggestion by the European Commission on the priority level,
and the estimated impact of the tolerance on the CO2 emission according to the following
labeling system for recommendation:
A = integrate as soon as possible into European transposition of the WLTP & propose for
integration (possibly with some minor amendments) into WLTP GTR phase 1b
B = investigate further for integration in WLTP GTR phase 1b (the result of these
investigations could also be that the respective correction is not applied)
C= investigate further for integration in WLTP GTR, probably within a time frame
exceeding phase 1b (the result of these investigations could also be that the respective
correction is not applied)
D= no further investigation since effect appears to be small and/or very complex to
address
Correction type Recommendation Comment

(reference in the report)
2.2 Deviation from target A The method for addressing the

speed (including battery issue is fully developed in the
SOC correction) report, relevant impact on CO2
emissions of up to 5% (deviation
from target speed and battery SOC
correction)
2.3 Quality of reference fuel B Impact on CO2 emissions still to be

investigated
2.4 Inlet air temperature B Impact on CO2 emissions for diesel

and humidity vehicles seems to be very low, for
gasoline vehicles relevant up to 2%
2.6 Temperature from D Very small impact on CO2

preconditioning and soak emissions, < 0,4%
2.7 Inaccuracy of road load B (withdrawn) Several options for addressing the
setting on the chassis dyno issue are available, relevant impact
on CO2 emissions of up to 3%
2.9 Deviation from C There seems to be a relevant

designated gear shift points influence on CO2 emissions,
however there are no ideas yet how
the issue could be addressed
4.1 Vehicle preparation for A There is a relevant influence of the

coast down, toe-in wheel alignment on road load
prescription coefficients, the requirement is easy
to implement
28
4.2 Vehicle conditioning for B or C There is a relevant influence of the
coast down: tyre pressure wheel alignment on road load
monitoring/control coefficients, the requirements
suggested are not so
straightforward to implement
5.1 Ambient weather B or C There is a relevant influence of

conditions at coast down: these parameters on the air drag
temperature, air pressure, measured at coast downs. In the
water content of the air current WLTP GTR there is already
a correction for the air density, but
this may not be sufficient.
5.2 Wind corrections at B or C Albeit the current WLTP GTR

coast down already contains a wind correction
further restrictions on side wind and
gustiness may be necessary.
5.3 Road condition of coast C The road surface of the test track
down test track (surface seems to have a significant
roughness, gradient, influence on the road load
undulation) parameter F0. It should therefore be
envisaged to either require a
minimum road "roughness" or to
correct road loads measured at a
given test track against a "standard"
road surface. However, the
investigation of relevant roughness
parameters and "standard" road
surface values is quite complex.
6.2 Rotational inertia A The suggested correction is very

correction (when evaluating simple to implement and provides a
the coast down test) more accurate result for CO2
emissions.
Table 3: Correction parameters, priorities and impact on CO2
Implementation into phase 1b GTR

It was recommended by the IG that corrections labelled with ‘A’ be implemented in the GTR
in phase 1b, and that the feasibility for implementation of ‘B’ items would be investigated. All
items ‘C’ were considered to be out of scope for phase 1b. For the deviation from the target
speed curve (item 2.2) a separate taskforce was started by Japan, see paragraph 3.4.5.7 on
Speed trace tolerance / drive trace index. However, a drive trace energy correction was
postponed to phase 2.
The other ‘A’ labelled items were concluded as follows:
a) The proposal on wheel alignment was adopted for implementation in the GTR (item 4.1).
b) A correction for the rotational inertia by weighing the test tyres was rejected (item 6.2).
In response to the issue of the inaccuracy of the road load setting (item 2.7) a proposal was
adopted to limit the time gap between warm-up and chassis dynamometer setting to 120
29
seconds, and a maximum of 60 seconds between consecutive coast-down runs for the dyno
setting procedure. In addition, a separate action by Audi was initiated to assess the
tolerances in the road load determination procedure, with the aim to reduce the tolerances
where possible. This information is included in paragraph 3.4.5.5 (Review of coastdown
tolerances).
All the other proposed correction algorithms to normalize the test results were postponed to
phase 2, mainly because there was no time available to validate these methods and there
was a lack of information and data on the effects on electrified vehicles. For phase 2 it has
yet to be decided which of these items are taken up into the scope for further analysis.
3.4.5.4 Number of tests

During phase1a there was no consideration on the number of tests needed for the type
approval process and on how to determine the final type approval value from the tests. To
address these issues, a taskforce was formed which was led by Takashi Fujiwara (OICA,
Honda).
The current Regulation R101 allows for a 4% CO2 tolerance, which means that if the CO2 test
result during type approval test is within 4% of the manufacture declared value, the declared
value will be accepted as the type approval value. Originally intended to reduce the testing
burden in the case that a vehicle is slightly modified, this tolerance is now used as a loophole
to artificially declare a lower CO2 emission as the actual vehicle performance. Therefore it
was necessary to close this loophole by tightening the type approval system on this point. At
the same time this will increase the representativeness of the test result, which helps to
produce reliable consumer information.
Though the taskforce had the scope to provide a technical solution, the ‘number of tests’
issue proved to have a political component as well. This political discussion was largely
driven by the different way in which the type approval process takes place in Europe and
Japan. While type approval testing in Europe takes largely place under responsibility of the
manufacturer and is only witnessed by the type approval authority, the Japanese TAA is
much more in control over the tests.
The discussions in the taskforce therefore mainly focused on the CO2 tolerance value
(referred to as ‘dCO2‘). The European Commission proposed to introduce a ‘safety margin’
which requires manufacturers to demonstrate a better CO2 than the manufacturer declared
value at type approval. Japan initially proposed a tolerance of 1.8%, but later proposed to
take the CO2 tolerance out completely as a compromise solution. However, the European
Commission could not agree to abandon their requested safety margin. There were long
controversial discussions taking place in the several task force and informal working group
meetings, but no agreement could be achieved on a harmonized CO2 tolerance value.
Acknowledging the differences between the regional type approval systems, it was finally
decided during the last WLTP-IG meeting in Tokyo that the CO2 tolerance value would be an
option for the CP. Even though this leads to further disharmonization between the regions, it
can be seen as a an acceptable solution by considering that the same stringency of the type
approval process in different regions would actually require different tolerance values.
These are the main conclusions that were agreed at the end of phase 1b:
a) Remove the 4% CO2 tolerance. The tolerance value will be determined by each
Contracting Party (CP), but dCO2 has to be within a range of -1.0% to +2.0%.
b) Electric energy consumption, all electric range and pure electric range are added for
evaluation of the performance of electrified vehicles, and a 0% tolerance is allowed for
any of those parameters.
c) Criteria pollutant limits should fulfilled during each of the type approval tests.
30
The process of the number of tests as described in paragraph 1.1.2. of Annex 6 consists of
the following steps:
step 1: The manufacturer declares the required values according to Table A6/1.
step 2: Perform type approval test(s) according to Figure A6/1 (flowchart).
Step 3: If the criteria according to Table A6/2 are fulfilled, the declared values are accepted
as type approval values for the total cycle.
step 4: The phase specific values are determined based on the test results, and are
adjusted by the distance between the type approval value of the total cycle value
and average of the total cycle test results.
The following statistical data were employed to derive the range of allowed CO2 tolerance.
 The test-to-test variation (one sigma) of CO2 emission is 0.9%. This value was confirmed
by tests performed both in Europe and Japan.
 Under the assumption this test-to-test variation, the expected number of tests during type
approval testing will be 2.5 on average in case of European proposal (i.e. -1.0% tolerance
for the first test and -0.5% tolerance for the second test), and 1.8 on average in the case
of the compromise proposal (i.e. 0.0% tolerance for all test).
The expected number of tests as a function of the chosen dCO2 values for the first and
second test are shown in Figure 7.
vehicles need second tests 3,0
Expected Number of tests
120% 2,5
vehicles need three tests 2,3
100% Expected number of tests 2,1
86,7%
1,8 2,0
80% 71,1% 71,1% 67,9%
1,4 55,7%
60% 1,0 1,1 50,0%
40% 35,5% 1,0
28,9% 25,0%
20% 13,3%
2,3% 0,0% 6,2%
0,8%
0% 0,0
dCO2_2 [%] 1.8 1.0 0.5 0.0 0.0 ‐0.5 ‐0.5
dCO2_1 [%] 1.8 1.0 0.5 0.0 ‐0.5 ‐0.5 ‐1.0
JPN 1st Compromised EU

proposal proposal proposal
Assumptions for calculation;
• Average (µ) = 0
• Standard deviation (σ) = 0.9 %
• Judgment for second test is based on the average value of first and second tests.
(i.e. σ for second test = 0.9/root (2) %)
Figure 7: Expected number of tests for ICE vehicles as a function of dCO2
31
During phase 1b it proved not possible to incorporate criteria for NOVC-FCHV vehicles. It is
foreseen to add these during WLTP phase 2, in which case they could simply be added to
Table A6/1 and A6/2.
3.4.5.5 Review of coastdown tolerances

During phase 1b a need was identified to review the tolerances allowed for the different road
load determination methods offered in Annex 4. The main purposes were to tighten the
tolerances where possible, to make the requirements more explicit, and to align the
tolerances between these methods. A proposal was prepared by BMW in July 2015, with a
large number of suggested improvements. Most of these improvements were accepted
without any further discussion. The remaining ones were discussed and agreed during a face
to face meeting.
These are some examples of the improvements that were agreed12:
 Adding frequencies at which parameters should be measured (speed, torque,
temperatures, pressure, wind direction, etc.)
 Deleting double tolerances, keeping the most stringent one
 Setting time windows for stationary anemometry wind speed criteria
 Specifying tyre pressure per axle
 Corrections for measurement equipment installed to the vehicle exterior
 Setting restrictions to the amount of rejected pairs of coastdown measurements.
On two issues it was not possible to reach agreement on the suggested improvement:
a) The limitation of the split run coastdown to a maximum of 3 parts and conditions to
ensure a smooth connection of these parts in par. 4.3.1.3.4.
b) The limitation of the atmospheric temperature to 30°C in par. 4.1.1.2
There was one other issue introduced in phase 1b that should be mentioned here, which is
closely related to this review of tolerances. This concerns the selection of reference speeds
for road load determination. It was decided to set fixed reference speed points eliminating
variation of the resulting road load coefficients by the choice of the reference speed points
and evaluation range. Reference speed points now start at 20 km/h and go up in fixed
incremental steps of 10 km/h. These increments were earlier free to choose, but limited to a
maximum of 20 km/h. The highest reference speed depends on the applicable test and on
the maximum vehicle speed. Since the number of reference points is increased, this means
the second order polynomial road load function is more accurately defined. At the choice of
the manufacturer he may also elect higher reference speeds –up to a maximum of 130 km/h-
to use the same road load measurement for type-approval in different regions with a different
applicable cycle.
3.4.5.6 Fuel consumption calculation

Since fuel consumption cannot be measured directly without installation of measuring
devices in each tested vehicle, the fuel consumption is calculated from the measured
emissions of hydrocarbons, carbon monoxide, and carbon dioxide. For each of the reference
fuels listed in Annex 3, specific H/C and O/C ratios are provided in the calculation formulas.
12
See document WLTP-12-26e at https://www2.unece.org/wiki/display/trans/WLTP+12th+session for a
complete overview.
32
A general equation to calculate fuel consumption for any other test fuel using the actual H/C
and O/C ratios is included as well.
The calculation of fuel consumption for individual vehicles within the interpolation family
follows the same interpolation method as applied for CO2 emissions, based on the fuel
consumption of vehicle H and vehicle L. Differences of HC/CO levels of vehicles within the
Interpolation family were considered as of minor significance. Determination of phase
specific values follows the principle of CO2 interpolation.
The fuel consumption calculation is included in paragraph 6 of Annex 7.
3.4.5.7 Speed trace tolerance / drive trace index

One of the main purposes of WLTP is to reduce the flexibilities which are allowed as test
tolerance.
During Phase1a, WLTP has already adopted many new elements to reduce the testing
flexibilities, such as the soaking and test room temperature, test mass determination, the
vehicle warm-up procedure, road-load calculation formula and so on. “Normalization” was
also one of items for discussion and concrete normalization methods were studied to correct
measurement results for any used tolerance, see paragraph 3.4.5.3.
Along the same line, the WLTP Technical Secretary(TS) has proposed during Phase1b to
implement the “drive trace index” which can be applied for all type of vehicles in order to
reduce the testing flexibility regarding the drive trace tolerance13.
WLTP IG requested to establish a Task-Force(TF) on the drive trace index to work out a
proposal for adoption. Its members mainly consisted of industry experts and a number of
meeting were held to develop the final proposal.
The driving technique during the test (smooth or rough) within the drive trace tolerance has a
significant impact on fuel consumption and CO2.
This is illustrated in Figure 8 and Figure 913,15.
13
Refer to document WLTP-10-31e at https://www2.unece.org/wiki/display/trans/WLTP+10th+session
33
Figure 8: Examp ples of diffe
erent drivin
ng behavio he speed trrace tolerance: a
or within th
‘devil’ manufactu urer trying
g to improvve CO2 pe erformancee and an ‘angel’
manufa acturer following the trace as well
w as poss sible
10,0
rough driving
8,0 within tollerance
6,0
4,0
∆CO2 (g/km)
Approx.
2,0 no
ormal drivingg
15g/km
0,0
-2,0
-4,0
-6,0
-8,0 smooth driving
within tole
erance
-10,0
-10,0 -5
5,0 0,0
0 5
5,0 10,0
I
Index
Figure 9: Effect on CO2 off normal, ‘smooth’

‘ and ‘rough’ driving within
a w the speed
olerance
trace to
34
This leads to an increase of the test-to-test variation but also to unfair competition. Since the
WLTC that was developed is a micro-transient type of test cycle, the current situation may
become worse since there is more potential gain in smooth driving.
On the other hand, Figure 9 also indicate that test-to-test variation is negligible when the
driving indexes are close to zero (‘normal driving’), which means the actual drive trace is
close to the prescribed cycle. Therefore, if an appropriate drive trace index(es) are chosen, it
can be expected that the flexibility caused by a smooth driving technique will be reduced.
The following elements were introduced and discussed within the Taskforce:
(1) Different drive trace indexes as a reference, according to the Table 4 below. Each
index is a kind of quantitative discrepancy between the actual drive trace and the
prescribed trace
(2) Keep drive trace tolerance to check the test validity but not showing this tolerance on
on the driver aid (the monitor that shows target and actual speed).
Index Name Description
ER Energy Rating Percentage difference between the total driven and

target cycle energy
DR Distance Rating Percentage difference between the total driven and

scheduled distance
EER Energy Economy Percentage difference between the distance per unit
Rating cycle energy for the driven and target traces
ASCR Absolute Speed percentage difference between the ASC

Change Rating
for the driven and target traces
IWR Inertial Work Rating percentage difference between the inertial work for the
driven and target traces
RMSSE Root Mean Squared performance indicator to meet the target speed trace
Error throughout the test
Table 4: Evaluated drive trace indexes
The calculation of the drive trace indexes was proposed to be done according to the
following procedure:
(1) Correct the actual drive trace data during homologation tests towards 10Hz (no more
than 10Hz and no less than 10Hz in order to be compatible with different laboratories)
(2) Apply a linear interpolation method of the prescribed drive cycle to convert it to 10 Hz
(3) Data filtering shall be done according to SAE J2951
(4) Each index is calculated according to SAE J295114
14
For the calculation formulas refer to slide 8-11 of document WLTP-12-27
35
Both ACEA and JAMA did some data evaluation studies on measured vehicles to find out if
these drives speed trace indexes would qualify as good indicators of the driving behavior
during the test. The results of these studies were presented in the Taskforce15.
Since no agreement on the specific criteria for these indexes was reached, the Taskforce
had to make the decision not to define specific criteria at this stage and to apply all possible
index values as a reference. It was also decided that the drive trace tolerance would not be
shown on the driver’s aid monitor, to avoid that this tolerance would be exploited during the
test.
Since drive trace indexes are now included in the GTR as a reference parameters, the
following future scenario is foreseen for Phase 2:
a) Gather drive trace index data from homologation tests in a database
b) Select from the database the most suitable index(es) and accompanying index criteria to
check the test validity.
c) At the same time, study “Normalization” methods for differences between target and
actual speed (especially for electrified vehicles)
d) Consider which method is better from the view of eliminating flexibilities and testing
practicability.
e) Implement either drive trace index(es) with the specific criteria or normalization
procedures in the GTR.
3.4.5.8 Utility Factors

A conventional vehicle with an internal combustion engine (ICE) will only consume fuel, while
a pure electric vehicle (PEV) will only have an electric energy consumption. But hybrid
electric vehicles16 may have a combination of electric energy and fuel consumption during
the type approval test. These vehicles can be operated in two different driving modes:
a) Charge depleting mode, during which energy is drawn from the rechargeable energy
storage system (REESS), and a
b) Charge sustaining mode, during which the stored energy in the REESS remains on
average constant.
The extent to which a vehicle during real world operation is driven in either of these modes
depends on the following factors, related to the layout of the driveline and the characteristics
of the trips:
 The capacity of the electric energy storage system;
 The electric energy consumption of the vehicle while driving in charge depleting mode;
 The distance that the vehicle is able to drive in charge depleting mode (resulting from the
first two factors);
15
Summary results of these studies can be found in document WLTP-11-21e (slide12&13) and in
document WLTP-11-22e (slide 3to9) at https://www2.unece.org/wiki/display/trans/WLTP+11th+session
16
There is a distinction between two types of hybrid electric vehicles in the GTR: off-vehicle chargeable
(OVC-HEV) and the non off-vehicle chargeable hybrid electric vehicle (NOVC-HEV). The OVC-HEV
is also indicated as a plug-in hybrid on the automotive market.
36
 The length and frequency distribution of trips made with the vehicle;
 The (off-vehicle) charging frequency for the electrical energy storage system.
The share between driving in ‘charge depleting’ and ‘charge sustaining’ mode can be
calculated from these factors, and is expressed as the ‘Utility Factor’ (UF). The UF is
therefore defined as the ratio between the distance driven in ‘charge depleting’ mode divided
by the total driven distance. The UF can range from 0 (e.g. for a conventional vehicle or for
an HEV) to 1 (for a pure electric vehicle or OVC-HEV that is driven in charge depleting mode
only). It is not a constant value, but is a function of the measured range that was driven in
charge depleting mode on the WLTC.
Since the fuel and energy consumption, as well as the emissions, are very different between
the two driving modes, Utility Factors are needed in order to calculate weighted emissions,
electric energy consumption, fuel consumption and CO2 values. UFs are based on fleet data
and driving statistics such as average daily trip length, average speed, road type distribution,
etc. From these data, a Utility Factor (UF) curve can be generated which facilitates a
weighting between the measured values of pollutant emissions, electric consumption, CO2
emissions and fuel consumption for the two driving modes (‘charge-depleting’ and ‘charge-
sustaining’).
During the discussions on the Utility Factors in phase 1b of WLTP, it became clear that there
was no consensus on a harmonized UF curve. This is largely a result of the fact that driving
statistics may differ significantly between the world regions, and they have a large effect on
the UF curve. Instead of having one uniform UF curve in the GTR, each contracting party
may develop its own UF curve based on the regional driving statistics. However, it was
decided that at least the methodology for the determination of driving statistics and the
development of regional Utility Factors should be harmonised. Appendix 5 of Annex 8
prescribes the methodology which is mainly based on SAE J2841 (Sept. 2010, Issued 2009-
03, Revised 2010-09). The UF curve itself is parametrized into 10 coefficients, listed in Table
A8.App5/1 of the that appendix.
Appendix 1 of this Technical Report describes the methodology that was applied to
determine the UF curve for the European Union in detail, and is intended to provide a
template for the UF curve determination in other regions.
3.4.5.9 Additional pollutants

The work of this taskforce was structured according to the objectives that were set for phase
1b of WLTP:
1) To demonstrate the feasibility to measure ammonia at the vehicle exhaust with an online
measurement method;
2) To describe measurement and calibration procedures, as well as calculations, based on
existing legislation and on the output from laboratory procedures led by the AP subgroup,
in particular for the pollutant emissions of ethanol, formaldehyde and acetaldehyde.
3) Drafting GTR text protocols and procedures including new measurement, technologies
and proposing new on-line methods.
This paragraph will report on each of these objectives separately.
37
Ammonia
The phase 1a version of the GTR describes effective methods for measuring Ammonia from
LD vehicles. The feasibility of these methods was assessed during phase 1b by validation of
testing procedures
An experimental validation phase was performed for the new driving cycle (WLTC) in the
Vehicle Emission Laboratory (VELA) at the European Commission Joint Research Centre
(EC-JRC Ispra, Italy). This was done to understand the feasibility of measuring some new
pollutants in the gaseous phase exhaust of LD vehicles and eventually, how to incorporate
the text to the GTR during phase 1b.
Conclusions of the work of the AP Taskforce about the measurements of NH3 in LD exhaust
were drawn and transposed into a draft text for the GTR. The document with the information
reporting the validation phase for different analytical instrumentation measuring ammonia
from LD exhaust during WLTC from the campaign was uploaded to the UNECE website17.
Summary of Validation phase results for NH3
Four light duty vehicles were tested as part of the Validation Phase (VP). The raw vehicles’
exhaust gas was analyzed in real-time using different instruments (FT-IR, Quantum Cascade
Laser Infra-Red Spectrometer-QCL-IR and an integrated photo-acoustical analyzer with a
Quantum Cascade Laser).
The obtained average ammonia concentrations and the emission profiles revealed that the
three instruments were suitable to measure ammonia from the vehicles raw exhaust. The
results showed that all instruments were in good agreement, presenting no significant
differences. The three instruments also presented very good reproducibility. The results
indicate that temperature of the sampling and analyzer is not important as long as there is no
condensation.
The following was achieved on NH3 measurements in the gas phase of LD vehicles’ exhaust.
1. The VP demonstrated that is perfectly feasible to measure ammonia at the vehicle
exhaust with an online method guaranteeing the reproducibility and repeatability of the
results.18
2. The VP confirmed that three instruments are validated as a measurement method for
NH3 in the GTR.
Ethanol, formaldehyde and acetaldehyde

A new validation Phase during phase 1b focussed on finding new and alternative on-line
methods for ethanol, formaldehyde and acetaldehyde to find out if they would qualify for the
WLTP GTR.
An intercomparison exercise of the WLTP test was conducted in the VELA laboratories
(JRC-IET Sustainable Transport Unit), aiming at measuring ethanol, formaldehyde and
acetaldehyde emissions from a flex-fuel light-duty vehicle using E85 fuel. All instruments
participating in the intercomparison allowed in situ measurements of these compounds
directly from the diluted exhaust gas at the CVS, as it was established in the scope of this
validation phase campaign.
17
https://www2.unece.org/wiki/display/trans/WLTP+6th+session, refer to documents WLTP-06-27e &
WLTP-06-2e
18
Reference: “Intercomparison of real-time tailpipe ammonia measurements from vehicles tested over the
new Worldwide harmonized Light-duty vehicle Test Cycle (WLTC)”. Environmental Science and
Pollution Research, 7450-7460, 2015.
38
Summary of Validation phase results for ethanol, formaldehyde and acetaldehyde
Measurements were done either in real time or immediately after the test. The measurement
and analysis of exhaust emissions over the WLTC was done by means of Fourier transform
infrared spectrometer (FTIR), proton transfer reaction-mass spectrometry (PTR-Qi-ToF-MS),
photoacoustic spectroscopy (PAS) and gas chromatography (GC). The measured
concentrations and the emission profiles revealed that all the used instruments are suitable
to measure these compounds from the vehicle’s exhaust (|Z-score| < 2). Results showed that
online systems can perform measurements from the vehicle diluted exhaust assuring the
reproducibility and repeatability of the results19.
The achievements reached during phase 1b for measuring ethanol, formaldehyde and
acetaldehyde in the gas phase of LD vehicles’ exhaust are described below:
1. AP task force found new alternative on-line methods for ethanol, formaldehyde and
acetaldehyde in addition to the classical methods already known for carbonyls (DNPH
cartridges) and for ethanol (impingers). Both are considered reliable reference
methods but quite time consuming.
2. Conclusions reached during VP in Phase 1b showed the possibility of measuring 3
additional pollutants (ethanol, formaldehyde and acetaldehyde) directly at the CVS
(diluted exhaust).
3. All new methods have been validated and proposed as alternative methods to be
included in the GTR
GTR drafting
The text referring to ammonia in the last version of the GTR (phase 1a, Annex 5 par.7.1.1)
was modified according to the conclusions of the Validation Phase. The measurement
methods of EtOH, formaldehyde and acetaldehyde were added to the GTR in the respective
annexes: Annex 5 (Instrumentation and Methods) and Annex 7 (Calculations).
3.4.5.10 Mode selection and predominant mode

Background
A vehicle can be equipped with different operational modes which determine how the vehicle
responds to the driver. For instance there can be a normal mode, an eco-mode and a sport-
mode to choose from. The GTR has to specify which mode the vehicle should be tested in.
Secondly, one of these modes may be automatically selected when the vehicle is started,
and can be seen as a ‘predominant mode’. For the conventional ICE vehicles the mode
selection was already covered in phase 1a20, but for the electrified vehicles this was still
under discussion. That is why this was considered an open issue for phase 1b. The
Subgroup EV was tasked with this issue.
19
Slides and progress report available at: https://www2.unece.org/wiki/display/trans/WLTP+12th+session
Document WLTP-12-23e by Audi presents applicable factors if ethanol content in the fuel is below 25%
20
For more information about the mode selection for ICE vehicles refer to the Transmission section in
paragraph 0 of this report.
39
Phase 1a mode selection
According to the phase 1a version of the GTR, the mode selection for testing the different
classes of electrified vehicles was defined as follows:
1. OVC-HEV (Selection of a driver-selectable mode in for a charge-depleting Type 1 test):
“The charge-depleting test shall be performed by using the most electric energy
consuming mode that best matches the driving cycle. If the vehicle cannot follow the
trace, other installed propulsion systems shall be used to allow the vehicle to best follow
the cycle.”
2. OVC-HEV, NOVC-HEV and NOVC-FCHV (Selection of a driver-selectable mode for a

charge-sustaining Type 1 test):
“For vehicles equipped with a driver-selectable operating mode, the charge-sustaining test
shall be performed in the charging balance neutral hybrid mode that best matches the
target curve.”
3. PEV (Selection of a driver-selectable mode in for a charge-sustaining Type 1 test):

“If the vehicle is equipped with a driver-selectable operating mode, the charge-depleting
test shall be performed in the highest electric energy consumption mode that best
matches the speed trace.”
Decision for phase 1b

During phase 1b, this issue was intensely discussed within the Subgroup EV. Reason for the
discussion was on the one hand the imprecise description of the mode selection in the
GTR15 (state of play end of phase 1a) and on the other hand the desire to bring the EV
section in line with conventional vehicles concerning the mode selection in case of the
existence of a predominant mode.
PEVs and OVC-HEVs tested in charge-depleting operating conditions have to drive
consecutive cycles for the range and electric energy consumption determination up until the
break-off criterion has been reached. Depending on the REESS capacity this may take a
long time for testing. To avoid testing in multiple modes, it would make sense to apply the
predominant mode for this, i.e. the mode which is automatically selected if the vehicle is
switched on. However, the predominant mode might not always allow the vehicle to follow
the prescribed test cycle. Therefore, an important question which had to be answered was
the prioritisation of choosing the predominant mode versus a mode which enables the
vehicle to follow the driving curve of the applicable test cycle.
The Subgroup EV requested a clear political guidance from the WLTP-IG during the meeting
in Stockholm. The IG members decided that the following prioritisation should be observed:
1. First priority is being able to follow the applicable driving cycle
2. Second priority is choosing the predominant mode.
Based on this political guidance, the Subgroup EV developed a precise description for the
selection of the driver-selectable modes. This was done in the format of flowcharts with a
decision-tree for the following vehicles/conditions:
 OVC-HEVs under charge-depleting operating conditions
 OVC-HEVs, NOVC-HEVs and NOVC-FCHVs under charge-sustaining operating
conditions
 PEVs
40
The flowcharts included in Appendix 6 of Annex 8 of the GTR will clearly guide the
manufacturer and the responsible authority to select the appropriate mode for testing.
3.4.5.11 Other taskforces

Only those taskforces that resulted in a modification or addition to the GTR were listed in
Table 2 and have been described in this report. However, it has to be mentioned that there
have been more taskforces in place. Most of them were rather informal, with the purpose to
tackle a small issue e.g. on the formulation of a definition.
There is one taskforce that should be mentioned here: the coasting taskforce. Coasting is the
technology that decouples the engine from the transmission during decelerations. The
engine is then stopped, or returned to idle speed. This can save fuel, but the reduction
potential depends on how the technology is used by the driver. It was claimed by OICA that
the strict speed trace of the test cycle would prevent the full potential of the coasting system
being exploited. Therefore a taskforce was initiated by the IG to develop a methodology that
would result in a fuel consumption benefit that would be representative.
The first suggestion -a modification of the testcycle- was not acceptable for the Contracting
Party of Japan. The next proposal was to apply a mathematical approach calculates the fuel
reduction potential. This led to controversial discussions on how the ‘average’ driver would
adjust his driving behaviour, and to what extent the fuel reduction related to the change in
driving behaviour could be attributed to the coasting technology. Finally, it had to be
concluded by WLTP-IG that no agreement on coasting can be found. The issue might be
reopened in phase 2 of WLTP, but only if there is a new proposal that will be able to meet the
earlier expressed concerns.
41
4 Test procedure development
4.1 General Purpose and Requirements
Increasing evidence exists that the gap between the reported fuel consumption from type
approval tests and fuel consumption during real-world driving has increased over the years.
The main driver for this growing gap is linked to the flexibilities available in current test
procedures, as well as the introduction of fuel reduction technologies which show greater
benefits during the existing cycle than on the road. Both issues are best managed by a test
procedure representing the conditions encountered during real-world driving. As explained in
the introduction this was one of the main objectives to initiate the WLTP development
process, apart from harmonization. By bringing the test conditions and driving characteristics
of the test as close as possible to how vehicles are used in practice on the road, the
fuel/energy consumption and emission levels of test and reality are most likely to correspond.
The results from such a representative test would then implicitly serve as an objective and
comparable source of information to legislators and consumers.
At the same time, striving for the most representative test conditions might conflict with other
important test aspects. There are a number of constraints that need to be observed for the
development of the test procedure, such as:
a) Repeatability
If the test is repeated under the same conditions and in the same laboratory, the test
result should be similar (within a certain tolerance for accuracy). This means that e.g. all
conditions at the start of the test (such as the battery state-of-charge) should be well-
defined. If it is difficult to control or measure a vehicle parameter, it will be necessary to
fix the start condition at a worst- or best-case value while in representative driving
conditions this parameter may always be somewhere in between. Some of the
‘representativeness’ of the test is then sacrificed to obtain the goal of repeatability.
b) Reproducibility
If the test is repeated under the same conditions in a different laboratory, the test result
should be similar (within a certain tolerance for accuracy). If results from all labs over the
world have to be the comparable, this sets restrictions to the test conditions and the use
of cutting-edge measurement instruments. For instance, the test temperature level
cannot be chosen too low, since there are also many laboratories in areas with high
ambient temperatures.
c) Cost-efficiency
Covering all the effects that test conditions and driving characteristics have on the fuel
consumption and emissions may increase the complexity of the test or even require
additional testing. The costs of a higher test burden will eventually be charged to the
consumers, so there is a need to strike a balance between test effort and quality of the
results. Additional testing can only be justified if variations in conditions have a significant
effect on the result. Therefore, some of the ‘representativeness’ of the test is
compromised to reduce the test burden. For example, the length of the test cycle is only
30 minutes, which is a challenging timeframe to contain all of the world’s driving
characteristics.
d) Practicability
A test procedure needs to be executable in a practical way, without asking unrealistic
efforts from the testing personnel and/or the test equipment. That would be the case, for
instance, if tyres were required to be run-in at a test track by a test driver until they have
worn down to a certain tread depth. Normally, such requirements will also have issues
relating to the other constraints such as the cost-efficiency. There may also be practical
restrictions or safety restrictions to the test vehicle itself, e.g. monitoring the temperature
42
in the catalyst, or monitoring the battery state-of-charge with current transducer clamps in
the engine bay.
The general purpose for WLTP was therefore to primarily aim at a testing procedure that is
most representative for real-world conditions, but within the boundaries of it being
repeatable, reproducible, cost-effective and practicable. During the discussions in the
development process, this often led to conflicts in choosing which method to apply.
4.2 Approach
For the development of the test procedures, the DTP sub-group in phase 1a took first into
account existing emissions and energy consumption legislation, in particular those of the UN
ECE 1958 and 1998 Agreements, those of Japan and the US Environmental Protection
Agency Standard Part 1066. A detailed overview of the regional emission legislations that
were studied for the UN GTR is included in Appendix 3. These test procedures were critically
reviewed and compared to each other to find the best starting point for the draft text of the
UN GTR. The development process focused in particular on:
a) updated specifications for measurement equipment towards the current state-of-art in
measurement technology;
b) increased representativeness of the test and vehicle conditions, in order to achieve the
best guarantee for measuring a fuel/energy consumption that is similar as for average on-
road conditions;
c) ensure the capacity to deal with current and expected technical progress in vehicle and
engine technology in an appropriate and representative way. This particularly involves
the test procedure for electrified vehicles.
As such, the GTR text was updated and complemented by new elements where necessary.
For this technical report it would be too comprehensive to list all the modifications that were
introduced. General updating activities -such as bringing the accuracy requirements of the
instrumentation to the current state of the art- need no further clarification and fall outside of
the scope of this Technical Report. Instead, the important changes that have contributed the
most in achieving an improved and representative test procedure will be identified and
explained.
Paragraph 4.3 generally outlines the main improvements in the GTR. The modifications that
need some more clarification will be detailed in Paragraph 4.4.
4.3 Improvements in the GTR

As a result of extensive analyses and discussions among the involved stakeholders, the
WLTP GTR has managed to improve on many aspects of the existing emissions testing
procedures. These include:
a) The use of state-of-the-art measurement equipment with tightened tolerances and
calibration techniques to take advantage of advancements in measurement
technology (including additional pollutant emissions such as NO2, N2O, NH3, ethanol,
formaldehyde and acetaldehyde);
b) More stringent requirements imposed on the test vehicle and test track with the
intention to determine a representative road load;
c) New or improved procedures to measure emissions, electric range and
fuel/hydrogen/energy consumption of (hybrid) electric vehicles, as well as to
determine the effect of other future drive train technologies;
43
d) Improved methods to correct measurement results for parameters related to fuel
consumption and CO2 emissions (e.g. test temperature, vehicle mass, battery state of
charge).
On a more detailed level, the following list shows the main improvements on specific aspects
of the testing methodology which have contributed to increase the representativeness or
usefulness of the test results:
 Instead of declaring one CO2 value for an entire family of vehicles (as currently required
by EU legislation) each individual vehicle within a vehicle family will receive a CO2 value
based on its individual mass, rolling resistance and aerodynamic drag, as determined by
its standard and optional equipment. In WLTP, this first was called the ‘combined
approach’ and later renamed into the ‘interpolation method’. It considers the combined
CO2 influences of mass, rolling resistance and aerodynamic performance characteristics.
 The test-mass of the vehicle is raised to a more representative level, and is made
dependent on the actual carrying capacity of the vehicle by including a percentage of the
maximum vehicle load.
 Instead of using discrete inertia steps, the simulated inertia by the chassis dynamometer
corresponds exactly to the vehicle test mass.
 The battery state-of-charge at the start of the test is set to a representative yet repeatable
starting point. This is achieved by requiring a fully charged battery to be partially depleted
by first driving a WLTC as preconditioning cycle.
 The difference in battery state-of-charge over the cycle is monitored and the fuel
consumption is corrected according to the change in battery state-of-charge over the
cycle (upon exceeding a certain threshold).
 The soak and test temperature in the laboratory is modified from a range of 20 to 30°C
(as is currently prescribed in the NEDC procedure) to a setpoint of 23 °C. No systematic
deviation is allowed from this setpoint.
 Requirements and tolerances with respect to the road load determination procedure are
strengthened and improved:
o The test vehicle and tyre specifications must be similar to those of the vehicle that
will be produced;
o Test tyre preconditioning are more stringent (tread depth, tyre pressure, run-in,
shape, no heat treatment allowed, etc.) to more closely match the tyre conditions
on production vehicles;
o Use of on-board anemometry will be permitted, and the correction method applied
for wind during the coast-down method is improved (both for stationary wind
measurement as for on-board anemometry);
o Special brake preparation to avoid parasitic losses from brake pads touching the
brake discs will be prevented by a mandatory brake procedure prior to the test;
o Wheel alignment settings are specified (set to a worst-case setting or according to
the prescribed value for normal on-road use)
o Test track characteristics (e.g. road inclination) will be more stringent to reduce
influences on the road load determination.
 Instead of the ‘table of running resistances’ (the ‘cookbook’ of road load values that can
be used if the road load for a vehicle has not been determined by track tests), a formula
for calculating road load is provided, based on related vehicle characteristics.
 Additional road load determination methods are added, e.g. the torque-meter method, the
on-board anemometry method, the road load matrix family and the wind tunnel method.
44
 Wind tunnel criteria are added, both for the wind tunnel method as for the delta Cd.A
determination, including provisions to approve the wind tunnel.
 An interpolation method for the calculation of the road load within a ‘road load family’ is
included.
 A formula for the calculation of fuel consumption based on the CO2 and pollutant
emissions are added, including the interpolation of the fuel consumption.
 The GTR text is more robust on various testing details (e.g. the torque-meter method for
road load determination)
 Definitions in the GTR, e.g. on mass, reference speeds, etc. have been improved for
more clarity and to ensure unambiguous interpretation.
 Measurement procedures are added for additional pollutants, i.e. NO2, N2O, NH3,
ethanol, formaldehyde and acetaldehyde.
 Electric and hybrid vehicles are separated from conventional vehicles with only an
internal combustion engine, and dedicated test procedures have been developed for
these vehicle types. Range, fuel/hydrogen/energy consumption, and emissions of
(hybrid) electrified vehicles are defined in all-electric, charge-sustaining, and charge-
depleting mode, and weighted by utility factors (where applicable).
 For pure electric vehicles (PEV) and hybrid electric vehicles (HEV) the provisions for test
preparation and preconditioning as well as for the tests were modified with respect to
existing regulations on the following aspects:
o REESS preparation
o REESS charge balance correction
o Test procedure, separately for:
 OVC-HEV,
 NOVC-HEV,
 PEV,
 OVC-FCHV,
 NOVC-FCHV.
o Calculations of whole cycle and (where applicable) phase-specific results for:
 Emission compound calculations,
 CO2 and Fuel Consumption Calculations including an interpolation method,
 Electric Energy Consumption Calculations including an interpolation method,
 Electric Range including an interpolation method.
o Mode selection for driver-selectable modes
o Cycle-downscaling and capped speed provisions for PEVs
o A shortened test procedure for PEVs
 Test equipment and calibration procedures were improved and/or supplemented in order
to better reflect the technical progress and current state of the art, particularly on the
following items:
o Cooling fan specifications (increased dimensions, decreased tolerances of the
velocity of the air of the blower),
o Chassis dynamometer (provisions for 4WD were added, the general requirements
were aligned with US 1066),
45
o Exhaust gas dilution system (subsonic venturi (SSV) or an ultrasonic flow meter
(USM) were added),
o Emission measurement equipment (also for the additional pollutants),
o Calibration intervals and procedures (calibration and recheck before and after each
test instead of each bag analysis),
o Reference gases (tolerances reduced from 2% to 1%).
 WLTP post-processing procedures that specify the calculation order of the of the output
values.
4.4 New concepts of the GTR

The main improvements introduced by the GTR have been identified in the previous
paragraph. In some cases it was sufficient to tighten a tolerance, or add a simple
requirement. For other improvements it was necessary to develop a whole new approach,
leading to a new concept in the GTR. To give a more detailed explanation on the background
and the underlying principles, this paragraph will outline the main new concepts that were
introduced.
4.4.1 Interpolation method

One of the key objectives of WLTP, as specified in par. 4.2, is to develop the test cycle and
test procedure in such a way that the resulting CO2 emission and fuel consumption is
representative for real-life vehicle usage. One barrier to achieve that goal, which was
identified early in the development process, is the fact that tests are executed on single
vehicles while the results of these tests are used to type-approve a whole family of vehicles.
The vehicles in one family would mainly differ from each other in terms of options selected by
the customer that lead to differences in mass, tire/wheel rim combinations and vehicle body
trim and/or shape. It was considered valuable to find a method that would attribute CO2 to
individual vehicles within the family in an appropriate way.
First of all, it was recognised that testing only one vehicle does not provide sufficient
information. At least two different vehicles within the family have to be tested to determine a
difference in CO2 that can be attributed to vehicle characteristics: one vehicle to the ‘worst-
case’ side and preferably one to the ‘best-case’ side to allow good coverage of all vehicles in
the family. Within the GTR these test vehicles are referred to as vehicle H and vehicle L
respectively. It was also agreed that pollutant emission standards should be met by all
vehicles within the family.
The next challenge was to attribute the difference found in CO2 between vehicle H and L to
vehicles in between. There is not a parameter available that single-handed correlates well to
the increased CO2 as a result of differences in mass, aerodynamic drag and rolling
resistance. As a first candidate, the mass of the vehicle was proposed as a parameter for
interpolation between vehicle H and L. Analysis of such an interpolation method lead to
unacceptable errors. This is easily understandable by considering that some options only
add mass, while others (e.g. spoilers, wider tires) only have a marginal effect on mass but
add considerable aerodynamic drag and/or rolling resistance.
The final breakthrough in this discussion arrived with the insight that it is the energy needed
at the wheels to follow the cycle which has a nearly direct effect on the CO2 of the test
vehicle, under the assumption of a relatively constant engine efficiency for vehicle L and H.
The cycle energy is the sum of the energy to overcome the total resistance of the vehicle,
and the kinetic energy from acceleration:
46
Ecycle = Eresistance + Ekinetic
With:
Eresistance = road load force F(v) multiplied by distance.
Ekinetic = vehicle test mass TM multiplied by acceleration and distance
These energy components are summed for each second of the cycle to form the total cycle
energy demand. Please note that if Ecycle is negative, it is calculated as zero.
The total resistance force F(v) follows from the road load determination procedure, as
outlined in Annex 4, and is expressed as a second order polynomial with the vehicle speed:
F(v) = f0 + f1.v + f2.v2
With:
f0, f1 and f2 being the road load coefficients which are found by regression of the polynomial
to the road load determination results.
The key elements for success of this method are that:
a) the difference ∆CO2 between vehicle L and H correlates well to the difference in cycle
energy ∆Ecycle, and
b) differences in mass, rolling resistance and aerodynamic drag due to vehicle options
can be translated into independent effects on f0, f1 and f2 and consequently into
∆Ecycle.
This last statement can be assumed fulfilled by considering the following arguments:
 The kinetic energy responds linearly to the mass of the vehicle.
 f0 responds linearly to the tyre rolling resistance and the mass of the vehicle.
 f1 has nearly no correlation to the mass, rolling resistance and/or aerodynamic drag
and can be considered identical for vehicles L and H.
 f2 responds linear to the product of aerodynamic drag coefficient Cd and vehicle
frontal area Af.
Consequently, if the values for mass, rolling resistance and aerodynamic drag are known for
vehicles L, vehicle H and every individual vehicle of the interpolation family, the difference in
cycle energy ∆Ecycle can be calculated with respect to vehicle L, and from the interpolation
curve the ∆CO2 is derived . This so-called interpolation method is illustrated in the figure
below for an individual vehicle with a ∆Ecycle which is 40% of the difference in cycle energy
between vehicle L and H.
The general principle of this CO2 interpolation method is described in par. 1.2.3.1 of Annex 6.
The mathematical representation is found in the formulas of par. 3.2.2 and section 5 of
Annex 7. Please note that the method is applied for each cycle phase separately (Low,
Medium, High and Extra-High).
47
CO2
CO2 for an individual vehicle with
40% ∆E due to ∆m, ∆aero and ∆RR
(with respect to best-case vehicle)
∆CO2,cycle
∆Ecycle
TML TMIndividual TMH Cycle

RRL RRIndividual RRH Energy
AeroL AeroIndividual AeroH
Figure 10: Example of the CO2 interpolation method applied for road load relevant
vehicle characteristics.
4.4.2 Vehicle selection

In a first attempt to specify test vehicle H for the CO2 vehicle family, the vehicle with the
highest mass, the highest rolling resistance tyres and the highest aerodynamic drag was
proposed. This seemed a sensible approach to describe a worst-case vehicle until it was
recognised that the vehicle with the highest mass may not be fitted with the worst-case tyres
and vice versa. Specifying such a worst-case vehicle could then lead to a non-existing
vehicle. The definition for vehicle selection in par. 4.2.1 of Annex 4 was therefore chosen to
be described in a more functional way: “A test vehicle (vehicle H) with the combination of
road load relevant characteristics (i.e. mass, aerodynamic drag and tyre rolling resistance)
producing the highest energy demand shall be selected from the interpolation family.” If in
the example above the influence of tyre rolling resistance on the energy demand is higher
than that of the mass and aerodynamics, the vehicle with the worst-case tyres is selected as
vehicle H. Consequently, there are no specific requirements as to what the test mass,
aerodynamic drag and rolling resistance are for test vehicle H, since that is implicitly stated in
paragraph 4.2.1.1. The same approach is followed for the selection of the best-case test
vehicle L, but then of course aiming at the combination of road load relevant characteristics
producing the lowest energy demand.
48
4.4.3 Interpolation/extrapolation range
The accuracy of the interpolation method for CO2 has been validated by 2 vehicle
manufacturers using their detailed in-house simulation models. The CO2 and Ecycle for
vehicles L and H were determined, and used to interpolate the CO2 of vehicles in between.
Comparing the interpolation results with the simulation results for intermediate vehicles of the
family demonstrated that the interpolation method is accurate well within 1 g/km of CO2 up to
a ∆CO2 of more than 30 g/km21. On the basis of these results the methodology was accepted
and the allowed interpolation range was set to a maximum of 30 g/km or 20% of the CO2 for
vehicle H, whichever is the lower value. The latter was needed to prevent that low CO2
emitting vehicles would receive a relatively large interpolation range. Also a lower range limit
of 5 g/km between vehicle L and H was set to prevent that test-to-test measurement
inaccuracies have a large influence on the course of the interpolation line. Finally it was also
agreed that the interpolation line may be extrapolated to both ends by a maximum of 3 g/km,
e.g. to include future vehicle modifications within the same type approval. However, the
absolute interpolation range boundaries of 5 and 30 g/km may not be exceeded. This
interpolation range does not apply for vehicles which have been tested according to the road
load matrix family approach (refer to paragraph 5 of Annex 4), which need a wider rang. It is
assumed that the safety margin built in the calculation of the road load will implicitly limit the
interpolation range.
The allowed interpolation/extrapolation range is specified in 1.2.3.2 of Annex 6.
4.4.4 Vehicle test mass

The mass of the test vehicle in UN-ECE Regulation 83 was found to be lower than in real-life
conditions. It is based on the so-called mass in running order (MRO), which is the sum of the
mass of the empty vehicle, the standard equipment (including spare wheel), at least 90% of
the fuel tank filled, and a mass of 75 kg to represent the weight of the driver. Any additional
mass due to the optional equipment and/or the carrying of passengers and luggage is not
taken into account. This definition can be found in the Special Resolution on Consolidated
Resolution on the Construction of Vehicles (R.E.3)22
For WLTP it was decided that the test mass of the vehicle should also include a
representative share of these missing elements. Based on some elementary studies and
calculations, the agreed compromise was that the test mass (TM) would be determined by
the sum of the following mass contributions23:
a) The empty mass of the vehicle (to make use of the definition in the Special Resolution,
this is defined as the MRO minus 75 kg),
b) The mass of the driver (75 kg),
c) An additional constant mass of 25 kg, related to after-sales equipment and luggage,
d) A variable mass that depends on the carrying capacity of the vehicle (‘maximum vehicle
load’). Depending on their category and/or anticipated usage (decided at regional level)
the mass representative of the vehicle load will be 15 or 28% of the difference between
the technical permissible maximum laden mass and the sum of the mass contributions of
a) to c) and the mass of the optional equipment as defined in par. 3.2.8., and
21
See document WLTP-DTP-LabProc-238
22
See document ECE/TRANS/WP.29/78/Rev.2
http://www.unece.org/trans/main/wp29/wp29wgs/wp29gen/wp29classification.html
23
See document WLTP-DTP-08-02e
49
e) The mass of optional equipment (factory installed equipment that is selected by the
customer24).
The difference between the test mass of vehicle H (TMH) and vehicle L (TML) corresponds to
the mass difference due to the installed optional equipment on these vehicles.
The actual mass of the test vehicle is checked before the road load determination is started,
and needs to be equal or higher than the target test mass. During the test phase this mass
may change, e.g. due to the fuel consumed. After the procedure has been completed the
vehicle’s mass is measured again, and the average of these measurements will be used as
input for the calculations (TMH,actual respectively TML,actual).
The vehicle test mass is defined in 3.2.25 of part II and is referred to in paragraph 4.2.1.6 of
Annex 4. A graphical presentation of the mass definitions and how they relate to one another
to build the test mass is provided in paragraph 3.4.5.2 of this report (see the section on
Masses).
4.4.5 Vehicle coastdown mode and dynamometer operation mode

There are two special modes the vehicle can be equipped with, that are specifically
developed for the purpose of being able to test the vehicle:
a) Vehicle coastdown mode: This mode is needed when the road load determination
procedure uses the coastdown principle, while the verification criteria cannot be met due
to non-reproducible forces in the driveline (e.g. parasitic losses in electric engines used
for propulsion). By activating the vehicle coastdown mode, the driveline components that
generate these non-reproducible forces should be mechanically and/or electrically
decoupled. The vehicle coastdown mode has to be activated both during the road load
determination procedure as on the chassis dynamometer.
b) Vehicle dynamometer operation mode: This mode is used to be able to drive the vehicle
normally on a single-axis chassis dynamometer. If the vehicle is front wheel driven, the
rear wheels are not rotating during the test. This might trigger the electronic stability
program (ESP) system of the vehicle, which response would render the test result invalid.
The vehicle dynamometer mode is only used when the vehicle is tested on the chassis
dynamometer.
Both these special modes are not intended to be used by the customer and should therefore
be ‘hidden’. They could be activated by a special routine e.g. using vehicle steering wheel
buttons in a special sequence pressing order, using the manufacturer’s workshop tester, or
by removing a fuse. Both modes should not activate, modulate, delay or deactivate the
operation of any part that affects the emissions and fuel consumption under the test
conditions.
The requirements for vehicle coastdown mode can be found in paragraph 4.2.1.8.5 of Annex
4, and for the dynamometer operation mode in paragraph 1.2.4.2.2 of Annex 6.
24
Since manufacturers cannot be held responsible for what is fitted to the production vehicle after it has
left the production line, any items fitted by the car dealership and other after-sales equipment is not
included in the mass of the optional equipment. This should however not create an incentive for
manufacturers to shift the installation of vehicle options from the factory to the dealer. If this would
become a common practice for the future, appropriate measures should be taken to avoid this loophole.
50
4.4.6 Tyres
The rolling resistance coefficient (RRC) of a tyre has to be measured according to Regulation
No. 117-02, or a similar internationally-accepted equivalent, and aligned according to the
respective regional procedures (e.g. EU 1235/2011). The UN GTR also introduced a
classification scheme, identical to EU Tyre Labelling Regulation 1222/2009. There are two
reasons for having a classification table:
a) The rolling resistance coefficient determination procedure is complicated, and known
to have inaccuracies. By introducing classes with a range of RRC’s which all receive
the same class value, the inaccuracy of this determination procedure takes no effect.
b) Since the GTR has introduced the CO2 interpolation method, every individual vehicle
will receive its own CO2 value. During the production, manufacturers could switch
from one tyre supplier to another. If the other tyres have a slightly different RRC, a
situation could occur that two completely identical vehicles (except for the brand of
the tyres fitted) would receive a different CO2 rating value. With the classification this
situation is prevented, as long as the different tyres fall into the same class.
The influence of the class width on the CO2 emissions was investigated. The difference in
measured CO2 between the actual RRC and the RRC class value was found to be smaller
than 1.2 g/km per ton of vehicle mass25.
For the calculation procedure that establishes the ‘slope’ of the CO2 interpolation line, the
actual RRC values are used as an input, not the class values. At the point when the
individual CO2 values are calculated for vehicles in the family, the RRC class values are
used. See paragraph 4.4.24
The tyre selection and the accompanying classification table can be found in paragraph 4.4.2
of Annex 4.
4.4.7 On-board anemometry

The Annex 4 Task Force was asked by the IG to better understand the background of the on-
board anemometry method and its associated calculations. This should include –if
considered necessary- the development of applicable criteria which provide statistical
grounds for the validation of the resulting measurement data.
Task force discussions and in-depth bilateral reviews with on-board anemometry experts
concerning the method’s source material, SAE J2263, led to the joint proposal that was
developed during phase 1b and adopted at the 12th IG meeting. Extreme cases of the
method’s parameters were studied to evaluate sensitivity, and a few deviations from the SAE
method (and phase 1a GTR text) were introduced to enhance the method for WLTP
implementation. The main changes to are the following:
 The option for contracting parties to opt for increased wind tolerances was removed
from the GTR, as those wind tolerances were outside of the allowable winds in SAE
J2263, and the applicability of the method’s calculations were at risk under those
conditions.
 In addition, overall wind speed tolerances were reduced slightly in an effort to further
reduce potential test to test variation. The tighter tolerances fall within the guidelines
set by the SAE J2263 (DEC2008) standard, ensuring its continued applicability.
25
See document WLTP-DTP-LabProcICE-140
51
 Once calculations are complete and the data is corrected to standard conditions, the
resulting force equations must satisfy new convergence criteria.
Concerning the last point, it was determined that the statistical accuracy requirements of the
stationary method were not applicable to the on-board method, since the output of the
method is a quadratic force equation instead of the gated times from the stationary method.
As such, the evaluation of the resulting forces using this convergence check was developed
to ensure a level of statistical relevance within the dataset.
The method for measuring wind with on-board anemometry is included in Annex 4,
paragraph 4.3.
Following the method’s adoption during the 12th IG meeting there should not be any
outstanding items remaining for Phase 2.
4.4.8 Default road load factors

In case of small production series or if there are many variants in one vehicle family, it may
not be cost-effective to do all the necessary road-load determination work by measurements.
Instead, a manufacturer may elect to use a default road-load factors. In UNECE Regulation
83 a table with road load coefficients is included (‘table values’), which are only related to the
reference mass of the vehicle, regardless of the vehicle size. It was agreed to develop a new
proposal for this table, with the following improvements26:
a) The table should be based on existing road load data, and should be oriented
towards the "worst" case. More concrete, it should represent the 5% vehicles with the
highest running resistances, rather than an "average" figure, in order not to create an
incentive to apply the default values for vehicles that have a higher than average road
load.
b) The table should use vehicle parameters as input which have a relation to the road
load of the vehicles
c) The specified load parameters will be used as target coefficients for the chassis
dynamometer setting, in contrast to Regulation 83 where the table values are
intended as set coefficients for the dynamometer.
A detailed study and a statistical analysis was performed by TNO on a dataset of road-load
factors which led to a formula for the road load factors, rather than a table27. The formula is
based on the vehicle’s test mass, and the product of vehicle width and height as an indicator
for the size of the vehicle. The formulas for the determination of the default f0 and f2 road load
coefficients can be found in paragraph 5.2.2 of Annex 4.
4.4.9 Road load matrix family

The Road Load Matrix Family (RLMF) was developed as an additional road load
determination method to facilitate low-volume vehicles for which the test effort of measuring
a vehicle L and H is too high, but on the other hand the default road load values would be too
pessimistic. More specific, the foreseen vehicle types to make use of this method are –
amongst others- large vans and multi-stage vehicles. To target these types of vehicles, the
26
See document WLTP-DTP-13-05 https://www2.unece.org/wiki/display/trans/DTP+13th+Session
27
See document WLTP-DTP-14-07 https://www2.unece.org/wiki/display/trans/DTP+14th+Session
52
scope ffor applicatiion of the R
RLMF methhod was se
et to vehicle
es with a minimum
m tecchnically
permisssible maximum laden mass
m of 300
00 kg.
Rather than measu uring the rooad load of the two vehicles at th he extreme sides of the e family,
the RLMMF is based d on a sing gle measure ement of a representa ative vehicle
e of the fam
mily, and
‘extrapo
olating’ this by conside ering the diffferences of
o relevant road
r load parameters,
p i.e. test
mass (TTM), tyre ro olling resista
ance (RR) and frontal area (Af). Since
S the extrapolation
e n of one
measurrement to either sides is less acc curate as ann interpolation between the extre emes, an
addition
nal safety margin
m was built into the method byb using a base
b vehiclee with a wo
orst-case
aerodynnamic drag, and by using u conse ervative corrrelation factors for th
he influencee of the
parame eters on the road load coefficients.
c .
Much o of the devellopment wo ork focusedd on the co orrelation fa
actors. It waas clear that these
factors should be different
d forr upward annd downward correlatio on, and tha at the safetyy margin
had to bbe similar to
o either sidees. This approach shoould ensure that the furrther away fromf the
measurred vehicle,, the more likely it iss that the actual road d load is overestimat
o ted (and
consequ uently also o the resp pective CO O2). As a result, an incentive e is included for
manufacturers to apply
a one of the standa ard road loaad methods s. The requirement of a similar
safety mmargin to both
b ends should
s enco
ourage man nufacturers to select a test vehicle in the
mid-range of the fa amily. In thee case that the deviatiion from the
e actual roaad load wou uld bring
an unaccceptable COC 2 disadva antage to th
he manufactturer, he co ould choose e to split the
e vehicle
family, o
or use one of
o the otherr road load determinati
d on methodss.
The me ethod is illusstrated in Figure
F 11, and
a shows that the exttrapolated road
r load (red line)
for othe
er vehicles in
n the familyy would follo
ow the uppe
er area of th
he actual road load ban
ndwidth.
Figure 11: Upward

d and down
nward extrrapolation for
f the road
d load mattrix family
By extraapolation of
o the road load to veh hicle L and
d H, the tarrget road looads for me easuring
these vehicles at the
t chassiss dynamometer can be e found. Th he test vehicle is teste
ed at the
vehicle L and the vehicle
v H ro
oad loads, and
a the CO2 results are e used to drraw an interrpolation
line for CO2 again nst cycle energy. For any of the e other veh hicles in the
e RLMF, th he cycle
energy will be calcculated from m the extra
apolated road load, an nd the calcculated CO2 follows
from thee interpolation methodd. Note that the CO2 innterpolation line does not
n have the e kinked
shape of
o red line in
n Figure 11..
A detailed descripttion on the developmen
d nt of the RL
LMF method nterpolation method
d and the in
ented in App
is prese pendix 2.
The me
ethod of the RLMF itself is included
d in the GTR in chapte
er 5 of Anne
ex 4.
53
4.4.10 Torque meter method
Background
The torque meter method was included in the phase 1a version of the GTR. At the time it
was also acknowledged the method should be reviewed in phase 1b by the road load experts
and a validation should take place to provide justification on the equivalence to other road
load determination options.
Improvements in phase 1b
Apart from editing modifications to make the GTR text more robust on that part, the review of
led to the following improvements:
 Additional speed points incremental steps of 10 instead of 20 km/h, to allow a more
accurate least squares regression curve (valid for all road load determination options)
 Specification of wheel torque measurement accuracy for the whole vehicle defined.
 Determining the dynamic radius of the tire at 80 km/h and check this to limit the
difference between on-road and chassis dynamometer testing
 A wind correction factor is has been added. The road load curve is now corrected by a
wind compensation factor: w2 3.62 c2 v2w ; this was not included in ECE83.
 Compensation for speed drift ensures a more correct value of the torque measurement
result
 A procedure was added to convert the torque based running resistance curve into a
force-based road load curve on the chassis dynamometer (see paragraph 8.2.4 of Annex
4)
Validation
The torque-meter method was validated by Ford, and the road load curves were found to be
in good agreement with the coastdown test results.
The following steps were taken to prove equivalency results between coast down method
and torque method:
1. Vehicle ‘A’ was tested at Lommel Proving Ground by the coast down method and torque
method using exact the same tires, tire pressure and ride heights. Test results for both
methods used for further steps were selected based on testing in similar weather
circumstances.
2. Wind tunnel testing was performed to evaluate the aerodynamic difference (Cd.A)
between the vehicle with and without torque transducers.
3. A recalculation of the torque method road loads was performed towards the same
conditions as the coast down results, and to correct for weight and Cd.A differences
4. A dyno setting was performed to the road load curve that was recalculated in step 3.
5. A coast down on the dyno was performed to determine the coast down times.
6. The road load forces were determined from the coast down method and torque meter
method.
Note that the road load tests at Lommel Proving Ground, the aerodynamic tests, and the
dyno setting procedures were witnessed by TÜV.
54
Figure 12: Validattion results
s for the to
orque meter method
Table 5:
5 Validattion results orque meter method
s for the to
The ressults of this validation exercise

e e shown in Figure 12 and
are a Table 5. 5 The green dotted
line in F
Figure 12 shhows the toorque meterr method road load, whhile the red line is for th
he coast
down m method road load28.
The connclusion from this validation exerccise was tha
at for this ve
ehicle the diifferences between
b
the forces derived out of the coast
c down method and d torque me eter method d were gene erally
28
See document
d WL LTP-11-15e att https://www
w2.unece.org//wiki/display//trans/WLTP+
+11th+session for
more detailed validation resultss
55
within 4
4% at lower speeds, witth a declinin
ng tendencyy towards higher
h speedds. The torqque
meter mmethod showwed a consistently high her road loa
ad. The abssolute force differencess was 13
N for sp
peeds over 80
8 km/h, annd smaller forf the lower speed ran nge.
Finally, a procedurre was adde ed to transfform the runnning resisttance curve e determine ed by the
torque meter
m method into a ro oad load cuurve. This iss achieved by coasting g down a ve ehicle at
the chaassis dynam mometer wh hich was prreviously se et to reproduce the torrque-based running
resistan
nces. Of course this prrocedure ca an only be applied if the vehicle is capable to coast
down inn a repeatab ble way (i.e. no unrepre esentative parasitic
p lossses in the drivetrain). If that is
not thee case, the e road loa ad coefficieents are calculated
c from the running resistance
coefficie
ents, taking into accouunt the dynaamic radius of the whe eels and a default
d valuee for the
drivetrain losses of
o 2%. Thiss transforma ation proceedure is described in paragraph 8.2.4 of
Annex 4 4. The torqu
ue meter me ethod itself can be founnd in paragraph 4.4 of Annex 4.
4.4.11 Win
nd tunnell method
This me ethod deterrmines the road load by using a combinatio on of a wind tunnel an nd a flat
belt29 orr a chassis dynamome eter. Within the R&D aactivities of manufactur
m rers the winnd tunnel
has beccome a widely used in nstrument. Up
U until the GTR was developed it was onlyy defined
in some e standards. Within pha
ase 1b the method
m hass been workked out in detail, and va
alidation
measurrements hav ve been perrformed.
Motivattion
The neeed for the wind tunneel method was
w alreadyy expresse
ed during phase 1a off WLTP.
Weatheer conditions in most parts
p of the
e world maake coast down
d testing
g on the rooad only
possible
e on a limited amount of days. Fo
or examplee in Germanny on-road testing is limited to
roughly 100 days per
p year.
o foreseen that
It is also t optimizzing the fuel efficiency will increas
singly depennd on improoving the
aerodyn namic perfo
ormance of the vehicle e. To evalu uate aerody ynamic inno ovations pro
operly, a
measurrement method of high her accuraccy is neede ed, because e the influen
nce often searched
s
for mayy be in the order
o of the inaccuracyy of the coasst down me ethod. The wind
w tunnel method
allows for
f accuratee measurem ment of the physical ve ehicle drag in the absence of any external
influencces and no corrections
c and calculaations over the speed, time and mass.
m
29
This iis a circular belt
b that is driiven between two rollers. The T contact with
w the tire is on the flat partp of
the beelt which is suupported by an
a air bearingg, hence the name.
n This is similar to an
n on-road situation,
whilee the contact surface
s on a chassis
c dynam mometer is innfluenced by the
t radius of the roller. Eaach wheel
can hhave its own flat
f belt, and the
t entire flatt belt unit is supported
s by an air bearing g to measure the
reactiion forces whhile the vehiclle is restraineed. The sum of
o reaction forrces is the tottal rolling ressistance.
56
Other advantages of the wind tunnel method are:
 measurements can also be done at a higher rate than the on-road alternatives;
 the repeatability is much higher;
 no atmospheric influences like wind, sun, humidity, etc.
 less corrections that compromise the accuracy (mass, temperature, air density, wind,
measurement equipment, etc.);
 no influences related to the driver, the test track or traffic;
For these reasons the wind tunnel method was welcomed as a good alternative road load
determination method. The only problem was the lack of a robust measurement procedure
and appropriate wind tunnel criteria, apart from some available standards.
Description
The basic idea of this method is that the aerodynamic drag and rolling resistance of the
vehicle can be separately determined. The wind tunnel is used to measure aerodynamic
drag, expressed as the aerodynamic resistance coefficient multiplied by the frontal area:
Cd*Af. The combination of rolling resistance and the losses of the drivetrain (e.g. wheel
bearings) is measured separately on a flat belt or on a chassis dynamometer. The sum of
these two resistance components form the total road load as it would be measured on the
road.
There are several options within the procedure, such as the coast down procedure (as on
road) or a stepwise constant speed approach (as typically performed in today's development)
for the rolling resistance determination. The advantage of the stepwise (or stabilized)
approach is to not have any influence of rotational or inertia masses. The advantage of the
coast down procedure is to be closer to on road testing and to the chassis dyno setting, so if
there were any unknown dynamic effect it would have the same influence during the coast-
down on the road and the coastdown on the dyno, thereby levelling out.
Another option is the warm-up procedure. The vehicle can be warmed up by driving the
vehicle, quite similar to the on-road warm-up. There was also an option included to drag the
vehicle by the dynamometer. This would eliminate the monotonous work and effort for the
driver of the vehicle. Due to the significant lower power transferred through the drivetrain
when the vehicle is dragged by the dynamometer, a higher warm-up speed is applied for this
option in order to arrive at a similar warm-up of the vehicle that is warmed-up under its own
power by a driver.
Any of these alternatives have to be confirmed and approved via a comparison to on road
testing before they may be used. As there is no direct link to on-road testing, it was agreed to
add a validation procedure. Every two years a correlation program has to be performed on
similar vehicles as intended to be type-approved. The road load of these vehicles will be
determined on the road and within the facilities (wind tunnel, flat belt/chassis dynamometer),
and the equivalency between the results has to be demonstrated. On average, the cycle
energy calculated from the road load may not deviate between these methods by more than
5% for a single vehicle, and more than 2% as an average of 3 vehicles.
Testing the rolling resistance on the chassis dynamometer requires an additional correction,
as due to the radius of the roller the rolling resistance of a tyre on the dyno is higher
compared to driving on a flat surface. A general correction formula is already available
(based on an old ISO standard), but was found to not be accurate for every tyre. The data of
an additional measurement series and the validation data produced by UTAC was used to
develop a conservative default formula, for the GTR. There is also a possibility included to
develop a more accurate formula, in close cooperation with the approval authority.
57
Validati
tion and jus
stification
To asse ess the validity of the e method and a the inccreased acccuracy, a large measurement
programm was perfo ormed by UTAC.
U The ey applied the
t combinnation of a wind tunne el and a
chassis dynamometer. After the measu urements were
w conclu
uded the saame vehicle es were
transferrred to VW tot assess th he validity of
o the methood using the
e combinatiion of a win
nd tunnel
and a flaat belt. The
e validation program inccluded:
 6 cars
 4 tracks
 2 wind tunn nels
 2 roller chassis dynam
mometer (two
o methods: deceleratio
ons and stabilized spee
eds)
 1 flat belt dyynamomete
er
The fina
al results sh
howed a goo
od quality of
o the test exxecution, an
nd the concclusions werre:
 There is a high
T h variatio
on of on-roa
ad results (eespecially due to the diifferent test tracks
t
that have been used);
 T repeata
The ability of the
e wind tunnel results iss very good;;
 T
There is a small
s system
matic deviation betwee en coast dowwn and winnd tunnel me ethod,
mostly the wind
w tunnel method yieelded a lowe er road load
d result.
Apart frrom one ve ehicle (N1-vvehicle) the systematicc deviation was smalle er than 10N
N. Some
results o
of the valida
ation progra
am are show wn in Table
e 6, Table 7,, Figure 13 and Figure 14.
The ove erall differen
nce in cycle
e energy demand betw ween the win
nd tunnel method (with chassis
dynamo ometer) and d the coastdown method was -0 0.8%, within a range of -2.0% to t 1.0%.
Includinng outliers the range iss between -4.7%
- and +2.2%.
+ The
ese variations are in th
he same
order o of magnitud de of the differences
d found betw
ween the coastdown
c measurem
ments on
different test trackss. As a concclusion, the
e wind tunneel method -e
either with the
t chassis dyno or
the flat belt- was coonsidered an
a acceptab ble road load
d determinaation methood.30
Table 6:
6 Cycle energy
e dem
mand comp
parison of wind tunne
el and coas
st down me
ethod
30
See foor more inforrmation documments WLTP P-10-14e and WLTP-10-155e at
https:://www2.unecce.org/wiki/ddisplay/trans/W
WLTP+10th+ +session
58
Table 7: Absoluute and rela ative rangee between the averagges of measurements s done
at each
h test faciliity in N and
d as a perc
centage (=b
bias of the methods)
Figure 13: Compaarison of ro

oad load curves
c for tthe coastdown metho od on the tracks
t
and th
he windtunnnel + chasssis dyno method (d deceleratio
on and steepwise
methodd). Dashed
d lines show
w confiden
nce intervals for the tracks
59
Figure 14: Compa arison of ro
oad load curves
c for tthe coastdown metho
od on the tracks
t
and the
e wind tunnel + flat belt
b method d
The validity of the alternative warm-up was

w measurred as well. Figure 15 shows the effect of
different (dyno drriven) warm m up strate
egies on the t mechanical drag of the ve
ehicle in
comparrison to a waarm-up by driving
d the vehicle
v itsellf.
Figure 15: Evalua

ation of the
e effect fo s on mechanical
or differentt warm-up strategies
drag
The form
mula for the n of the rolling resistance force du
e correction ue to the roller radius was
w also
evaluate
ed. The ressults are sho
own in Figu ure 16. The correction formula sug ggested by the ISO
proceduure proved to be incorrect (purple e dashed lin ne). Lowerinng the coeffficient of 1.0 in the
formula to 0.2 ressulted in a good matcch to the fo orce measu ured on the e flatbelt. This
T was
d in the GTR
included R as a defa
ault coefficie
ent for the ccorrection.
60
Figure 16: Evalua
ation of the
e correctio
on formula
a for the ro
oller radius
s of the ch
hassis
dynam
mometer
pment process
Develop
The winnd tunnel method
m was already described in some
s existing standard ds. Howeveer, these
standarrds rely on good engin neering judg
gement and d can there efore not be e applied as
a a test
proceduure for the GTR
G as succh. This me
eant that the
e text of the
e existing standards
s had to be
completted in a much high her level ofo detail, e.g.
e by se etting criterria and sppecifying
requirem
ments. This s also included the neeed for a coorrelation program
p bettween on-ro oad and
wind tunnnel tests.
During the process, the GTR R text was developed within a small sub-grroup and in
n a very
construcctive mann
ner to creaate a robusst test proccedure with
h guidance how to appply the
method and to perfform the tessting.
The nee ed for the inclusion off the wind tunnel
t methhod was ac cknowledgedd during ph
hase 1a,
but in th
he absence e of sufficien
nt validation
n data it waas decided to postpone e the adopttion until
phase 1b.
1 This gavve an opporrtunity to se et up a valid
dation test program,
p an
nd use the results to
developp the text for the GTR.
Within the
t taskforcce on the wind
w tunnel method dooubts were raised tow
wards the va
alidity of
wind tun
nnel resultss, especiallyy from Japa
anese side. As Japan has no concept of round robin
comparrison of win nd tunnel results like e in Europee, the con
ncerns were e well und
derstood.
Therefo
ore the follow
wing precau utions have
e been taken
n:
 wind tunnell criteria havve been scrrutinised an
w nd tightened
d where pos ssible,
 t approva
the al of the faccilities via a correlation with on roa
ad testing was
w added, and
a
 t
two validation studies were
w execu uted (by UTA AC and by VW).
Having delivered the

t required validation n data and a robust description
d of the metthod, IG
agreed on adoptin ng the win nd tunnel method
m e 10th IG m
witth the flat belt at the meeting.
Addition
nal testing was
w neede ed for the correction
c fu
unction on the chassiss dynamommeter, so
that parrt was adopted later at the 12th IG meeting.
The win
nd tunnel me
ethod is inccluded in pa
aragraph 6 of
o Annex 4.
61
Windtunnel criteria
It should be mentioned here that a windtunnel can be used for two purposes in the GTR:
a) to determine the ‘delta Cd.A’ between options to the vehicle exterior and/or
bodyshapes for the purpose of interpolation between vehicle L and H, and
b) to determine the overall Cd.A of the whole vehicle to derive the target road load
coefficients, i.e. the windtunnel method described in this paragraph.
The basic windtunnel criteria are laid down in paragraph 3.2 of Annex 4, but due to the
differences between these purposes the criteria for the wind tunnel method are more
stringent (see paragraph 6.4.1).
The reasons for these different criteria are as follows:
1. The difference between the delta Cd.A of vehicle L and H is much smaller than the overall
Cd.A of the whole vehicle. Therefore the absolute effect of an error in the determination of
the delta Cd.A has less consequences.
2. The sum of the delta Cd.A for the set of options on vehicle H is aligned by the Cd.A
difference between vehicle L and H. This means that any error in the measurement is
largely compensated.
For these reasons a bigger solid blockage ratio can be accepted for the wind tunnel used for
the delta Cd.A determination, and a higher deviation is allowed between front and rear
pressure coefficient. Also the blockage due to the vehicle restraint system has no influence,
because its influence levels out during the delta Cd.A determination.
4.4.12 Alternative delta Cd.A determination

For the interpolation method on CO2 as described in paragraph 4.4.1 there is a need to
determine the variation in the value of Cd.A for each vehicle option that has an influence on
the aerodynamic performance of the vehicle. In the GTR this is referred to as the delta Cd.A
determination, which is an input for the calculation of the cycle energy for an individual
vehicle. Examples of vehicle options which’ aerodynamic resistance would have to be
determined are wheel rims and tires, spoilers, adjustable vehicle height system, grille
shutters, etc.
It was acknowledged by the Annex 4 taskforce that:
1. Variations in the delta Cd.A or vehicle options are in the same order of magnitude as the
measurement tolerance. This makes it virtually impossible to determine an accurate
value for the delta Cd.A by performing e.g. a coast-down with and without the option
installed on the vehicle. Only the wind tunnel method may be sufficiently accurate to
measure this due to the absence of uncontrollable influences.
2. The determination of the delta Cd.A for all the options in a vehicle family may take a lot of
effort in the windtunnel, and is therefore time consuming and costly. At the same time,
not all manufacturers may have unlimited access to a windtunnel.
3. There are simulation methods available which are able to accurately determine the
influence on aerodynamic performance for different body styles and options installed at
the vehicle exterior.
For this reason an alternative method was proposed which –under strict requirements- would
allow the calculation of the delta Cd.A by e.g. computer simulations based on the method of
Computational Fluid Dynamics (CFD). The basic principle for this alternative method is that it
should always be validated by demonstrating equivalency with measured aerodynamic
results. Therefore, the following requirements and restrictions were set to this method:
a) The method may only be used after agreement by the responsible authority, and after
fulfilling the other requirements and restrictions.
62
b) It has to be demonstrated that the method has an accuracy of ±0.015 m2 delta Cd.A.
c) The method has to be validated, not only by demonstrating the accuracy requirement, but
also to yield similar flow patterns, air velocities, pressures and forces.
d) It can only be used for those kind of aerodynamic influencing parts (e.g. wheels, body
shapes, cooling system) for which equivalency was demonstrated.
e) The evidence of equivalency is presented in advance to the responsible authority, for
each road load family (if a simulation method is used) or by a correlation test programme
(if a measurement method is used).
f) Only the wind tunnel method is allowed to be used for the equivalency demonstration.
g) The method may not be applied for vehicle options with a delta Cd.A that is more than
100% higher than the option for which equivalency was demonstrated.
h) Whenever the simulation model is changed or updated, the validation needs to be re-
demonstrated.
Note that the alternative delta Cd.A method may only be used to determine the difference in
aerodynamic drag, it is not allowed to evaluate the absolute aerodynamic resistance of the
whole vehicle. For the measurement of the overall aerodynamic resistance e.g. the wind
tunnel method of paragraph 4.4.11 should be applied.
The alternative delta Cd.A method is described in paragraph 3.2.3.2.2.3. of Annex 7.
4.4.13 Road load family

The "Road Load Family" is a concept which allows to calculate road load coefficients instead
of measuring them. Within that framework, the interpolation is limited to a vehicle family with
similar characteristics but is independent for example of the vehicle's engine. Hence, a diesel
and a gasoline variant of the same vehicle model may be in the same "Road Load Family".
The method is based on a linear interpolation principle of the relevant road load properties:
aerodynamics, rolling resistance and mass. The effect of these properties is calculated into a
cycle energy value, quite similar to the approach for road load and CO2 calculation within the
‘Interpolation family’.
Motivation
The consequence of bringing in the concept of the interpolation family leads to an increase in
the test effort for road load determination because for every Interpolation Family at least two
vehicles ("High" and "Low") have to be tested. At the same time, the interpolation family
approach offers the use of a road load interpolation method based on relevant parameters.
This gives an opportunity to create a road load family that is larger than the interpolation
family, mainly by attributing the effect of the engine by means of a difference in vehicle mass
and –if appropriate- aerodynamic drag difference.
Scope
The following family criteria are specified in the GTR:
 same drivetrain and gearbox;
 limits to n/v ratio 25% (with respect to the most common installed transmission type);
 limits to interpolation range min. 4%, max. 35% cycle energy (based on vehicle HR);
 some additional provisions for electrified vehicles.
63
This me eans that different
d eng
gines (diesel, gasoline
e, different displaceme ents) can be
b in the
same R Road Load Family, bu ut different types of drivetrains
d (
(e.g. two-w
wheel or fou
ur-wheel
drive) or gearboxess (MT/AT) will
w be in diffferent Road
d Load Fam milies.
These ffamily criteria are described in parr. 5.7 of partt II of the GTR.
Validati
tion and jus
stification
Within tthe conceptt of the Inte
erpolation method
m (see
e par. 4.4.1 port) it was already
1 of this rep
confirme ed that road
d load and CO
C 2 have a linear respponse to diffferences in aerodynam mic drag,
rolling re
esistance and
a mass.
Differen
nt engines have
h no direct influencce on road load, apart from the parameters
p that can
be interrpolated (aeerodynamiccs, mass). This
T is valid
d for all pow
wertrains where
w the engine
e is
decoupled from th he drivetrain during ro oad load determinatio on. Therefore, as long g as the
drivetrain -starting at the cluttch and en nding at the e wheels- iss the same e, the road load of
different vehicles within
w that family can be calcula ated by inte erpolating the three ro
oad load
relevantt parameterrs, i.e. aeroddynamic draag, mass and rolling re esistance. See
S Figure 17.
Figure 17: Road load releva

ant compon
nents of the
e drivetrain
n for an ICE
E vehicle
Apart frrom this tecchnical argu

umentation, a validationn by testing g was considered nece
essary in
order to
o verify the linearity of this approa
ach and to establish
e a maximum range.
r The IG gave
the man ndate to BMMW to perform some ro oad load tessts for this purpose.
p
Due to the restrictiions of vehicle availab bility and weeather condditions only four vehiclles were
tested. Two vehiclles were se elected to represent a vehicle High H and Lo ow of a rannge that
typicallyy would en ncompass a road load vehicle family. f Thee other two o were sele ected in
between n vehicle L and H. The e first two fo
ormed the b basis for the
e interpolatio
on, based on
o which
the road d load for th
he last two vehicles co ould be calcculated. By comparing the measu ured and
calculatted road loa ads, the acccuracy of the road loa ad interpola ed.31 The
ation could be validate
vehicle selection is shown in n Figure 18 8 and the results
r are presented in Figure 19. The
vehicless were all rear-wheel driven, equipped with h the same e automatic c transmission, and
their n/vv ratio was within
w 11%..
31
It shoould be acknoowledged thatt any such diffference may also be attribbuted to the innaccuracy of the
coastddown method d itself. To eliiminate that influence
i as much
m as possiible, the coasstdown proceddure was
done twice, and thhe results werre averaged.
64
Figure 18: Vehicle
es selected
d for road load family
y validation
n tests
Figure 19: Results of the roa

ad load fam
mily validattion tests
For this particularlyy wide range of vehicle

es the validaation results
s show a ve ery good agreement
e calculated
with the d interpolatio
on line. Gennerally, the accuracy is o 0.5% of the cycle
s within 0 to
energy, with a ma aximum abssolute errorr of 0.08 MJ.M Over th he WLTP te est cycle th
his error
would ccorrespond to a diffe erence in COC 2 of app proximatelyy 0.5 g/km.. Therefore e it was
concludded that the approach of o the road load family was validatted, having an accuraccy that is
at least equal to thee coastdowwn method.
Develop
pment process
As the formulas were
w alread
dy available from the e Interpolattion Familyy, the deve
elopment
mainly ffocused on the family criteria, the
e maximum m range andd a robust drafting
d texxt for the
GTR. AAlso the description of the
t test veh hicles "High" and "Low"" was reworked and immproved,
65
in order to have a robust definition and a clear basis for the interpolation. The proposed
range by BMW of 4 to 35% of the cycle energy for vehicle H was considered acceptable.
The method was finally adopted at the 10th IG meeting32. It was accepted as a method which
significantly reduces testing effort without changing the accuracy of the results and is
therefore a clear improvement of the emission legislation, compared to the ones existing in
today's legislation worldwide.
The road load family is described in par. 4.2.1.3 of Annex 4.
4.4.14 Manufacturer’s responsibility on road load

The concept of ‘manufacturer responsibility’ on road load is also a new concept to the GTR,
not so much being a measurement or calculation concept but more like a principle. This
statement in paragraph 3 of Annex 4 needs to ensure that despite the variety of road load
measurement methods provided in the GTR and the tolerances allowed within these
methods, the road load reported for an individual vehicle should be confirmed and not
underestimated.
The gtr contains different methods to determine the road load of a vehicle, based on different
measurement options and calculation options:
 Coast down with stationary anemometer
 Coast down with on-board anemometer
 Torque meter method
 Wind tunnel with flat belt
 Wind tunnel with chassis dynamometer
 Road load family
 Road load matrix family
 Default road load parameters
Even though the measurement methods are developed to arrive at an accurate road load by
setting appropriate tolerances, accuracies and precisions, the road load values of a vehicle
may depend on the (combination of) method(s) and calculation(s) chosen. This choice of
method is up to the manufacturer. A selection of methods with the intention to determine
road load values that underscore the real world road load of production cars should be
avoided. Therefore the following text was included in par. 3 of Annex 4:
“The manufacturer shall be responsible for the accuracy of the road load coefficients
and will ensure this for each production vehicle within the road load family.
Tolerances within the allowed road load determination, simulation and calculation
methods shall not be used to underestimate the road load of production vehicles. At
the request of the responsible authority, the accuracy of the road load coefficients of
an individual vehicle shall be demonstrated.”
This statement basically ensures that if the road load of a production vehicle was verified by
the responsible authority, its road load would have to be in agreement with what was
declared at type approval.
32
See document WLTP-10-17-rev1e at https://www2.unece.org/wiki/display/trans/WLTP+10th+session
66
Since neither Conformity of Production or In-Service Conformity requirements are included in
this version of the GTR, the proposed wording was selected with care. It was not able to
agree on a reference road load determination method, and this issue should be further
discussed in Phase 2 of WLTP.
4.4.15 Alternative vehicle warm-up procedure

The WLTC based warm-up procedure takes 30 min time and adds 23 km on the odometer
provisions. To reduce this effort it was decided that there was a need for an alternative warm
up procedure, but this would only be accepted if could be demonstrated that it would yield at
least a similar warm-up of the vehicle. The alternative warm-up procedure would only be
valid for vehicles within the same road load family.
To demonstrate equivalent warm-up at least one vehicle representing the road load family
has to be selected and warmed up on the dynamometer according to the alternative
procedure. After this warm-up the dynamometer load setting is determined. The alternative
warm-up procedure is considered valid if the calculated cycle energy demand within each
cycle phase is equal to or higher than the energy of the same phase driven with
dynamometer load settings according to a warm up with a WLTC. The details of the
procedure and its equivalency have to be reported to the responsible authority.
4.4.16 REESS charge balance (RCB) correction for ICE vehicles

Under Regulation 83, the vehicle battery is normally fully charged at the start of the test. The
state of charge upon completion of the test will always be lower than 100%, which means
that effectively the energy drawn from the battery has been consumed over the test cycle.
Or, more scientifically correct, the engine did not have to restore the charging energy though
providing mechanical energy to the alternator.
Early in the WLTP process, this was recognized as an issue which has an unrealistic effect
on the fuel consumption at type approval, and whose influence is too high to be ignored33.
As a first step towards a representative test procedure, the battery state-of-charge at the
start of the test was changed from fully charged (NEDC) to a representative start value. This
is achieved by driving a preconditioning WLTC with a fully charged battery at the beginning.
Secondly, a pragmatic approach was developed to monitor and correct a significant
difference in battery charge over the cycle. The idea is to correct the fuel consumption and
CO2 emissions towards a zero charge balance, i.e. no net energy drawn from or supplied to
the battery. Please note that the term used for battery in the GTR is ‘REESS’ –
Rechargeable Electric Energy Storage System, and the ‘REESS Charge Balance is
abbreviated to RCB. The difference in energy level of the battery over the cycle is expressed
as ∆EREESS.
During the test, the battery current is monitored by a clamp-on or closed type current
transducer. This signal is integrated over the whole duration of the cycle to deliver the RCB.
If this RCB is negative (charge is reduced) and exceeds a specified threshold, the fuel
consumption will be corrected. This threshold is laid down in the RCB correction criteria table
A6.App2/2, and is based on the ∆EREESS divided by the equivalent energy of the consumed
33
See the report by Helge Schmidt and Ralf Johannsen: Future Development of the EU Directive for
Measuring the CO2 Emissions of Passenger Cars - Investigation of the Influence of Different Parameters
and the Improvement of Measurement Accuracy” - Final Report, 14 December 2010 (listed as document
WLTP-DTP-LabProcICE-038)
67
fuel. In the case that it is below the specified criteria (0.5% for the complete WLTC cycle
including the Extra-High phase), no correction needs to be applied.
The correction of the CO2 will be applied for every cycle phase independently (Low, Medium,
High and Extra-High). It is calculated by considering the ∆EREESS per cycle phase, an
assumed alternator efficiency of 0.67, and the combustion process specific Willans factor.
The Willans factors are expressing the engine’s efficiency in terms of the positive work of the
engine against the CO2. Under the driving conditions of the WLTC, the Willans factors will
remain relatively constant for small variations in cycle or load, and therefore provide a good
basis for correction. The corrected fuel consumption is expected to correspond to a WLTC
with zero charge balance.
The correction method for the RCB is outlined in Appendix 2 of Annex 6. The procedure for
the REESS charge balance correction of electrified vehicles is described in paragraph
4.4.18.
4.4.17 Electrified Vehicles

In the GTR a separate annex is dedicated to electrified vehicles (Annex 8). The electrified
vehicles are separated into the following groups according to their propulsion concepts:
 Pure electric vehicles (PEV)
 Hybrid electric vehicles, further subdivided into:
o Not off-vehicle charging hybrid electric vehicles (NOVC-HEV),
o Off vehicle charging hybrid electric vehicles (OVC-HEV)
Since it was not possible to determine appropriate parameters for the calculation of a rated
power value, the electrified vehicles could not be classified according to the method applied
to ICE vehicles. Instead, all Annex 8 vehicles are classified as Class 3 vehicles and therefore
the WLTC Class 3a or 3b driving curve is the reference cycle (depending on their maximum
speed). Consequently, different specifications for the cycle versions and the provisions for
vehicles that cannot follow the trace had to be elaborated. A ‘system power’ definition for
electrified vehicles is currently being developed by the EVE Informal Working Group. If that
work is completed in due time, this definition will be used for classification of electrified
vehicles into Class 1, 2 and 3 during WLTP phase 2.
The test procedure for monitoring the electric power supply system, defining the specific
provisions regarding the correction of test results for fuel consumption (l/100 km) and CO2
emissions (g/km) as a function of the energy balance ∆EREESS for the vehicle batteries, is
different from that for ICE vehicles (REESS = Rechargeable Electric Energy Storage
System). This procedure is referred to as the REESS charge balance (RCB) correction
method. All installed REESS’s are considered for the RCB correction of CO2 and fuel
consumption values. The sum of ∆EREESS is the sum of each REESS’s RCB, multiplied by the
respective nominal voltage.
New range tests for OVC-HEVs and PEVs are specified. Vehicles with manual transmission
are driven according to the manufacturer’s instructions, as incorporated in the manufacturer's
handbook of production vehicles and indicated by a technical gear shift instrument.
The vehicles are tested by the applicable WLTC and WLTC city phases (low and medium
only) in both charge-sustaining and in charge-depleting mode. This means that electrical
range as well as fuel consumption and CO2 emissions are determined for the whole cycle
and the low and medium speed phase cycle separately. Via the Utility Factor (UF), which is
dependent on the electric range in charge-depleting mode, the CO2 emissions and fuel
consumption results of the CS and CD test are transformed into a weighted average.
68
For the electric range determination of OVC-HEVs and PEVs the GTR contains completely
new requirements with respect to existing regulations. The break-off criteria for the electric
range tests were modified on the basis of the results from the validation 2 phase of the
WLTP development.
For NOVC-HEV with and without driver-selectable operating modes the RCB correction for
CO2 and fuel consumption measurement values are required. The RCB correction is not
required for the determination of emissions compounds.
4.4.18 RCB correction for OVC-HEVs, NOVC-HEVs and NOVC-FCHVs

The RCB correction for hybrid electrical vehicles which are tested according to Annex 8 have
a different correction procedure as used for conventional vehicles because they have more
than one battery while the energy content of the traction battery is much higher.
Background
The RCB correction for hybrids was already developed within phase 1a but a clear demand
was identified to further discuss this during the WLTP phase 1b. This decision was taken in
order to improve on the procedure, make it more robust and to be able to perform a deeper
analysis of the discussed approaches. This was considered essential as the determined
correction coefficient is not only required for the correction of whole cycle test results but also
for the determination of the phase specific values – see paragraph 4.4.20.
Phase specific values can also be determined by correcting each phase with a phase
specific correction coefficient. But due to the vehicle operation strategy it is not always
possible to determine in each and every phase a positive and negative charging balance,
which is a prerequisite for the correction coefficient determination.
In phase 1a, only a procedure under cold conditions was developed, which means that the
vehicle is starting in ambient temperature conditions at each correction coefficient
determination test. Ambient temperature conditions can be reached by soaking the vehicle
as defined in the GTR for a time period of 12-36 hours. This procedure was already applied
in the past but has proven to be very time consuming due to the long soak period in between
the tests. Therefore a more practical solution would be welcomed.
The main questions to be answered were defined as follows:
a) Under which conditions does an REESS energy change-based correction of the charge-
sustaining fuel consumption and CO2 mass emission have to be applied
b) How should the procedure for the correction coefficient determination be properly
defined?
c) Which boundary conditions for the correction coefficient determination tests should be
defined?
These questions were addressed by the Subgroup EV in phase 1b.
Application criteria for the RCB correction

The conclusion of the discussions within Subgroup EV level was that a correction is only
required if the REESS has been discharged and the correction criterion ‘c’ between the
69
e value of the REESS
absolute S electric energy
e cha
ange and th
he fuel ene
ergy is high
her than
0.5%34.
her cases a correction may be om
In all oth mitted and the uncorreccted valuess may be ussed. This
is graphhically illustrrated in Figure 20.
Figure 20: Graphiical illustra

ation of the
e applicatio
on criteria for
f RCB co
orrection
Figure 220 is only re
eferring to whole
w cycle
e test resultss. Individua
al phases ha
ave to be corrected
c
irrespecctive of the energy change, at le east if these
e values arre required by the Con ntracting
Party.
Proced
dures for th
he correctio
on coefficie
ent determ
mination
During p
phase 1b, Subgroup
S E experts discussed
EV d in
ntensively a new appro oach with re
espect to
the procedure for the determ mination of the correcttion coefficcient. This new
n approaach is a
determination proccedure unde er warm con
nditions, whhich can be e selected by
b the manu ufacturer
as an alternative option to the procedure under cold conditions..
The corrrection pro
ocedure und der warm co onditions was
w reviewed and evalu uated during phase
1b by tthe membe ers of the WLTP Sub bgroup EV.. For this purpose bo oth VW an nd BMW
providedd simulation
n and meassurement re e group.35 The
esults to the T results of the evalu uation of
this proocedure showed rob bust and repeatable values fo or the corrrection co oefficient
determination due to the repro oducible conditions and vehicle be ehaviour. An
A additionaal benefit
of the procedure un nder warm conditions is the fact that this procedure is le
ess time connsuming
because e no soak period
p is neccessary in between
b the
e required tests.
34
c is thhe ratio betweeen the absoluute value of the
t REESS ellectric energyy change and the energy coontent of
the coonsumed fuell. The limit off 0.5% appliees for an applicable test cy ycle consistingg of a L,M,H
H and ex-
H phaase; it is 1.0%% if the appliccable test cyccle consists off a L,M and H phase, and it is 1.5% if the
t
appliccable test cyccle consists off a L and M phase
p only. Refer
R to Appenndix 2 of Annnex 8.
35
For ddetails refer to
o document WLTP-SG-EV
W V-06-11e and document WLTP-SG-EV
W -08-02 at
https:://www2.unecce.org/wiki/ppages/viewpag ge.action?paggeId=23101485
70
Similar to the proceedure undeer cold cond
ditions, the m
manufacture er is allowed to set the
e state of
charge of the trac ction REESS for the correction
c c
coefficient d
determinatioon, with the
e aim to
trigger a positive or
o negative delta REE ESS over th he test. The e break timee during wh hich this
REESS S adjustmen nt takes pla
ace should be less tha an 60 minuttes, and the same bre eak time
should bbe applied for
f each of the tests for reason of repeatabilitty.
For thee procedure e under wa arm conditions the manufacture
m r has to ensure
e thesse warm
conditio
ons prior to each drive en cycle forr the correcction coefficcient determ
mination in order to
arrive at
a repeatable e results. Iff necessaryy the manuffacturer may y conduct an
a additiona al warm-
up proccedure before each tesst. In that ca ase, the samme warm-up has to be e applied to
o each of
the testss required for
f the corre ection coefficient determination.
Both the
e procedure
e under colld condition
ns and the p
procedure under
u warm
m conditionss can be
applied for NOVC--HEVs and OVC-HEVss. The sam me principle is also be applied forr NOVC-
FCHVs..
The flow
wcharts in Figure
F 21 sh
how the seq
quence of activities
a witthin the procedures un
nder cold
and warrm condition
ns.
Figure 21: Flowchharts of procedures for the e determin

nation of the corre
ection
coeffic
cient underr cold and warm
w cond
ditions
Both prrocedures are repeated until the set of measureme

m nt results fulfil the boundary
b
conditio
ons for the correction
c co
oefficient de
etermination.
Bounda
ary conditiions for the
e correction
n coefficie
ent determiination
The RCCB correctioon function is basically determined d by the slope of the lin
near regression line
through the test re esults, with
h CO2/FC on o the vertiical axis annd the energy balance e of the
REESS S on the horrizontal axiss (∆E , ). The acc curacy of th
his slope caan be incre
eased by
adding more test results,
r but is also sen
nsitive to the placemen nt of these points. The
erefore it
was deccided to apply the folloowing two-step approacch in order to receive a set of tessts which
are meaaningful for the correction coefficie
ent determinnation:
1. The first step re
equires at le
east five tes
sts (random
mly placed) with
w two critteria two fullfil
71
2. The second steep requiress only threee tests but with
w additio onal criteria to ensure that the
me accuracyy is provided
sam d as with fivve randomlyy placed tes
sts
In the ffirst step, thet manufa acturer has to be provide at lea ast a set of five testss for the
correctio on coefficie
ent determin nation to the responsib
ble authority
y. The set of
o test resu
ults have
to fulfil tthe following
g criteria:
e test with ∆E
a) The set shall contain at least one , 0 and at least o
one with
∆E , 0
b) The difference in CO2 betw ween the teest with thee highest ellectric enerrgy change and the
test with
w the highest positivve electric energy
e channge, both arre the outerr tests relate
ed to the
electtric energy change,
c sha
all be equall or more than 5 g/km.
The criteria for the first step arre shown fo
or an the exxample vehicle in Figure 22.
Figure 22: Graphiical representation off the criteriia for the fiirst step (at least 5 tests)
In the second step p, the requirred number of tests can be reduce ed to three test if the following
f
criteria are
a fulfilled with respecct to the pla
acement of these
t tests::
a) The difference in CO2 betw ween two ad djacent mea asurementss, related to o the electricc energy
channge during the test, shall be less than
t or equal to 10 g/kkm.
b) The difference in CO2 betw ween the te est with the highest neg gative electrric energy change
c
and the test witth highest positive
p elecctric energy change shall not be le ess than 5 g/km.
g
c) In addition to b) the test with the highest negative e electric ennergy chang ge and the test
with
h highest poositive electrric energy change
c shalll not be witthin the regiion defined by
∆
1%
% 1%.
d) The test in betw
ween the test with the highest
h neg
gative electrric energy change
c and the test
with st positive electric energy change shall be witthin the regiion defined in b)
h the highes
and c).
The critteria for the

e second sttep are shoown for 3 exxample veh
hicles in Fig
gure 23. Th
he areas
defined by b) and c)c are highlighted in bro
own.
72
Figure 23: Graphiical representation off the criteriia for the second
s step
p (3 tests)
After inttense discussions and careful con

nsideration of these criiteria they were
w finally adopted
GTR in the Tokyo mee
for the G eting.
The proocedure foor RCB co orrection off OVC-HEV
Vs, NOVC--HEVs and
d NOVC-FC
CHVs is
describe
ed in Appen
ndix 2 of An
nnex 8.
4.4.19 Shorten
ned test procedure
p e for PEV range test
The tesst procedure e developedd in phase 1a to deterrmine the raange of a pure
p electricc vehicle
(PEV) requires
r to drive the applicable cycle conssecutively until
u the ve
ehicle is no o longer
capablee to follow th
he prescribed speed trrace. This procedure
p c take a lot of time, a
can and also
has a re
epeatability issue. Therefore a shortened tesst procedure e with a calculation meethod for
the rangge determin nation of PEVs
P was proposed
p in
n phase 1b. This meth hod provide es better
repeata
ability on thhe test resuults. This new
n method dology will also reducce the testt burden
conside
erably.
Repeattability issu
ue
With thee consecutive cycle te est procedu ure of phasse 1a, the test
t will finis
sh at an unndefined
point off the applic
cable test cycle
c at the
e moment that the usa able electric c energy has been
depletedd. The acttual vehicle e speed anda accelerration at thhat point (and
( thereffore the
demand ded electrical power fro om the REE ESS) is not the same from
f test to test. The electricity
e
cut poin
nt by the vehicle
v conttrol system m is sensitivve to the actual
a electric power demand,
d
hence the driver be ehavior in te
erms of acccelerating a
and braking may influen nce the testt results.
This cauuses a poorr repeatability of the phhase 1a test method off driving connsecutive cyycles.
Test prrocedure
The me osed in phase 1b determines the range of a PEV by a combinatio
ethod propo on of the
following:
 a sh
hortened tesst procedure
e (STP) to determine
d the usable battery
b enerrgy (UBE), and
a
 a ca
alculation ap
pproach to determine
d the
t pure eleectric range.
73
The fun nction to obtain the purre electric range
r over the whole cycle
c (PER
RWLTC) is dettermined
as follow
ws:.
UBE
PER
EC ,
where :
UBESTP is the usable batttery (REES
SS) energyy determine ed from the e beginningg of the
ned Type 1 test proced
shorten dure until the
e break-off--criterion ha
as been rea
ached.
ECDC,WLLTC is the weighted
w ele
ectric energ
gy consump ption for the applicable WLTP test cycle of
segmen nt 1 and 2 o
of the shorte
ened test procedure (ssee Figure 24)
2
In orderr to shorten
n and simplify the testt proceduree duration, a test sequuence with a higher
consummption of ele ectricity from the REE ESS was prroposed to determine the usable e battery
energy. This test sequence
s w
would reduce the length
h of the tesst procedure
e due to this higher
energy consumptio on, and is sh hown in Fig
gure 24.
Figure 24: Sequence of the shortened test proce

edure for PEVs
ortened testt procedure (STP) conssists of the following se

The sho egments:
 Seggment 1 is used
u to mea
asure the ele
ectric energ gy consumpption at a coold start andd at a
highh SOC levell of the REE ESS. Segme ent 1 has a repetition of
o L and M phases
p at th
he end,
to differentiate between co old and hot phases of L and M.
 Seggment 2 is used
u to mea
asure the ele
ectric energ gy consumpption at a loww level of SOC
S of
REE ESS.
 The constant speed cycle in the midd dle of segme ents 1 and 2,
2 CSCM, is s intended to o
deplete the REESS more rapidly than n by driving the normal applicable cycle. The length
of th
his segmentt depends on o the REES SS capacityy.
 The constant speed cycle at the end of segmentt 2, CSCE, is s intended to
t deplete th he
remaining enerrgy from the e REESS (th his is limited
d to a maxim
mum of 10% % UBE), until the
breaak-off criteriion has bee
en reached.
36
By integ
grating the measured energy
e fro
om the REEESS over thhe whole ST
TP, the totaal usable
battery energy UBESTP is derived. The seelected spe
eed of the CSC
C segme ents is the same
s for
both and should haave a minim
mum of 100 km/h.
36
Note that the curreent from the REESS
R needss to be measuured, but for thhe voltage thhere are other
alternnative methodds: either a measurement,
m or
o a nominal voltage, or thhe on-board signal
s for REE ESS
voltagge (see Annexx 8, Appendixx 3).
74
The purre electric ra
ange PERWLTC
W is obtained not by the actual distance
d driiven during this test
sequencce but by thhe calculatio
on formula provided. Due
D to the constant
c eneergy deman nd at the
CSCE segment,
s th
he influencee of the ele
ectricity cutt by vehicle
e control syystems at the final
moment on test results is minimized. As a co onsequence e, this meethod yieldss better
ability than the method provided in the phase 1a version of the GTR
repeata R.
ary conditiion to use shortened

Bounda s test
When a PEV has an a expectedd range equal to or lon nger than 3 applicable
e WLTP test cycles,
the sho
ortened tes
st procedurre should be b applied. In the ca ase the Exxtra High Phase
P is
exclude
ed from the applicable cycle, this condition iss replaced by a bounddary of 4 ap
pplicable
WLTP test cycles.
If the exxpected range is shorte uld be applied.
er, the conssecutive cyccle test proccedure shou
These ccriteria are specified
s in table A8/3 of Annex 8.
S energy de
REESS eterminatio
on
The REEESS energ gy is determmined by measuring
m he current and voltage of the RE
th EESS in
each phhase. Curreent transduccers are cla
amped on the
t cables that are directly connected to
the REEESS. Alternatively, the on-board current
c measurement data
d may bee used. In th
his case,
the accu
uracy of the
ese data shaall be demo
onstrated to
o the respon
nsible autho
ority.
Voltage
e measuring g equipmen nt is require e at the terminals of REESS.
ed to meassure voltage
Alternattively, the on-board voltage
v meeasurement data mayy be used. In this ca ase, the
accuraccy of these data shall be
b demonsttrated to thee responsib
ble authorityy. For NOVC C-HEVs,
NOVC-F FCHVs and d OVC-HEV Vs, the nom minal REESSS voltage may be usedu instead
d of the
measurred voltage.
tion of the shortened

Validati s test procedure
Main disscussion po
oint for the new proposal was the e of results between the phase
e difference
1a and 1b methods. Especially the impa act of the se
elected constant speed d on the rannge was
ned. In orde
question er to take care of the concerns,
c A
ACEA and JAMA
J providded data to
o support
the STPP method, both
b by mea asurements and simula ations.
Figure 25: Validattion data fo

or shorten test proced ded by ACEA (Renau
dure provid ult)
75
Figure 2 25 shows the variation n of the pure electric range
r again
nst a selectted constan nt speed
for the CCSC segme ents. On the vertical axis
a the calcculated rangge is shown n as a ratio
o against
the rangge determin ned by the consecutive e cycle testt. The rangee graduallyy decreasess with an
increasiing constan nt speed. Thhe differencce of range
e between th he shortene ed and consecutive
test is 1.3% at 120 0km/h. The variation
v width in rangee against th
he constant speed is within
w 1%
between n 80 km/h and 120 km m/h. As a conclusion,
c the STP yields
y a slig
ghtly worsee electric
range, b but is fairlyy close to th
he outcome e of the con nsecutive cycle
c test re
esult. Note that the
speed ofo the CSC segments
s s
should be 100 km/h or higher acco ording to thee GTR.
Figure 226 shows th he variation of the pure
e electric ra
ange agains
st constant speed
s of a different
PEV ass in Figure 25. The range variatio on clearly sshows same tendencyy. The difference of
range between
b the shortene ed and con nsecutive te est was abbout 1.8 kmm at 120km m/h. The
variation
n width in ra
ange againsst constant speed wass below 2 km m of the ratiio between 80 km/h
and 120 0 km/h.

or shorten test proced ded by ACEA (BMW)
dure provid
A similaar evaluation
n on anothe
er vehicle iss presented
d in Figure 27
2 which alsso shows thhe range
variation
n of a PEV against thee selected constant
c speeed. The raange variatio
on shows th
he same
tendenccy as for th he other veehicles. Th
he differencce in range e between the shorten ned and
consecu utive test is
s about 1.2
2% at 120kkm/h. The variation
v width in rangge against constant
c
speed iss below 1%% between 808 km/h and d 120 km/h.

or shorten test proced
dure provid
ded by ACEA (VW)
Figure 228 shows an

a additiona al result provided by JA
AMA to see e the impactt of constan
nt speed
variation
ns on electtric range. The variation width inn the range e against co
onstant spe eed was
0.6% be etween 80kkm/h and 120km/h.
1 The same variation
v alsso accounts
s the UBE and the
energy consumptio on. Figure 28 only sh hows resultss from the shortened test proced dure, so
there is no comparrison to the results for the
t consecu utive test.
76
or shorten test proced
dure provid
ded by JAM
MA
5 through Figure 28 it can

he Figure 25
From th c be conccluded that:
a) tthe results from
f the ST
TP have a good
g ment to the consecutive test result;
agreem
b) tthe impact of
o the seleccted constan
nt speed on
n the test re
esult is not significant,
s
ggenerally within
w 1% be
etween 80 and
a 120 km//h;
c) tthe difference between
n the shorte
ened test pro ecutive test in
ocedure and the conse
rrange is below 2% up to
t constant speeds of 120 km/h;
d) tthe shortened test proccedure yield
ds slightly le ble results consistently..
ess favorab
Since th he STP hass considerab
ble benefitss in terms off increased repeatability and redu
uced test
effort, it was accep
pted as an attractive
a meethod for electric range
e determina
ation.
A possible item to o be discusssed in phhase 2 is th
he applicab
bility of cappped speedd to the
method. PEVs whic ch have a capped
c speed have a longer rangee because the
t energy demand
is less. The test burden for ccapped speeed PEVs could
c be efffectively be
e decreased
d by this
method. However,, the appliccability of this
t method
d to capped d speed PEVs has not n been
discussed during phase
p 1b.
4.4.20 Phase-s
specific values
v forr EVs
Backgrround
During tthe develop pment of the e phase 1a GTR a req quest to obttain phase-sspecific parrameters
for elecctrified vehiccles was made
m by thee Contractinng Party of Japan. Pha ase-specificc means
separate paramete ers for the lo
ow, mid, higgh and (opttionally) extra high phaase of the WLTC,
W in
additionn to the oveerall cycle re
esults. Thiss request was driven byb the desire to compa are more
than only the overa all parameters betwee en different vehicle typees, includin
ng conventio onal ICE
vehicless. This shou uld enable the
t customer to compa are the CO2 emission and the fue el and/or
electric consumptio on also for driving
d in different area
as (urban orr extra-urba
an areas).
While thhese phase e-specific parameters
p had been available fo or conventio onal vehicle
es since
the beginning of phhase 1a, thiis was not the
t case forr electrified vehicles. The
T main re eason for
that is because thet test prrocedure itsself is diffe
erent betwween the OVC-HEVs
O (charge
ng and (CD
depletin D) and a ch harge sustaaining (CS) test) and PEVs (range test). A second
important reason iss that the higher batterry capacity of OVC- an nd NOVC-H HEVs underr charge-
sustaining operatioon conditionns may cause these ve ehicles to drive
d individ
dual phasess with an
SOC im mbalance, because
b th
he chargingg or discha arging durinng a phase dependss on the
operatioon strategy. So while the vehicle e may drive e SOC neutral over th he whole cyycle, the
phases within the cycle may show a non n-neutral SOC. If this potential immbalance would
w not
be correected for ea
ach individuual phase, the phase specific fue el consumpttion would have an
offset ea
ach time tha
at an imbala ance occurss.
77
An overview of the phase-specific parameters that are available for the different EVs is
presented in Table 8, Table 9 and Table 10.
Phase specific values for PEVs

The PEV test procedure to determine the range consists of a certain amount of consecutive
driven cycles using the consecutive cycle procedure (CCP) or the shortened test procedure
(STP). This procedure is explained in the previous paragraph 4.4.19. For the PEV the
approach was to find a mathematical methodology that delivers accurate phase-specific
values without additional testing by driving the same phase consecutively until the battery is
depleted.
A new method that weights the respective electric consumptions of the same phase within
each of the cycles was evaluated. This methodology calculates a weighting factor for each
phase based on the ratio between used energy over that phase and the total usable battery
energy. This weighting factor implicitly includes physical impacts such as the warm up of the
vehicle and the efficiency behavior of the traction battery. Hence, this method leads to a
similar phase specific electric energy consumption and range compared to a vehicle being
tested by driving consecutively the same phase. This evaluation was validated through range
measurements and simulations37 and was then agreed by the EV subgroup in phase 1b.
The parameters available for PEVs are listed in Table 8. Phase-specific values are included
where an ‘x’ is marked under Low, Mid, High and ExHigh.
WLTC
(Low + WLTC
Mid + city
High + (Low +
Parameter exHigh) Mid) Low Mid High ExHigh Explanation
Electric energy consumption determined from the recharged

EC x x x x x x
energy and the equivalent all electric range
EAC x Recharged electric energy
PER x x x x x x Pure electric range
Table 8: Parameters for PEVs
Phase specific values for NOVC-HEVs

As explained above, it is important to take care about a potential non-neutral electric energy
charging balance over one phase for NOVC-HEVs. Therefore it was concluded by the
Subgroup EV that an RCB correction for each phase needs to be applied. This correction
methodology ensures a proportional fuel consumption correction over the phase to the
charged or discharged electric energy during the charge-sustaining test.
The parameters available for NOVC-HEVs are listed in Table 9. Phase-specific values are
included where an ‘x’ is marked under Low, Mid, High and ExHigh.
37
For more information on the validation refer to documents WLTP-SG-EV-09-14, WLTP-SG-EV-06-09-
rev1, and WLTP-SG-04-10 at https://www2.unece.org/wiki/pages/viewpage.action?pageId=23101485
78
WLTC
(Low + WLTC
Mid + city
High + (Low +
MCO2,CS x x x x x CO2 determined from the charge‐sustaining (CS) test
FCCS x x x x x Fuel consumption determined from the CS test
Table 9: Parameters for NOVC-HEVs
Phase specific values for NOVC-HEVs

The same need for RCB correction on each phase of course also applies for the OVC-HEVs
charge-sustaining test. However, the NOVC-HEVs are tested in charge-depleting mode as
well, and these additional parameters make the determination of phase-specific parameters
even more complex. For some of the parameters a weighting according to the utility factors
has to be applied (see paragraph 3.4.5.8). The group decided to exclude these from the
phase-specific calculation. The main reason is that the utility factors are not available at a
phase-specific level, which means that it is not sensible to calculate phase-specific weighted
values. Furthermore, the non-weighted phase-specific values already meet the requirement
of being comparable to conventional and pure electric vehicles.
Some more investigations had to be done to determine the phase-specific electric energy
consumptions and electric ranges by a calculation methodology from the charge-depleting
test results. Due to the primary requirement to deliver parameters that can be compared with
the electric energy consumption and electric range of PEVs, the group focused on the
parameters EC (electric consumption) and EAER (equivalent all electric range). Supported
by simulations38 it was shown that a similar weighting approach as applied for the PEVs
leads to sufficiently accurate values, which can also be interpolated for individual values.
The parameters available for NOVC-HEVs are listed in Table 10. Phase-specific values are
included where an ‘x’ is marked under Low, Mid, High and ExHigh.

WLTC
(Low + WLTC
Mid + city
High + (Low +
MCO2,CD x CO2 determined from the charge‐depleting test (UF weighted)
MCO2,CS x x x x x CO2 determined from the charge‐sustaining (CS) test
MCO2,weighted x Utility factor weighted CO2 determined from the CD and CS test
FCCD x Fuel consumption determined from the CD test (UF weighted)
FCCS x x x x x Fuel consumption determined from the CS test
38
For more information on the validation refer to documents WLTP-SG-EV-05-08, WLTP-SG-EV-08-05-
rev1, WLTP-SG-EV-09-08, WLTP-SG-EV-09-13 at
https://www2.unece.org/wiki/pages/viewpage.action?pageId=23101485
79
Utility factor weighted fuel consumption determined from
FCweighted x
the CD and CS test
Electric energy consumption determined from the CD test

ECAC,CD x
(UF weighted)
Utility factor weighted electric energy consumption determined

ECAC, weighted x
from the CD test
Electric energy consumption determined from the recharged

EC x x x x x x
energy and the equivalent all electric range
EAC x Recharged electric energy
RCDC x Charge‐depleting cycle range
All electric range determined from the CD test (distance until first
AER x x
engine start)
Equivalent all electric range determined from CD and CS test
EAER x x x x x x
(pure electrically driven distance)
Actual charge‐depleting range determined from CD and CS test

RCDA x*
(distance driven in CD operation)
Table 10: Parameters for OVC-HEVs
4.4.21 Interpolation method for electrified vehicles

Background
During the development of the phase 1a version of the WLTP GTR an interpolation method
was introduced for conventional vehicles that enables the calculation of individual CO2
emission and fuel consumption values based on the specific cycle energy demand of an
individual vehicle. Basis for the interpolation is the measurement of two extreme vehicle
configurations regarding their fuel consumption/CO2 emission within one vehicle family. To
ensure the accuracy between interpolation and measurement, vehicle family criteria had
been defined. For more information on the interpolation method see paragraph 4.4.1.
The aim of the Subgroup EV was to adopt a similar interpolation methodology -tailored to
electrified vehicles- to be also capable to calculate vehicle-individual values for these
vehicles39. To identify which modifications might be necessary to the existing method the
group decided to evaluate this separately for NOVC-HEVs, OVC-HEVs and PEVs. Originally
the need for this vehicle classification was based on the fact that the main component-based
criteria for the vehicle family building are different between these vehicle groups. For
example it is important to focus on the electric components of all electrified vehicles for the
family building but in the case of NOVC- and OVC-HEVs one has to consider the ICE as
well. Since OVC-HEVs can be driven in charge sustaining and charge-depleting operation,
the methodology has to take care about much more parameters having to be interpolated.
39
For an overview of which values are determined in the GTR for EVs, see the tables in paragraph 4.4.20
80
Interpolation method for NOVC-HEVs
Due to minor differences between the test procedure of conventional vehicles and NOVC-
HEVs the evaluation started with this vehicle type. The road load and interpolation family
criteria were extended with the electric components that might have an impact on road load,
CO2 emission or fuel consumption but are not covered by the cycle energy based
interpolation. The CO2 interpolation range within one family compared to conventional
vehicles was reduced to avoid the potential risk of non-linear effects; an additional test with a
vehicle in the middle of the outer ones of the family (regarding the cycle energy) is required if
the CO2 interpolation range should be extended above 20 g/km. This is described in
paragraph 4.5.1 of Annex 8.
Interpolation method for OVC-HEVs

Since OVC-HEVs have to conduct two tests under different test conditions (charge-depleting
and charge-sustaining), the number of values to be interpolated is much larger than for other
vehicle categories. This variety in parameters and the fact that some values are calculated
from both tests leads to more complex handling of cycle- and phase-specific values.
Therefore it is not always possible -or only under certain conditions- to interpolate the
parameters that are determined for OVC-HEVs. Hence the following amendments were
necessary:
a) The charge-depleting cycle range RCDC and the actual charge-depleting cycle range
RCDA are excluded from the interpolation method due to their non-linear behaviour.
b) The all-electric range AER can only be interpolated if it fulfils a specific criterion.
c) An additional restriction for the application of the interpolation method is introduced.
Ad a): The charge-depleting cycle range RCDC is a discontinuous parameter because it is

defined as the number of complete cycles driven in CD operation multiplied by the cycle
distance. This means that a different number of cycles within one family leads to a jump from
x*23.3 km to (x+1)*23.3 km. The second parameter to be excluded is the actual charge-
depleting range – RCDA. This describes the distance at which the REESS is fully depleted and
the vehicle is only capable to continue in charge-sustaining operation. This parameter cannot
be interpolated due to the rising power demand (coming from vehicle L towards vehicle H),
while the available electric power is the same within one family. This is illustrated by the
following example. Coming from vehicle L to vehicle H the logical response for individual
vehicles is that the RCDA first will start to decline due to higher electric energy consumption.
This relation is linear until the power demand exceeds the available electrical power of the
driveline. This will trigger the ICE to assist the electric motor, so for this individual vehicle
also energy from the combustion engine is used to follow the drive cycle. This leads to an
increase of the RCDA. For the remaining vehicles towards vehicle H it depends on the
operation strategy what the RCDA value will arrive at. Due to this non-linearity the RCDA is
excluded from the interpolation.
Ad b): Consider the following example. Vehicle L has just sufficient electric power to fulfil the
cycle without the ICE having to assist. This means that the first engine start of vehicle L will
not take place until the REESS has been depleted. The other vehicles in the family would
have an engine start in each of the cycles at the point(s) where the electric power is not
sufficient to follow the prescribed speed trace. This leads to a discontinuity in the AER that
prevents an accurate interpolation. However, this situation may not always be the case.
Therefore a criterion was developed to detect if a discontinuity is present or not. This criterion
is the ratio of AER to RCDA, which should not differ more than 0.1 between vehicle L and H. If
this criterion is met, the interpolation of AER is permitted, otherwise the worst-case AER
value applies to the whole family. This is described in paragraph 4.5.7.1 of Annex 8.
Ad c): An additional restriction for the interpolation is that the number of whole cycles driven
in the CD test should not differ more than 1 between vehicle L and H. On the one hand this
81
requirement allows to build an interpolation family even if the number is not the same for all
vehicles, and on the other hand restricts that the interpolation range is so wide that the
linearity is compromised.
All other parameters listed in Table 10 can be interpolated without further requirements.
Interpolation method for PEVs

For the pure electric vehicles (PEVs), the ICE-based interpolation family criteria had to be
converted from those that apply to a conventional driveline to those that apply to the “electric
machine”, “electric converters” and the “REESS”. The PEV relevant parameters “electric
consumption – EC” and “pure electric range – PER” are well suited for interpolation because
the relation between cycle energy demand and EC is also linear. The PER also responds
linear because it depends on the recharged energy, which will be constant as the same
REESS required to be used throughout the interpolation family. These linear relations are
independent from applying the consecutive cycles testing method or applying the shortened
test procedure. To ensure the linearity of the PER for the CCP it was concluded in phase 1b
that it should be calculated from the electric energy consumption and the usable battery
energy, rather than just measuring the range from the test directly. Otherwise a non-linearity
could be introduced because the energy consumption itself depends on the specific phase
considered.
Validation
The development of the interpolation method and the additional required criteria and
restrictions took a lot of effort by the Subgroup EV participants. During the course of phase
1b the group produced evaluations of measurement data and performed simulations to
substantiate the proposed interpolation methods40. In the end they could all agree to the
approaches described in this paragraph, and the methods were adopted.
In phase 2 of WLTP the group will focus on the interpolation method and criteria for FCHV.
4.4.22 End of PEV range criteria

Background
According to the GTR phase 1a, the range test for PEVs is terminated when the break-off
criterion is reached, which means that the vehicle is not capable to follow the prescribed
speed trace for 4 consecutive seconds or more41. For vehicles with a speed cap (i.e. a
maximum speed limiter) lower than the maximum speed of the applicable WLTP test cycle
this would result in a non-representative pure electric range. This is because the break-off
criterion would already be reached during the first cycle, even though the REESS is not yet
depleted. The Subgroup EV was tasked to develop a solution for this issue.
Discussions during phase 1b and adopted solution
The discussions first focused on PEVs but soon extended to OVC-HEVs, which also have a
purely electrically driven range. This is referred to as the all-electric range AER, and this
range would also be unrepresentatively small for OVC-HEVs with a capped speed.
40
For more information about the validation refer to documents WLTP-SG-EV-05-02, WLTP-SG-EV-06-
04, WLTP-SG-EV-06-05, WLTP-SG-EV-08-04, WLTP-SG-EV-08-05, WLTP-SG-EV-09-02 at
https://www2.unece.org/wiki/pages/viewpage.action?pageId=23101485
41
Refer to paragraph 3.4.4.1.3 of Annex 8
82
One of the issuess during the e discussio
ons was that a manufacturer wh ho has dessigned a
vehicle for urban coonditions annd applies a speed capp at e.g. 90km/h would d get penalizzed by a
mall electric range, for example
very sm e only 17 km. This
T unrepre esentative electric
e rang
ge would
not servve as a useeful consum mer informaation either, since the driver would not experience a
range oof just 17 km
m but would d be able to
o drive maybe 150 km or more (ju ust as an exxample).
Therefoore it was cle
ear that a solution
s had to be foundd for this iss
sue.
Anotherr concern of
o this capp ped speed is that it consumes
c less energyy during th he cycle,
becausee the energgy demand is reduced d at a lowe er speed. At
A the same e time, thatt vehicle
would b
be driving a shorter disttance during
g the test, which
w is also
o not repressentative.
Taking these conccerns into acccount, a methodology
m y was deve eloped to lengthen the cycle to
such an n extent th
hat the cap pped speed d cycle covers the same distan nce as the
e normal
(uncappped) cycle. During thiss elongationn the vehicle e is driven at its higheest (capped) speed.
This approach is coonsidered representati
r ive for real--life driving, since a cap
pped speedd vehicle
a-urban area
in extra as would have
h to drivve longer att its maxim mum speed to cover th he same
distancee.
Figure 229 shows how
h this lengthening off the capped cycle is ta
aking place
e for differen
nt speed
caps. Each cycle shown
s has the
t same overall
o distance. Note that
t the elongation is done
d per
individual phase.
Figure 29: Capped speed cy es for differrent speed caps

ycle profile
This topic was intensively discussed

d a there ha
as ad been tw
wo opposite
e positionss by the
Contraccting Parties
s of Europe and Japan.
Position
n of the Eurropean Com
mmission was that this methodolo b applied for each
ogy should be
capped speed and d in any ph
hase where the capped speed wo ould modifyy the speed d profile.
The possition of Ja
apan was not to apply this metho odology at all,
a motivated by theirr starting
point th
hat the cyccle should not be mo odified, in o
order to ennsure that test
t resultss remain
comparrable and therefore havve to be bassed on the ssame cycle.
Due to tthese oppos
site position
ns a regiona
al solution was
w implem
mented in the
e GTR as fo
ollows:
83
For Europe:
If the (capped) maximum speed of the vehicle is lower than the maximum speed of the
applicable WLTP test cycle, Europe will apply the capped speed cycle with a proportional
elongation of the cycle to arrive at the same cycle distance.
For Japan:
If the maximum speed of the vehicle is lower than the maximum speed of the applicable
WLTP test cycle, Japan will abstain from driving the applicable WLTC. Only the WLTCcity
results will be reported.
The disharmonization between Japan and Europe is fairly limited because for Japan the
‘Extra-High’ phase is excluded from the applicable WLTC. Effectively this means that there is
only a difference between Europe and Japan for vehicles with a capped speed below the
maximum speed of the ‘High’ phase (i.e. 97.4 km/h). Taking the speed trace tolerance into
account, this speed border is further reduced to 95.4 km/h.
The capped speed approach is also reflected in the context of the selection of the driver-
selectable mode, which is described in paragraph 3.4.5.10.
The capped speed cycle modification can be found in paragraph 9 of Annex 8.
4.4.23 FCV test procedure

The NOVC-FCHV test procedure was developed for the phase 1b version of the GTR. It is
basically the same procedure as for NOVC-HEVs, but replaces the measurement of CO2 by
a method to determine the hydrogen consumption of NOVC-FCHVs.
Typical methods used today to measure hydrogen consumption are the following:
a) Gravimetric method:
The weight of the consumed hydrogen is measured as a weight difference of an
external hydrogen tank before and after the test.
b) Flow method:
The integrated value of a hydrogen flow through a tube between the tank and the fuel
cell system is measured.
c) Pressure method:
The pressure decrease of the hydrogen tank is measured, and calculated into a
hydrogen consumption.
The gravimetric method provides a direct way to measure the amount of consumed
hydrogen, while the flow and pressure method need to be calculated and are influenced by
ambient conditions. For the phase 1b version of the GTR the gravimetric method is therefore
prescribed as the primary method. The measurement procedure is largely based on the
procedure described in ISO 23828.
At the request of the manufacturer and upon approval of the responsible authority the
consumption may be measured using either the pressure method or the flow method as an
alternative to the gravimetric method. In this case, the manufacturer has to provide technical
evidence that the method yields equivalent results.
In order to obtain a sufficient degree of accuracy with the pressure and the flow method it is
required to give special attention towards e.g. the temperature management of the test tank
and the preparation/calibration of the high accuracy flow meter. The pressure and flow
methods are also described in ISO 23828, which can be used as a basis for these
requirements.
84
Just as for NOVC-HEVs also NOVC-FCHVs have to be corrected towards a neutral charging
balance if they do not meet the tolerance criteria. More information on the RCB correction
procedure can be found in paragraph 4.4.18. As the configuration of the power train of
NOVC-FCHVs is similar to that of (N)OVC-HEVs, this means that the hydrogen consumption
of NOVC-FCHVs needs to be corrected for the electric energy change of all REESSs.
The NOVC-FCHV test procedure is described in paragraph 3.5 of Annex 8, and the RCB
correction is included in Appendix 2 to Annex 8.
Due to the time constraints of phase 1b and the lower priority that FCVs received, not all the
open issues could be solved. Therefore the scope of WLTP phase 2 should include the
following issues:
 Test procedure for OVC-FCHV
 Interpolation approach for NOVC-FCHV and OVC-FCHVs
4.4.24 WLTP post-processing

Within the "Drafting Taskforce" (see paragraph 3.4.1), which was in charge of implementing
editorial changes to the GTR, the following problem was identified: For historical
reasons, every correction, such as RCB correction, Ki-factors or averaging of tests was
handled separately. Therefore it was not clear, in which order which correction should be
applied. Especially, it was unclear how to apply corrections on fuel consumption, because
that is based on CO2 and criteria emissions, which are both subject to correction
requirements. In addition some of the references were incorrect, due to the fact that the
correction steps were developed in parallel.
This called for the need of putting the calculation steps into a logical order, provide a
complete overview of the post-processing procedure in the GTR, and to set the references
accordingly.
Motivation
The requirement of applying corrections is obvious, because test results can only be
comparable if they are corrected towards standard conditions. But as the order may have a
slight influence on the end result (due to fact that some corrections are additive yet others
are multiplicative), this needs to be specified to avoid confusion between industry, authorities
and organizations performing in-use tests. An addition bonus is that a clear overview makes
references easier and the list of the corrections more transparent.
Description
The need for including an order into the corrections is due to the interdependency between
the following issues:
a) Calculation of phase specific values;
b) Calculation of fuel consumption out of CO2 and criteria emissions;
c) Additive corrections, e.g. the Ki factors (creating non-linearity if the order is changed);
d) Averaging of tests;
e) Concept of a "declared value";
f) Regional options (e.g. 14°C test in Europe, different declared value concept).
As there will be always a small error induced when the order of calculation steps is changed,
the following priority was decided:
85
ulate criteria
1) Calcu a emissionss and CO2.
2) Calcu
ulate fuel co
onsumption based on 1).
1
Apart frrom the requirement that the en nd result should

s be meaningful
m and accurrate, the
following objectivess were also
o strived for::
 E
Enable an alignment
a w calculattions for hyb
with brid vehicless;
 E
Enable regional correcction(s) within one step
p (a placeho
older in the GTR);
G
 R
Reduce unn
necessary calculation
c a correcttion efforts,
and
As a ressult of the last
l ded to shift the fuel consumption calculation towards
point, itt was decid
the end of the calculation proccess.
The finaal post-proccessing schheme that was
w adopte ed is shownn in the schheme of Fig gure 30.
The chaarge-sustain ning calcula
ations for NOVC-
N and OVC-HEVss and ICE vehicles
v ha
ave been
aligned as much as a possible.. The orderr of applying the calcu
ulation/correection stepss is from
top to bottom.
b Thee small columns on the e right show
w the outputt values of each step. For fuel
cell veh
hicles the same
s proce
ess can be applied, butb in that case
c the mass
m emissions are
replacedd by fuel co
onsumption..
Figure 30: Post-p

processing scheme with
w the orrder of callculations and correc
ctions
within the GTR fo
or ICE and HEV
Validati
tion and jus
stification
To checck the validity of the prroposed po ost-processing, an Exceel-tool was provided to o enable
stakeho
olders to cheeck the orde er of the se
equence andd the effect this has on
n the resultss.
The pro oposal of Figure
F 30 wwas concluded to deliver meaningful and sound
s resu
ults, and
thereforre no further validation was consid
dered necesssary.
Develop
pment process
From thhe moment that this isssue was id dentified theere was a broadb suppport of clariffying the
calculattion/correction order within the GT TR itself. Affter the firstt starting no
ote in summ mer 2015
the deve elopment was
w mainly done
d via e-m
mail exchan nge and the e final propoosal was adopted at
86
the 12th IG meeting. Due to the short timeline, the drafting text was agreed shortly after that
meeting in October 2015.
The scheme for post-processing is included in Table 7/1 in paragraph 1.4 of Annex 7. For the
calculations in charge-sustaining condition of hybrid electric vehicles (NOVC-HEVs and
OVC-HEVs) it can be found in Table A8/5 and A8/6.
Due to the fact that for fuel cell hybrids (NOVC-FCHVs):
a) the interpolation method for will be handled in phase 2,
b) a calculation of fuel consumption is not necessary because it is measured directly, and
c) the Ki-correction is not applicable,
some of the steps shown in Figure 30 are removed and/or amended. This alternative post-
processing scheme is shown in Table A8/7 of Annex 8.
The post-processing scheme for the calculation of electric ranges, electric consumptions and
weighted parameters for OVC-HEVs and PEVs will be discussed in phase 2.
4.5 GTR structure

The GTR covers every aspect on emission testing to the last detail and consequently it has
become a large document. For someone who is not familiar with it, the amount of information
contained in the GTR can be overwhelming. Even though a clear structure was used, not all
of the test requirements are always found at the place where they would intuitively be
expected. As an introductory guide for those that are relatively new to the GTR, this
paragraph summarizes the contents of the Annexes which are related to the test procedure.
Annex 1 and 2 are missing in this overview since they are covered by the technical report on
the DHC2.
4.5.1 Annex 3 – Reference fuels

The structure of annex 3 has to be seen as temporary. In phase 1 of the GTR development it
is merely a re-formatted list of the specifications of reference fuels that are in current usage
in the Contracting Parties. This serves two purposes, one is to provide technical specification
values to reference in the calculation formulae throughout the GTR and the second is to offer
specifications to Contracting Parties in the future in an attempt to prevent further
disharmonisation.
In conclusion, the list of reference fuels included in the Annex 3 serve as a guideline, albeit
non-binding.
The structure can and probably will change with any attempt to harmonise reference fuels in
later phases of WLTP.
4.5.2 Annex 4 - Road and dynamometer load

This Annex describes the determination of the road load of a test vehicle and the transfer of
that road load to a chassis dynamometer. The road load is a 2nd order polynomial
approximation of the vehicle's losses determined by using one of the available methods.
In this paragraph the options and the procedure are briefly outlined and explained.
87
General requirements
Road load can be determined using the coast down method, torque meter method and the
wind tunnel method. In addition, road load may be estimated at a (conservative) default
value, or may be ‘extrapolated’ from a measured representative vehicle.
To compensate the effects of wind on the road load determination procedure, the wind
conditions need to be measured. Two methods are possible: using stationary anemometry
alongside the test track (in both driving directions if the track has an oval shape), or by using
on-board anemometry. The latter method has more relaxed limitations towards the maximum
wind speeds under which it is allowed to determine the road load.
The temperature window within which the road load determination tests take place is
specified as 278 to 313 K (5 to 40°C), but on regional level Contracting Parties may deviate
up to +/- 5 K from the upper limit, and/or lower the range to 274 K.
Vehicle selection
Vehicle H is selected for the road load determination, being the vehicle within the CO2
vehicle family with the combination of road load relevant characteristics (i.e. mass,
aerodynamic drag and tyre rolling resistance) producing the highest cycle energy demand
(see also par. 4.4.2 of this report). If the manufacturer wants to apply the CO2 interpolation
method, additionally the road load is also measured on vehicle L. This is the vehicle within
the CO2 vehicle family with the combination of road load relevant characteristics (i.e. mass,
aerodynamic drag and tyre rolling resistance) producing the lowest cycle energy demand.
Aerodynamic drag
Any movable aerodynamic body parts has to operate in the same way as they would do
under conditions encountered in the Type 1 test (test temperature, speed, acceleration,
engine load, etc.). A moveable spoiler for stability at higher speeds, as an example, may
move out or retract in the same way as it would do on the road. However, this requirement is
not intended to be ill-treated to determine an unrealistic low road load. If such practices are
observed or suspected, appropriate requirements will have to be added at a later stage.
For the determination of aerodynamic drag differences within the vehicle family a windtunnel
has to be used. However, not every windtunnel may be fitted with a moving belt, which is
needed to properly establish the drag of different wheel rim/tyre combinations. In such cases,
the manufacturer may alternatively propose a selection based on wheel rim/tyre attributes
(see 4.2.1.2 of Annex 4). If the wheel rim/tyre selection for vehicle H is done by this
alternative approach, the CO2 regression method cannot be used for the wheels, and the
worst-case wheel rim/tyre combination is applied for all vehicles within the vehicle family.
Vehicle preparation
The test mass of the vehicle is measured before the road load determination procedure
starts, and is verified to be equal or higher than the specified test mass. After the road load
determination procedure is finished, the mass of the vehicle is measured again. The average
of the mass before and after testing is used as input for the calculation of the road load curve
(see also paragraph 4.4.4 of this report).
The selected vehicle needs to conform in all its components and settings (e.g. tyre selection,
tyre pressures, wheel alignment, ground clearance, vehicle height, drivetrain and wheel
bearing lubricants) to the corresponding production vehicle. It is allowed to be run-in for
10,000 to 80,000 km, but at the request of the manufacturer a minimum of 3,000 km may be
used.
88
If the vehicle is equipped with a vehicle coastdown mode (see paragraph 4.4.5 of this report),
it needs to be activated both during the road load determination procedure as during tests on
the chassis dynamometer.
The tyre tread depth needs to be at least 80% of the original tread depth over the full width of
the tyre, meaning that the outer shape of the worn tyre is similar to that of a new tyre. This
requirement needs to be checked before starting the road load determination procedure. To
prevent that the tread depth is further reduced by all of the testing activities, this
measurement is only valid for a maximum of 500 kilometres. After this 500 kilometres, or if
the same set of tyres is used for another vehicle, the tread depth has to be checked again.
Tyre pressure is set to the lower limit of the tyre pressure range specified by the
manufacturer for the specific tyre, and is corrected if the difference between ambient and
soak temperature is more than 5 K.
Vehicle warm-up
If the vehicle is tested on the road or at a track, it is warmed up by driving at 90 % of the
maximum speed for the applicable WLTC (or 90 % of the next higher phase if this is added to
the applicable cycle). Before the warm-up it will be decelerated by moderate braking from 80
to 20 km/h within 5 to 10 seconds. This procedure prevents any practices to reduce parasitic
losses from brake pads touching the brake discs.
Measurement procedure options

The GTR provides in five different methods that can be used to determine the road-load of
the vehicle:
a) Coastdown method: A vehicle is accelerated to a speed above the highest reference
speed, and is decelerated by coasting down with the transmission in neutral.
b) Torque-meter method: Torque meters are installed at the wheels of the vehicle, and
the torque is measured while the vehicle travels at constant reference speeds.
c) Matrix family method: The road-load is measured on one representative member of a
family, and ‘extrapolated’ to other family members by considering the difference in the
dominant road load parameters.
d) Windtunnel method: The aerodynamic drag of the vehicle is determined in a
windtunnel, and the rolling resistance is added by measurement on a flat belt or a
normal chassis dynamometer.
e) Default road load: Instead of measuring the road load, the manufacturer may choose
to use a ‘default road load’ which is based on vehicle parameters
The road load is presented as a second order polynomial approximation of the vehicle's
losses when dragged or when it is coasting. In general road load has to be determined in the
speed range of the applicable test cycle, but due to regional deviations also up to higher
speeds, to use a test result for more than one region42.
An overview of the available road load determination options and references to the
paragraphs describing the procedure and the results is provided in Table 11.
42
For example Japan does not include the extra-high phase of the WLTC in their applicable test cycle.
89
Road load determination methods and their Reference Reference Road load
options and alternatives to method to result coefficients
Coast down with stationary anemometry 4.3.1. 4.3.1.4.5. f0 , N

and 4.5. f1, N/(km/h)
coast
- with or without split runs f2, N/(km/h)²
down
Coast down with on-board anemometry 4.3.2. 4.3.2.6.7. f0 , N
(road- (with different possible positions of the and 4.5. f1, N/(km/h)
based) anemometry) f2, N/(km/h)²
- with or without split runs
Measurement of running resistance using 4.4. 4.4.4. and c0, Nm

the torque meter method 4.5. c1, Nm/(km/h)
torque
c2, Nm/(km/h)²
meter
(road- - if coast down on dynamometer according f0 , N

based) to 8.2.4. has been performed f1, N/(km/h)
f2, N/(km/h)²
Calculation for road load for a road load 5.1. 5.1. f0 , N

matrix family f1, N/(km/h) *)
matrix
f2, N/(km/h)²
(road-
based - based on coast down or torque meter c0, Nm
+calc.) measurement c1, Nm/(km/h) *)
c2, Nm/(km/h)²
default Calculation of default road load based on 5.2. 5.2. f0 , N

vehicle parameters f1, N/(km/h) *)
(calc.)
f2, N/(km/h)²
Measurement of road load within labs by 6. 6.7.3. f0 , N

wind tunnel and a dynamometer f1, N/(km/h)
wind
f2, N/(km/h)²
tunnel
- with a flat belt dynamometer
(lab- - - with stabilised speeds or with

based) deceleration
- - - with warm-up by driving or warm-up by

dragging the vehicle
- with a roller chassis dynamometer plus

correction function
- - with stabilised speeds or with

deceleration
- - - with warm-up by driving or warm-up by

dragging the vehicle
*) This coefficient is set to zero for this method
Table 11: Overview of available road load determination methods and options, with
reference to paragraphs in the GTR on the procedure and the results.
90
The characteristic differences between these methods are shown in Table 12
Method Coast down Torque meter RL matrix Default RL Wind tunnel

family
Focus/scope passenger passenger Large vans for small passenger vehicles

vehicles vehicles, above 3 tons series
max. laden
wheel hub
mass
motor
Measured velocities and wheel torque (extrapolate Nothing is air drag and
value times during at constant measured measured drivetrain plus RR
coastdown speeds RL) losses
Positive and + well known + shorter test + balanced +cheapest + reproducible and
negative track compromise method weather
+ simple
attributes between test independent
measurement + measures + no test
effort and
equipment "real" road effort + accuracy
accuracy
load
- long test track - worst + suitable for secret
- slightly
needed - weather case road designs
worse road
dependent load
- weather load (safety - expensive
dependent - complex margin) equipment
process
- inaccurate
Table 12: Characteristic differences between the road determination methods
Measurement procedure – Coastdown method

The coastdown method itself can also be conducted in two different ways:
a) Multi-segment method with stationary anemometry (paragraph 4.3.1 of Annex 4)
b) On-board anemometer-based coastdown method (paragraph 4.3.2 of Annex 4)
Ad a): Reference speeds are selected over the speed range of the applicable cycle from 20
km/h upwards in steps of 10 km/h. The highest reference speed is 130 km/h or the reference
speed point immediately above the maximum speed of the applicable test cycle. The vehicle
is coasted down from at least 5 km/h above the highest reference speed to at least 5 km/h
below the lowest reference speed. Though it is recommended that coastdown runs are
performed without interruption over the whole speed range, it is allowed to split the runs (e.g.
if there is not sufficient length on the test track) while taking care that vehicle conditions
remain as stable as possible. Coastdown runs are repeatedly performed in opposite driving
directions until the statistical accuracy is satisfied. The coastdown time at each reference
speed is determined by calculating the harmonised time averages of runs (separately for
opposite directions). By taking the vehicle inertia into account, the deceleration curve can be
used to calculate the road load force for each reference speed. Vehicle inertia is calculated
by taking the average of the vehicle mass before and after the road-load determination
procedure, increased by the equivalent effective mass mr of wheels and other rotating
components. The sets of reference speeds and corresponding road load force are used to fit
a second-order polynomial regression curve with the road load factors f0, f1 and f2. This
procedure is done for both driving directions separately, and the average of the road load
factors is calculated from it. As a final step, the road load factors are corrected for the
average wind speed, actual test mass, temperature effect on rolling resistance and
deviations from standard temperature and pressure affecting the aerodynamic drag.
91
Ad b): The vehicle will be equipped with on-board anemometry to accurately determine the
wind speed and direction. During testing, the anemometer may be located on the centreline
of the vehicle via a boom approximately 2 meters in front, at the midpoint of the vehicle’s
hood (bonnet), or at least 30cm back of the windshield on the vehicle’s roof. The maximum
allowed overall average wind speed during the test activity is 7 m/s and peak wind speeds
should not exceed 10 m/s. In addition, the vector component of the wind speed across the
road shall be less than 4 m/s. The wind criteria were chosen to fall within the restrictions
specified in SAE J2263, with lower tolerances to decrease potential test variability due to
wind influence. The test procedure is similar as for a), but at least 5 coastdown runs are
performed in each direction. The results from the coastdown curves and the anemometry
data are combined in an ‘equation of motion’. In a complex calculation procedure the
parameters that define the road load curve are derived. The correction for wind is implicitly
included in this process, while the equation of motion is afterwards corrected to reference
conditions. For the test to be validated for WLTP, the results must pass the statistical
convergence requirement.
Measurement procedure – Torque-meter method

One alternative for coastdown testing method is the torque-metering method (see paragraph
4.4 of Annex 4), which has the following fundamental differences:
a) Instead of calculating the road load indirectly from the deceleration curve, the torque is
measured directly at the wheels (which can be translated into a resistance force with the
dynamic radius of the tyre). Therefore, this method can be applied with the vehicle at
constant speed. If a vehicle has non-reproducible forces in the driveline which cannot be
prevented by the coastdown mode, the torque meter method is the only method available
for road load determination.
b) Since the torque meter is usually installed between the wheel hub and tyre rim, all of the
resistances upstream in the driveline of the vehicle are not measured. The torque-meter
method therefore finds a lower resistance force than the coastdown method. To avoid
mixing up these forces, the coastdown method is said to determine the ‘total resistance’,
while the torque-meter method determines the ‘running resistance’. To obtain a proper
setting of the chassis dynamometer, the vehicle with torque-meters installed will be put
on the dyno, and the running resistances found on the track are reproduced. Once the
chassis dynamometer is set, a coastdown will be executed, from which the road load
factors can be derived for any subsequent testing purposes. Of course, if the vehicle has
non-reproducible forces in its driveline, the chassis dynamometer can only be set with
torque-meters installed.
The test procedure for the torque-meter method also involves the use of fixed reference
speeds from 20 km/h upwards in incremental steps of 10 km/h to a maximum of 130 km/h
(see paragraph 3.4.5.5). The vehicle is driven at each reference speed for a minimum of 5
seconds, while the speed is kept constant within a small tolerance band. Measurements are
repeated in opposite driving directions and compensated for speed drift, until the statistical
accuracy is satisfied. The sets of reference speeds and corresponding resistance torques are
used to fit a second-order polynomial regression curve with the running resistance factors c0,
c1 and c2, which describe the wheel torque as a function of vehicle speed. This procedure is
done for both driving directions separately, and the average of the running resistance factors
is calculated from it. As a final step, the running resistance factors are corrected for the
average wind speed, actual test mass, temperature effect on rolling resistance and
deviations from standard temperature and pressure affecting the aerodynamic drag.
92
Measurement procedure – Road load matrix family
The road load matrix family method is intended for vehicles produced in low-volumes, and its
scope is reduced to vehicles above 3 tons. The road-load is measured on one representative
member of a family, and ‘extrapolated’ to other family members by considering the difference
in the dominant road load parameters. This method is introduced in paragraph 4.4.9 of this
report, and is further detailed in Appendix 2 of this report.
Measurement procedure – Windtunnel method

The resisting force on a vehicle is a combination of the aerodynamic drag, and the rolling
resistance. The windtunnel method determines these resistances separately:
a) the aerodynamic drag of the vehicle is determined in a windtunnel, and
b) the sum of rolling resistance and drive train losses is measured on a flat belt or a chassis
dynamometer.
This method allows road load measurements to be independent from the weather conditions
and produces accurate, repeatable and reproducible results.
The method is described in paragraph 4.4.10 of this report.
Default road load

The third option for road load determination is to abstain from measurements on a track, by
using default values for the road load factors (see paragraph 4.4.7 of this report). This may
be a cost-effective alternative, especially in case of small production series or if there are
many variants in one vehicle family. The default road load values are based on the test mass
of the vehicle as an indicator for rolling resistance, and the product of vehicle width and
height as an indicator for aerodynamic drag. To prevent that these default values would
create an advantage over measured road load, they have been developed to go towards a
worst-case.
Preparation for the chassis dynamometer test

The first step in the chassis dynamometer test is to set the equivalent inertia mass. This
mass is the same as the average mass of the vehicle during the road load determination
procedure. In contrast to Regulation 83 there are no inertia steps, so the setting has to meet
the test mass exactly, or – if that is not possible – the next higher available setting. In case a
single-axis dynamometer is used, one pair of wheels is not rotating. To compensate for this,
the inertia mass is increased by the equivalent effective mass mr of the non-rotating wheels
(if that information is not available, this may be estimated at 1.5 per cent of the unladen
mass).
In the next step, both vehicle and chassis dynamometer are warmed up as indicated in the
GTR. The warm-up procedure for the vehicle is the applicable test cycle. Alternatively, the
manufacturer may use a shorter warm-up cycle for a group of vehicles, but only at the
approval of the responsible authority after demonstrating equivalency.
Chassis dynamometer load setting

The purpose of the chassis dynamometer setting is to reproduce the load that was found in
the road load determination process as close as possible. Since the resistance of a vehicle
on a chassis dynamometer is much different from being on the road, the aim is to let these
differences be compensated by the dynamometer setting. There are two sets of road load
93
coefficients specified (these are the coefficients that describe the second order polynomial
curve):
a) Target coefficients: road load that was determined on the road
b) Set coefficients: load that is set on the chassis dynamometer
The difference between these two loads is mainly caused by internal friction in the chassis
dynamometer, the different contact of wheels on rollers, and the absence of aerodynamic
drag.
The result of the chassis dynamometer setting is a second order polynomial, which
represents the difference between the target road load (f0,f1 and f2) and the losses of the
vehicle on the chassis dynamometer. Effectively the dynamometer will simulate the
difference compared to the on-road losses of the vehicle.
There are 2 different methods allowed in the GTR for the setting of the chassis
dynamometer, see Table 13.
Chassis dynamometer setting method Reference to

method
iterative The vehicle is accelerated under its own power. Coast General:
method down on and adjustment of the chassis dynamometer is paragraphs 7. and
repeated until a tolerance of 10 N on 2 consecutive coast 8.
down runs is met (after regression).
Specifically:
8.1.3.4.2.
- as an alternative a new (shorter) warm-up cycle may be
used when evidence on the equivalency to a WLTC
warm-up is provided; see paragraph 7.3.4.3.
fixed run The vehicle is accelerated by its own power, or by the General:
method chassis dynamometer. Executed by a software program, paragraphs 7. and
the dynamometer will perform 3 coast downs after a first 8.
stabilization and one dynamometer setting coastdown
Specifically:
run. The set coefficients are derived from the average of
8.1.3.4.1.
the 3 coast downs, and no tolerance is applied.
Table 13: Chassis dynamometer setting methods and alternatives in the GTR
If the road load determination was done by the torque meter method, identical torque meters
will be installed on the vehicle, and the settings are iteratively adjusted until the difference
between simulated and measured load satisfies a tolerance of ±10 N×r’ from the target
running resistance at every speed reference point.43 After the chassis dynamometer setting,
the running resistance is transformed into road load coefficients by a coastdown of the
vehicle on the chassis dynamometer, unless the vehicle is not suitable for coasting down.
This procedure is described in par. 8.2.4 of Annex 4.
There are 2 appendices to Annex 4:
43
r’ is the dynamic radius of the tyre in metres on the chassis dynamometer obtained at 80 km/h
94
Appendix 1: the process of a performing a coastdown on the chassis dynamometer, and how
to convert the measured road load forces at reference speeds into a simulated road load
curve (constants for the second order polynomial).
Appendix 2: the process of adjusting the chassis dynamometer load setting to match the
simulated road load to the target road load, separately for the coastdown method and the
torque-meter method (determination of the proper ‘set coefficients’).
4.5.3 Annex 5 – Test equipment and calibrations

In this annex the requirements for the test equipment, the measurement and analysis
equipment, calibration intervals and procedures, reference gases, and additional sampling
and analysis methods are specified. During phase 1b a critical review on the test equipment
and calibrations was performed. Clarifications concerning additional sampling and
measurement methods were included where necessary.
The test equipment requirements cover the cooling fan and the chassis dynamometer. The
cooling fan requirements specify performance, dimensions and number and location of
measurement points for the check of the performance. The position of the fan with respect to
the front of the vehicle was made more robust. The chassis dynamometer requirements are
based on existing regulations but are supplemented by requirements for vehicles to be tested
in four wheel drive (4WD) mode. The accuracy requirements of difference in speed and
distance covered within a test between the front and rear rollers were reviewed and
confirmed during phase 1b. The chassis dynamometer calibration concerns the force
measurement system, parasitic losses and the verification of road load simulation.
The measurement and analysis equipment requirements cover the exhaust gas dilution
system, the emissions measurement equipment and the necessary calibration intervals and
procedures.
A full-flow exhaust dilution system is required for emission testing. This requires that the total
vehicle exhaust be continuously diluted with ambient air under controlled conditions using a
constant volume sampler. A critical flow venturi (CFV) or multiple critical flow venturis
arranged in parallel, a positive displacement pump (PDP), a subsonic venturi (SSV), or an
ultrasonic flow meter (USM) may be used. The exhaust dilution system consists of a
connecting tube, a mixing chamber and dilution tunnel, dilution air conditioning, a suction
device and a flow measurement device.
Specific requirements are given for the connection to the vehicle exhaust, the dilution air
conditioning, the dilution tunnel, the suction device and the volume measurement in the
primary dilution system. Recommended systems are exemplarily described.
These requirements are followed by the specifications of the CVS calibration and the system
verification procedures.
The requirements for the emission measurement equipment include gaseous emission
measurement equipment, particulate mass and particulate number emission measurement
equipment. They start with system overviews and end with descriptions of recommended
systems.
The calibration intervals and procedures cover instrument calibration intervals as well as
environmental data calibration intervals and analyser calibration procedures.
In addition, Annex 5 describes several methods to measure non-limited gaseous exhaust
species. The methods include laser spectrometry and Fourier transform infrared (FTIR) to
measure NH3, gas chromatography to measure N2O and methods for ethanol, formaldehyde
and acetaldehyde.
95
4.5.4 Annex 6 – Type 1 test procedure and test conditions
This Annex describes the execution of the testing activities to verify emissions of gaseous
compounds (including CO2), particulate matter, particle number, and fuel consumption over
the Type 1 test, using the WLTC applicable to the vehicle family. The scope of Annex 6 is
restricted to internal combustion engine vehicles (ICE). Electrified vehicles, i.e. having a
battery used for driving the vehicle, are tested according to the procedure in Annex 8.
General requirements
Testing is done in a conditioned environment on a chassis dynamometer. Diluted exhaust
emissions are continuously diluted with ambient air by a constant volume sampler (CVS) and
a proportional sample of exhaust gas collected for analysis. Background concentrations in
dilution air are measured simultaneously for all emission compounds, as well as particulate
mass and number, to correct the measurement results.
The temperature in the test cell has a setpoint of 296K with a tolerance of ±5 K during
testing, at the start of the test it should be within ±3 K. The setpoint for the soak area is the
same with a tolerance of ±3 K. In all cases, the temperature may not show a systematic
deviation from the setpoint.
Test vehicle
For the emission test (‘Type 1’) at the chassis dynamometer the road load of vehicle H,
which was determined according to Annex 4, has to be set. If at the request of the
manufacturer the interpolation method is used on CO2 (see paragraph 4.4.1 of this report),
an additional Type 1 test is performed with the road load as determined at test vehicle L.
However, the CO2 interpolation method may only be applied on those road load relevant
characteristics that were chosen to be different between test vehicle L and test vehicle H. For
example, if both test vehicle L and H are fitted with the same tyres, no interpolation is
allowed for the rolling resistance coefficient. Refer to paragraph 4.4.3 of this report for the
allowed interpolation/extrapolation range.
Please note that this interpolation method only applies to the group of vehicles that fall into
the same ‘interpolation family’, whose criteria are specified by par. 5.6 in part II of the UN-
GTR. These criteria have been chosen in such a way that the emission and fuel consumption
behaviour of vehicles in the interpolation family are likely to be similar, e.g. same engine,
same transmission type and model, same operating strategies, etc.
The vehicle is placed on the chassis dynamometer, and if it is equipped with a ‘dynamometer
operation mode’ and/or a ‘vehicle coastdown mode’, these modes have to be activated for
the respective procedure (refer to paragraph 4.4.5 of this report). Auxiliaries such as an
airconditioning system and radio are switched off during the test.
The tyres fitted on the test vehicle should be of a type specified as original equipment by the
manufacturer, but it is allowed to increase the tyre pressure by a maximum of 50 per cent
above the specified tyre pressure. Since any differences in rolling resistance are implicitly
corrected by the chassis dynamometer setting, this will not affect the accuracy of the road
load, as long as the same pressure is used throughout the tests.
Vehicle preconditioning
The chassis dynamometer is set in accordance with the procedure described in Annex 4. For
reasons of reproducibility, the battery will be fully charged. To precondition the vehicle and
the battery, the applicable WLTC will be driven (preconditioning cycle). Additional
preconditioning cycles may be driven at the request of the responsible authority or the
manufacturer, to bring the vehicle and its control systems to a stabilized condition. For
96
example, if the vehicle is equipped with an automatic gearbox that slowly adapts to the
driving behaviour, multiple preconditioning cycles could be needed to let the algorithm of the
shifting strategy adapt to the WLTC. After preconditioning and before testing, the vehicle is
soaked for a minimum of 6 hours to a maximum of 36 hours in a conditioned environment
(soak area setpoint of 296 K ± 3 K) until the engine oil temperature and coolant temperature
are within ± 2 K of the setpoint.
Transmissions
For manual transmissions, the gear shift prescriptions according to Annex 2 have to be
fulfilled within a tolerance on the point of shifting of ± 1 second. If the vehicle is unable to
follow the speed trace it has to be operated with the accelerator control fully activated.
Vehicles with an automatic-shift or multi-mode gearbox have to be tested in the ‘predominant
mode’, but only if such a predominant mode is present and is agreed by the responsible
authority to fulfil the requirements of 3.5.10 in part II of the GTR. The results in predominant
mode are used to determine fuel consumption and CO2 emissions.
It should be avoided that the vehicle would automatically shift itself to another mode as the
predominant mode, as this could open the way for misuse. Therefore a requirement was
added to state that ‘a single mode that is always selected when the vehicle is switched on
regardless of the operating mode selected when the vehicle was previously shut down’.
If the vehicle has no predominant mode or the requested predominant mode is not agreed by
the responsible authority as a predominant mode, the vehicle shall be tested in the best case
mode and worst case mode for criteria emissions, CO2 emissions, and fuel consumption. The
results of best- and worst-case mode are averaged to determine fuel consumption and CO2
emissions.
Even if there is a predominant mode available, the vehicle still has to fulfil the limits of criteria
emissions in all forward driving modes, except for modes that are used for special limited
purposes (e.g. maintenance mode, crawler mode).
Type 1 test
The testing can start after the vehicle has been properly soaked (see ‘vehicle
preconditioning’). The vehicle is moved from the soak area to the test room, and placed on
the chassis dynamometer. All the necessary equipment for emission measurement,
particulate filter and particle sampling is prepared and/or calibrated prior to the test. The
vehicle is started, and the applicable WLTC is driven while the speed is kept within the
indicated speed trace tolerances - refer to paragraph 1.2.6.6 of Annex 6 for detailed speed
trace tolerances. Except for particulate filter sampling, all measurements of compounds have
to be available for each of the individual cycle phases (Low, Medium, High and Extra-High),
in order to accommodate regional weighting by the Contracting Parties. Particulate sampling
is done on one filter for the whole cycle or –again for regional weighting purposes – on one
filter over the first three phases, and one separate filter for the fourth phase.
Post-test procedures
Just prior to the analysis, the analyzers will be calibrated as prescribed. On completion of the
cycle phases, the bags containing the diluted exhaust gases will be analyzed as soon as
possible, in any event not later than 30 minutes after the end of the cycle phase. The
particulate filter is transferred to the stabilization room within one hour after completing the
test.
97
Annex 6 has two appendices:
Appendix 1: Emissions test procedure for vehicles equipped with periodically regenerating
systems.
If the emission limits applied by the Contracting Party are exceeded during a cycle by the
regeneration of periodically regenerating emission reduction system(s), these emissions may
be calculated into a weighted average. This is done by the Ki factor, which defines how the
elevated levels of emission compounds during cycles where regeneration occurs are
attributed to the emission performance on cycles without regeneration. Basically, the
procedure for Ki determination takes into account the number of cycles without regeneration
and the emission performance on those cycles, and compares this to the one (or several)
cycles where regeneration occurs with the corresponding elevated emission levels. The Ki
can be applied as a multiplicative or an additive factor. The procedure also provides a Ki
calculation method for vehicles with more than one regenerating emission reduction system.
Appendix 2: Test procedure for electric power supply monitoring system
The monitoring of the charge/discharge energy of the battery in conventional ICE vehicles is
described. If the battery discharge energy over the cycle is above a set limit, the CO2 mass
emissions and fuel consumption have to be corrected via a formula with default values on
alternator accuracy and a Willans factor. This RCB correction procedure is explained in detail
in paragraph 4.4.16 of this report.
4.5.5 Annex 7 – Calculations

In this annex the procedures are described to calculate the results from all the data collected
from the Type 1 tests, and to make the necessary corrections. The calculations that are
specifically related to electrified vehicles are not included in here; these can be found in
Annex 8.
First the diluted exhaust gas volume is determined and corrected towards standard
conditions. In the next step the mass emissions of all the monitored gaseous compounds are
calculated from the measured concentrations in the bags. These are corrected by the
concentrations already present in the dilution air. The final result is presented as mass
emissions in g/km for each of the cycle phases (Low, Medium, High and Extra-High).
The calculation procedure of the interpolation method to determine vehicle specific CO2
emissions and fuel consumption for individual vehicles within the CO2 vehicle family is also
included in Annex 7. A detailed overview of this calculation procedure is given in paragraph
4.4.1 of this report. As the interpolation method uses the energy demand over the cycle as
an input, a separate calculation method is included for this in paragraph 5 of Annex 7.
The remaining procedures in Annex 7 describe the calculation process to derive the mass
emission in mg/km of particulates from the collected mass on the filter, and the particle
number emissions in particles per km.
Based on the calculated emissions for CO2, HC and CO and test fuel properties, the fuel
consumption is calculated for each of the cycle phases and for the complete test. This is
included in paragraph 6 of Annex 7. For more information on the fuel consumption
calculations refer to paragraph 3.4.5.6 of this report.
98
4.5.6 Annex 8 - Pure electric, hybrid electric and fuel cell hybrid vehicles
This annex is dedicated to pure electric (PEV), hybrid electric (NOVC-HEV, OVC-HEV) and
compressed hydrogen fuel cell hybrid (NOVC-FCHV) vehicles, and is structured into the
following paragraphs, which will be briefly summarized:
1. General requirements
This sets the requirements of the test procedures for pure electric, hybrid electric and
compressed hydrogen fuel cell hybrid vehicles. It is pointed out that for vehicles tested
under Annex 8 the RCB correction procedure according to Appendix 2 of Annex 8 is
applied, as well as Appendix 3 of Annex 8 for the measurement of REESS current and
voltage. For conventional ICE vehicles the RCB correction procedure according to
Appendix 2 of Annex 6 is applicable. See also paragraphs 4.4.16 and 4.4.18 of this
report
Unless stated otherwise in Annex 8, all requirements of Annex 6 also apply to vehicles
tested according to Annex 8.
All Annex 8 requirements shall apply to vehicles with and without driver-selectable
modes, if not stated otherwise.
1.1. Units, accuracy and resolution of electric parameters
This prescribes the units used for the electric parameters, as well as the accuracy and
resolution requirements the measurement system has to fulfil.
1.2. Emission and fuel consumption testing
For vehicles tested according to Annex 8, the same measurement requirements have to
be fulfilled as for conventional ICE vehicles.
1.3. Units and precision of final test results
This sets the precision requirements for the final test result values and states that for the
purpose of calculation the unrounded values shall be used.
1.4. Vehicle classification
This specifies that all Annex 8 vehicles are classified as Class 3 vehicles and therefore
the WLTC Class 3a or 3b driving curve is the reference cycle (depending on their
maximum speed). Due to the downscaling procedure for PEVs and the capped speed
cycle modification for all Annex 8 vehicles, the applicable test cycle may differ from the
reference cycle.
1.5. OVC-HEVs, NOVC-HEVs and PEVs with manual transmissions
The vehicles shall be driven according to the manufacturer’s instructions, as
incorporated in the manufacturer's handbook of production vehicles, and as indicated by
a technical gear shift instrument.

2. REESS and fuel cell system preparation
This paragraph defines the run-in of the test vehicle in advance of the WTLP test
procedure.
3. Test procedure
3.1 General requirements
99
The applicable test cycles and requirements for the preparation of the test are described.
If the vehicle cannot follow the trace, the acceleration control shall be fully activated until
the required speed trace is reached again. Power to mass calculation and classification
methods shall not apply to these vehicle types (see also paragraph 1.4).
3.2 Test procedure for OVC-HEV
Requirements for the testing of OVC-HEV under WLTP conditions are specified,
including:
a) the operating conditions for both charge-depleting Type 1 test and charge-sustaining
Type 1 test procedure,
b) the preconditioning procedure,
c) soak procedure of the vehicle,
d) setting of the driver-selectable mode, both in charge-depleting and charge-sustaining
operating condition, and
e) end of test criteria (break-off criterion).
Charge-depleting Type 1 tests and Charge-Sustaining Type 1 tests may be driven
independent from each other but may also be combined (see Figure A8/1 in Annex 8).
3.3 Test procedure for NOVC-HEV

Requirements for the testing of NOVC-HEV under WLTP conditions are specified,
including:
a) the operating conditions for the Type 1 test procedure,
c) soak procedure of the vehicle, and
d) setting of the driver-selectable mode for the vehicle.
3.4 Test procedure for PEV

Requirements for the testing of PEV under WLTP conditions, including:
a) the applicable test procedure and its operating conditions,
c) soaking of the vehicle,
d) setting of the driver-selectable mode for the vehicle, and
e) end of test criteria (break-off criterion).
For PEVs with a higher range, a shortened test procedure (STP) is applied, from which
the electric range is calculated - see paragraph 4.4.19 of this report.
The electric range of OVC-HEVs is determined for the whole WLTC as well as for the
city cycle consisting of the low and medium phases only
3.5 Test procedure for NOVC-FCHV

Requirements for the test procedure of NOVC-FCHV under WLTP conditions, including:
a) the operating conditions for the Type 1 test procedure
c) soaking of the vehicle, and
100
d) setting of the driver selectable mode for the vehicle.
4. Calculations
This paragraph specifies the calculations of the test results, including gaseous emission
compounds, particulate matter emission and particle number emission, CO2 mass
emission, fuel consumption, electric energy consumption and range.
Gaseous emission compounds, particulate matter emission and particle number

emission
For NOVC-HEVs and OVC-HEVs, gaseous emission compounds, particulate matter
emission and particle number emission shall be calculated by the same requirements as
for conventional ICE vehicles according to Annex 7.
In addition, a calculation method for OVC-HEVs is applied to weigh the mass emissions
of gaseous compounds, particulate matter emission and particle number emission of the
charge sustaining and the charge depleting test according to the utility factor.
CO2 mass emission

For NOVC-HEVs and for OVC-HEVs under charge-sustaining operating condition, the
calculation procedures for the CO2 mass emission of the whole cycle but also for each
individual cycle phase are included. Where necessary, these results are corrected
towards a zero charge balance of the REESS according to Appendix 2 of Annex 8.
In addition, a calculation method for OVC-HEVs is applied to weigh the CO2 emissions of
the charge sustaining and the charge depleting test according to the utility factor.
Fuel consumption
For OVC-HEVs under charge-sustaining operating conditions, NOVC-HEVs and NOVC-
FCHVs, the fuel consumption will not be measured directly, but determined from the
gaseous emission compounds by the described post processing procedure for the
charge-sustaining values – see Table A8/5, /6 and /7 of Annex 8.
Charge-depleting as well as utility factor-weighted fuel consumption values are
calculated and determined by the calculation methods provided.
Electric energy consumption

For PEVs and OVC-HEVs, the determination of the electric energy consumption is
described. The electric energy consumption is determined for the whole cycle as well as
for each individual phase. Basis for the measured energy consumption is the measured
recharged electric energy from the mains, so as to include the charging losses.
For OVC-HEVs, there are also calculation methods provided for the utility factor-
weighted as well as the charge-depleting electric energy consumption.
Range
For PEVs, an electric range is determined which is referred to a the ‘pure electric range’
(PER). This range has to be provided for the whole cycle as well as for each individual
phase. This is calculated from the usable battery energy and the average energy
consumption over the cycle or phase.
101
For OVC-HEVs, there are three ranges to be determined:
a) All electric range (AER): the distance driven up to the first engine start.
b) Actual charge-depleting range (RCDA): the distance driven to the point where it was
not in a charge-depleting operating condition anymore and had entered into a
charge-sustaining operating condition.
c) Equivalent all electric range (EAER): the portion of the RCDA which was driven
electrically
The AER range has to be determined both for the whole WLTC and for the WLTCcity
cycle.
The EAER has to be determined for the whole WLTC, for the WLTCcity cycle and for
each individual cycle phase.
The RCDA has only to be determined for the whole cycle.
Interpolation of parameters for individual vehicles

Paragraph 4.5. of Annex 8 describes the interpolation method to calculate the values for
individual vehicles between vehicle H and vehicle L.
The basic concept of the interpolation approach is the same as that of conventional
vehicles but due to the interaction of the electric power train and conventional power
train (depending on the vehicle’s operation strategy) as well as the calculation schemes
to arrive at the output values, additional requirements have to be fulfilled. During phase
1b of WLTP this was investigated and evaluated for all of the Annex 8 output values.
Result of this investigation and evaluation was that for some values, the linearity
between vehicle H and vehicle L cannot be ensured in each and every case without
additional requirements. The required conditions to apply the interpolation approach are
further specified in this chapter.
One example on the case of NOVC- HEVs and OVC-HEV is the allowed CO2 mass
emission difference between vehicle H and vehicle L in charge-sustaining condition. This
range is limited to 20 g/km if only a vehicle H and L is measured and can be extended to
30 g/km if an additional vehicle M is measured.
Further requirements to complement the main body of Annex 8 is provided in the following
appendices:
Appendix 1 - REESS state of charge profile

This appendix is a visualisation of the different Type 1 test procedures for OVC-HEVs,
NOVC-HEVs, NOVC-FCHVs and PEVs. It contains figures showing example SOC profiles
for charge-depleting and/or charge-sustaining tests.
Appendix 2 - REESS energy change-based correction procedure

This appendix describes the procedure to determine the CO2 correction coefficient, which is
needed if a correction of the charge-sustaining Type 1 test CO2 mass emission for NOVC-
HEVs and OVC-HEVs is required. The correction procedure is mandatory for the
determination of the determination of the phase specific values. See also paragraph 4.4.18 of
this report
Also included is a correction procedure for NOVC-FCHVs with the determination of a fuel
correction coefficient as a function of the electric energy change of all REESSs.
102
Appendix 3 - Determination of REESS current and REESS voltage for NOVC-HEVs, OVC-
HEVs, PEVs and NOVC-FCHVs
This Appendix describes the required instrumentation and measurement methods to
determine the REESS current and the REESS voltage of NOVC-HEVs, OVC-HEVs, PEVs
and NOVC-FCHVs.
Appendix 4 - Preconditioning, soaking and REESS charging conditions of PEVs and OVC-
HEVs
This Appendix defines the procedure for the REESS and combustion engine preconditioning
in preparation of the test as well as the charging procedure of the REESS.
Appendix 5 - Utility factors (UF) for OVC-HEVs

This Appendix describes the formula and the coefficients of the regional UFs. Each
Contracting Party may develop its own UFs, but is recommended to apply the procedure of
SAE J2841. See also paragraph 3.4.5.8 of this report.
Appendix 6 - Selection of driver-selectable modes

This Appendix describes which mode should be selected for the Type 1 test procedure, for
which flowcharts are included. The prioritisation concerning the mode selection is as follows:
1. First priority is being able to follow the applicable driving cycle
2. Second priority is choosing the predominant mode.
In case of OVC-HEVs, the mode selection has to be evaluated for both charge-depleting and
charge-sustaining operation conditions. See also paragraph 3.4.5.10 and Appendix 1 of this
report.
Appendix 7 - Fuel consumption measurement of compressed hydrogen fuel cell hybrid

vehicles
This Appendix is describes the method to measure the fuel consumption of NOVC-FCHV.
See also paragraph 4.4.23 of this report.
103
4.5.7 Annex
A 9 – System
m equivale
ency
Other measureme
m nt methodss can be ussed for testting if they yield equivvalent resultts to the
testing methods de escribed in the GTR. To
T prove syystem equivvalency, the e accuracy and the
on of the ca
precisio andidate me ethod has to
o be equal or better thhan the refeerence methhod, and
this evid
dence will have
h to be based
b on sta
atistical datta.
w the differe
To show ence betwee
en accuracy
y and precission, please
e refer to Fiigure 31.
Figure 31: Differe

ence betwe acy and prrecision with respect to a refe
een accura erence
value
Guidancce for the correlation

c b
between an existing an
nd a candid d is provided in ISO
date method
5725 Pa
art 6 Annexx 8.
plementatio
The imp on of system
m equivalen ailed in a phase 2 of the
ncy will be further deta t GTR
develop
pment.
104
5 Validation of the test procedure
Within the WLTP development programme two validation steps were executed. The first
validation phase aimed at the assessment of the driveability of the WLTP cycles, these
results are included in the Technical Report on the development of the harmonised driving
cycle (DHC)44. This chapter will give an overview of the activities that were done in the
Validation phase 2, which was dedicated to test and validate the new elements in the test
procedure. All of the validation tests and analyses were executed during phase 1a timeframe
of WLTP. Apart from the Round Robin tests (see paragraph 3.4.4) there were no further
testing or validation activities included in phase 1b.
5.1 Validation Tests

5.1.1 Participants and vehicles, measured parameter
Validation phase 2 was executed between April 2012 and December 2012. All necessary
information concerning:
 Test plan,
 Parameter list and test procedure,
 Test sequences,
 Driving cycle schedules,
 Gearshift prescriptions for manual transmission vehicles,
 Data collection and delivery
were made available to the participants via JRC’s FTP-server.
For class 1 and class 2 vehicles the cycle version 1.4 was used, for class 3 vehicles the
cycle version 5 was applied. At the beginning of the validation 2 phase the gearshift
calculation tool version dated 16.04.2012 was used.
Some modifications on procedural issues needed to be performed during the validation 2
phase, based on the analysis of the results obtained so far. Table 14 gives an overview of
these modifications.
The most important modifications were made by the VP2 information package from 25 July
2012. For class 1 and class 2 vehicles the cycle versions 1.4 were replaced by cycle
versions 2 and the gearshift calculation tool from 16.04.2012 was replaced by the version
from 09.07.2012.
44
See document GRPE-68-03 http://www.unece.org/trans/main/wp29/wp29wgs/wp29grpe/grpeinf68.html
105
No. Date Filename Modification
1 19 April 2012 File_2 - Item 21:

Parameter_List_for_Validation_2_v7_
DTP_19-April-2012.xlsx Proportional fan
2 23 April 2012 File_1 - Validation2 Test Plan_23-April- Addition of TNO as Participating

2012.xls Lab (in box L5 and in Evaluation
Item “ICE Vehicle weight”)
3 23 April 2012 File_8 - WLTP_VP2_Participating Update of the List of Participating

Labs_list_23-April-2012.docx Labs (TNO – The Netherlands)
4 26 April 2012 File_6 - Data_collection_template_26- Addition of columns (related to

April-2012.xls adopted Gear Shift strategy) to
the “bag results test i *” pages
5 15 May 2012 File_DHC_B_ANNEX_15-May- New file - Addition of a “.doc” file

2012.doc with detailed instructions on how
to use the Gear Shift Evaluation
Tool
6 15 May 2012 File_3 - LabProc-EV-TestMatrix_from New file - Addition of the Test

ACEA_15-May-2012.xlsx Matrix for EV/HEV
7 15 May 2012 File_0 - Read me_15-May-2012.docx “Read me” file updated
8 09 July 2012 File_DHC_A - Driving Cycles_09-July- New version of Class 1 and Class
2012.xlsx 2 driving cycles
9 09 July 2012 File_DHC_B_gearshift_calculation_tool Gear Shift calculation tool updated

_09-July-2012.mdb and streamlined
10 09 July 2012 File_DHC_B_ANNEX_09-July- Revised explanatory note on how

2012.doc to use the Gear Shift calculation
tool
11 23 July 2012 File_8 - WLTP_VP2_Participating File updated

Labs_list_23-July-2012.docx
12 23 July 2012 File_9 - JRC_ftp_server_Owners_23- File updated

July-2012.xlsx
File_6.1 - New version of the excel template

Data_collection_template_lab_and_ve to report test results. The original
hicle_info_25-July-2012.xls file has been split in two files, now
13 25 July 2012 including also EV/HEV and
File_6.2 - PM/PN features
Data_collection_template_test_results
_25-July-2012.xls
14 25 July 2012 File_0 - Read me_25-July-2012.docx File updated
Table 14: Procedural modifications during the validation 2 phase
In total, 34 different laboratories, institutions and manufacturers participated in the validation

phase 2.
106
The results were delivered to the JRC server and then collected in an Access database. A
total number of 109 vehicles were tested in the validation phase 2. These can be categorised
into subgroups as shown in Table 15.
Vehicle subcategory number
Battery electric vehicle 6
Hybrid electric vehicle with Petrol ICE 3
Hybrid electric vehicle with Diesel ICE 1
Plug in hybrid electric vehicle with Petrol ICE 2
M1, class 1, Diesel 2
M1, class 1, NG 1
N1, class 1, Diesel 5
M1, class 2, Petrol 2
M1, class 3, NG/LPG 6
M1, class 3, Petrol 40
N1, class 3, Diesel 4
N1, class 3, Petrol 2
N1, class 3, NG 1
Table 15: Overview of the validation 2 vehicle sample categories and numbers
Information about the chassis dynamometers was delivered by 33 of the 34 participating

laboratories. For 19 laboratories it was possible to measure all 4 phases of the WLTC in one
test, because their test benches had 4 bag measuring devices. The other laboratories had
only 3 bag measuring devices. Most of them measured the first 3 phases (L&M&H) with a
cold start and then phases L, M and exH in hot condition in a second test. Some participants
measured different phase combinations in addition to the base test.
For the larger part of the vehicles only the basic tests were performed. The base test
consists of the WLTP test with a cold start at the test mass of vehicle H (TMH). For 92% of
the ICE vehicles an additional hot start test was performed. It was foreseen to repeat all tests
at least twice, so that three results could be used to assess the repeatability. Some
participants did additional tests with parameter variations.
The following parameter variations were performed:
 Four filter (one per cycle phase) and one filter tests (for all phases) for particulate
mass (vehicles 1 and 3),
 Gearshifts according to GSI and the calculation tool (vehicles 4, 5, 8, 10 and 102),
 Test mass and/or road load variations (16 vehicles, from 2 variants up to 4 variants),
 Different preconditioning tests (vehicles 19 and 43),
 Overnight soak with forced cooling (vehicles 43, 44, 53, 61, 67, 68, 69 and 70)
For the pure electric vehicles charge depleting tests were performed, in some cases with
different cycles or phase combinations.
An overview of the different cycle combinations and number of tests performed is given in the
following tables.
Table 16 shows the cycle allocation for PEV’s and hybrids. All hybrids and 4 of the 6 PEV’s
were tested with the class 3 cycles. Although its maximum speed was 145 km/h, vehicle 58
was classified as class 2 vehicle because the power to mass ratio was below 34 kW/t, if the
107
‘30 minutes power’ is considered as rated power. Consequently this vehicle was tested with
the class 2 cycles.
Vehicle 84 had a 30 minutes power of 28 kW. Using this value the vehicle was classified as
class 1 vehicle, although the maximum speed was 130 km/h. Consequently this vehicle was
tested first with the class 1 cycles. But since the discussions about the classification of PEV’s
was already ongoing at that time, additional tests were performed with the class 2 and class
3 cycles.
The EV subgroup finally concluded that a power-to-mass ratio determination is not yet
feasible for PEV’s due to the absence of a robust system power definition. Therefore it was
decided that all PEV’s are tested at class 3 cycles.
All class 1 and class 2 vehicles with ICE are from India. Table 17 shows that 5 of the 8 class
1 vehicles were tested with both cycle phases (low and medium), the remaining 3 were
tested with the low phase only, because the maximum speed was below 70 km/h.
All class 2 vehicles were tested with the class 2 cycle but without the extra high speed phase
(see Table 18).
All M1 class 3 vehicles were tested at all 4 cycle phases (see Table 19 and Table 20), while
1 of the 7 N1 class 3 vehicles was tested without the extra high speed phase (see Table 21).
Number of tests
WLTC, C 2, WLTC, C 2, WLTC, C 3, WLTC, C 3,
WLTC, C 1, WLTC, C 1, WLTC, C 2, V 2, WLTC, C 3, WLTC, C 3, V 5,
Veh_Cat engine_type IDveh V 1_4, V 1_4, V 5, V 5,
V 2, L&M V 2, L&M&L L&M&H&exH V 5, L&M L&M&H&exH
L&M L&M&H L&M&H L&M&H&L
BEV EM 58 70 36
BEV EM 59 48 12 30
BEV EM 77 5
BEV EM 80 8 12
BEV EM 84 50 37 6 10
BEV EM 108 43 12
PHEV Petrol OVC 60 22 35
PHEV Petrol OVC 65 4
HEV, class 3 Diesel, NOVC 104 3
HEV, class 3 Petrol NOVC 9 13
HEV, class 3 Petrol NOVC 78 2 2
HEV, class 3 Petrol NOVC 85 9
Table 16: Overview of tests for pure electric and hybrid electric vehicles
WLTC, C 1,
WLTC, C 1,
Veh_Cat engine_type IDveh V 2,
V 2, L&M&L
L&L&L
M1, class 1 DIESEL 87 6
M1, class 1 Diesel 101 6
M1, class 1 NG 86 6
N1, class 1 Diesel 89 6
Table 17: Overview of tests for class 1 ICE vehicles
108
WLTC, C 2,
WLTC, C 3, V 5,
Veh_Cat engine_type IDveh V 2,
L&M&H&exH
L&M&H
M1, class 2 DIESEL 88 6
M1, class 2 Petrol 35 6
N1, class 2 NG 2 12
Table 18: Overview of tests for class 2 ICE vehicles
WLTC, C 3, WLTC, C 3,
WLTC, C 3, WLTC, C 3, WLTC, C 3, V 5, V 5_1,
Veh_Cat engine_type IDveh V 5, V 5,
V 5, L V 5, L&L V 5, L&M L&M&H& L&M&H&
L&M&exH L&M&H
exH exH
M1, class 3 Diesel 82 2 4 17 27
M1, class 3 Diesel 83 4 10 16
M1, Class 3 DIESEL 94 3 3
M1, class 3 Diesel 102 2 12 14
M1, class 3 Diesel 14 3 3
M1, class 3 DIESEL 30 3 3
M1, class 3 DIESEL 31 3 3
M1, class 3 diesel 42 12
M1, class 3 diesel 61 18
Table 19: Overview of tests for class 3 M1 vehicles (Diesel ICE)
109
WLTC, C 2, WLTC, C 3, WLTC, C 3, V WLTC, C 3,
WLTC, C 3, WLTC, C 3, V 5,
Veh_Cat engine_type IDveh V 2, V 5, 5, V 5_1,
V 5, L&M L&M&H
L&M&H L&M&exH L&M&H&exH L&M&H&exH
M1, class 3 LPG 55 3 3
M1, class 3 NG 25 3 3
M1, class 3 NG 36 3 3
M1, class 3 NG 37 3 3
M1, class 3 NG 7 6
M1, class 3 NG 50 6
M1, class 3 Petrol 97 1 1
M1, Class 3 Petrol 99 3
Table 20: Overview of tests for class 3 M1 vehicles (NG or Petrol ICE)
110
WLTC, C 3, WLTC, C 3, V 5, WLTC, C 3, V
Veh_Cat engine_type IDveh V 5, V 5,
V 5, L&M L&M&H&exH 5, L&M&L
L&M&exH L&M&H
N1, class 3 Diesel 103 2 2

N1, class 3 Petrol 69 3 4
N1, class 3 Petrol 70 4 5
Table 21: Overview of tests for class 3 N1 vehicles
5.1.2 Evaluation issues

The following evaluation issues were discussed within the DTP subgroups:
 Soak Temperature Tolerances
 Soak with forced Cooling down
 Test Cell Temperatures
 Tolerances of Humidity during Test Cycle
 Tolerances of Emission Measurement System
 Preconditioning Cycle
 Preconditioning for Dilution Tunnel
 Speed Trace Tolerances
 Gearshift tolerances for manual transmission vehicles
 Monitoring of RCB of all Batteries
 Cycle Mode Construction
 Required Time for Bag Analysis
 Dilution Factor
 Dyno Operation Mode
Out of this longlist, the following issues will be discussed in this report based on the
validation phase 2 results:
 Overnight soak temperature,
 Test cell temperature and humidity,
 Speed trace violations,
 Charge depleting tests for PEV and OVC HEV
Other issues are not mentioned in detail here, such as the test mass influence, because the
tests results did not provide evidence that there was a need to modify the GTR on those
issues. The differences between the results for manual transmission vehicles with gearshifts
according to the on board GSI and the WLTP calculation tool were rather small and did not
show any trends.
111
5.2 Results
5.2.1 Overnight soak temperatures
The validation 2 results database contains temperature monitoring for 274 different overnight
soaks without, and 15 soaks with accelerated cooling. Figure 32 shows an example for
coolant and air temperature monitoring of 7 different tests on the same vehicle.
An extensive evaluation of the results led to the following specifications in the GTR:
“The soak area shall have a temperature set point of 23 °C and the tolerance of the actual
value shall be within ± 3 °C on a 5 minute running arithmetic average and shall not show a
systematic deviation from the set point. The temperature shall be measured continuously at a
minimum frequency of 1 Hz.”
70
air temp, test 1 air temp, test 3
60
air temp, test 8 coolant temp, test 1
coolant temp, test 3 coolant temp, test 4
50
temperature in °C
40
overnight soak temperatures, vehicle 4
30
20
10
0
0 5 10 15 20 25
time in h
Figure 32: Example of overnight soak temperature monitoring
5.2.2 Test cell temperatures

The next validation point was the variation of the test cell temperature during the tests. The
class 3 cycle was used for the evaluation. Figure 33 shows the time history of the test cell
temperature with the lowest variation, Figure 34 shows the case with the highest variation.
The variation ranges for all tests are shown in Figure 35.
Based on these results the following requirements were drafted for the GTR:
“The test cell shall have a temperature set point of 23°C. The tolerance of the actual value
shall be within ± 5 °C. The air temperature and humidity shall be measured at the vehicle
cooling fan outlet at a minimum frequency of 1 Hz.”
112
30 140
Idveh 3, test series 1, test_ID 15, start condition ID 1, WLTC class 3
120
25
100
20 amb_air_temp
vehicle speed in km/h
temperature in °C
oil_temp
80
15 v_act
60
10
40
5
20
0 0
0 300 600 900 1200 1500 1800
time in s
Figure 33: Best case of test cell temperature over all 4 phases of the class 3 WLTC
30 140
120
25
Idveh 15, test series 1, test_ID 3, start condition ID 1, WLTC class 3 100
20
temperature in °C
amb_air_temp
80
15 v_act
60
10
40
5
20
0 0
0 300 600 900 1200 1500 1800
time in s
Figure 34: Worst case of test cell temperature over all 4 phases of the class 3 WLTC
113
Figure 35: Test ce
ell tempera
ature variation range during
d clas
ss 3 WLTC,, all tests
5.2.3 Test
T cell humidity
y
Examples for the time history and the variances of test ce
ell humidity are shown
n in the
Figure 36 to Figure 38).
following figures (F
Based o
on these ressults the following requ
uirements were
w drafted
d for the GT
TR:
“The sp
pecific humiddity H of eitther the air in the test cell ntake air of the engine shall be
c or the in
such tha
at: 5.5 ≤ H ≤12.2
≤ (g H2O/kg dry air).
“Humidiity shall be measured continuousl
c mum of 1 Hz
y at a minim z.
Specificc humidity H shall be m
measurable to with a resolution of ±1 g H2O/kg
g dry air.”
114
60% 140
Idveh 1, test series 1, test_ID 1, start condition ID 1, WLTC class 3
120
50%
vehicle speed in km/h, amb_air_pressure
100
40% rel_humidity
relative humidity
v_act 80
30% amb_air_pressure kPa
60
20%
40
10%
20
0% 0
0 300 600 900 1200 1500 1800
time in s
Figure 36: Example for the time history of the test cell humidity over the class 3
WLTC
60%
50%
40%
relative humidity
vehicle 28, test cell humidity during WLTC
30%
20% test 1 test 2 test 3
10%
0%
0 300 600 time in s 900 1200 1500 1800
Figure 37: Examples for the time history of the test cell humidity over the class 3
WLTC
115
60%
50%
40%
relative humidity
30%
hum_ave ‐ 2*stddev humidity variance during tests
20%
hum_ave
hum_ave +2*stddev
10%
0%
1 2 3 7 8 9 10 11 12 13 15 16 17 18
vehicle number
Figure 38: Test cell humidity variances during the tests
5.2.4 Speed trace violations

The participants of the validation 2 phase delivered the time sequences of the measured
vehicle speed signal together with the set speed with 1 Hz resolution. The deviations of the
measured speed from the set speed were then calculated for all tests and
compliances/violations were calculated for the following tolerance bands:
 ±3 km/h, ±1 s,
 ±2 km/h, ±1 s,
Figure 39 shows an examples of the first 300 s of the speed traces of 6 tests for a
subcompact car with a power to mass ratio of 43.6 kW/t together with the set speed and the
tighter of the above listed tolerance bands (±2 km/h, ±1 s). No speed trace violations
occurred in either of these tests.
In most cases the drivers did not have problems to keep the actual speed within this
tolerance band. In some cases tolerance violations occurred due to lack of power (see Figure
40 and Figure 41).
Figure 40 shows the speed trace of the extra high speed part for a N1 vehicle with a petrol
engine retrofitted for CNG bi-fuel operation. Running on petrol, the rated power is 85 kW.
With a kerb mass of 2003 kg this leads to a power to mass ratio (pmr) of 42.4 kW/t, so that
this vehicle would be a class 3 vehicle, since the borderline between class 2 and class 3 is
34 kW/t.
When this vehicle was tested on natural gas, the rated power reduced to 68 kW, resulting in
a pmr value just below the borderline of 34 kW/t. The speed trace violations shown in Figure
40 would not occur if the vehicle had been tested on the class 2 cycle, since this cycle has
less demanding accelerations and a lower top speed.
116
140
v_min v_set v_max 2500101 2500102

120
2500201 2500202 2500301 2500302
100
veh 25, all tests
80
60
40
20
0
0 60 120 180 240 300
time in s
Figure 39: Example for speed trace and tolerance band for the class 3 WLTC
140
veh 2
120
100
80
v_min v_set v_max 2071 2072

60
40 2081 2082 2091 2092 2101
20 2102 2111 2112 2121 2122
0
1500 1560 1620 1680 1740 1800
time in s
Figure 40: Example for tolerance band violations for the extra high speed phase of
the class 3 WLTC (CNG fueled vehicle, pmr = 33.4 kW/t)
117
A more severe example is shown in Figure 41. This vehicle from India was tested on natural
gas, which obviously reduced the maximum power compared to the operation on petrol and
would therefore qualify as class 2 vehicle. In this particular case it would even not be able to
reach the top speed of the extra high speed phase of the class 2 cycle (123 km/h).
In addition to that, Figure 41 clearly shows that the driveability problems are not only related
to the top speed sections, but occur already around the cycle time of 1550 to 1560 s at a
vehicle speed of 80 km/h because the acceleration is too high.
A more detailed analysis of such driveability problems led to the downscaling method for low
powered vehicles, which is described in detail in the DHC report44.
Based on the results of the speed compliance/violation analysis the ±2 km/h at ±1 s tolerance
was concluded to be feasible, and was therefore implemented into the GTR.
Gearshifts did not cause driveability problems for manual transmission vehicles.
140
veh 37
120
100
80
60
v_min v_set v_max 37011 37012
40
20 37021 37022 37031 37032
0
1500 1560 1620 1680 1740 1800
time in s
Figure 41: Example for tolerance band violations for the extra high speed phase of
the class 3 WLTC (CNG fueled vehicle)
5.2.5 Charge depleting tests for PEV and OVC HEV

As already mentioned, charge-depleting tests were performed for 6 pure electric vehicles
(PEV) in the validation 2 exercise. Since it was not quite clear how to classify PEVs with
respect to vehicle classes, the cycle version allocation was interpreted differently by the
participating labs. One lab used the ‘30 minutes maximum power’ of the electrical motor and
classified the vehicles by calculating the power to (kerb) mass ratio based on that power
indicator.
This led to the situation that vehicle 58 with a kerb mass of 1860 kg and a peak power of 120
kW, but a 30 minutes power of only 60 kW, was classified as class 2 vehicle although its
maximum speed was 145 km/h. This vehicle could have easily driven the class 3 cycle, but
118
was only tested on the class 2 cycle in the version 1.4, which does not include an extra high
speed phase. With the 3 phases (Low, Medium and High) of the class 2 cycle the vehicle
was able to drive more than 250 km i.e. more than 17 cycles before the batteries were
depleted.
Two CD tests on this cycle were performed with vehicle 58. The cumulative discharge curves
are shown in Figure 42 and Figure 43. At first glance there seems to be a wide spread of the
energy consumption per cycle within a charge depleting test. For both tests the difference
between maximum and minimum discharged energy over one cycle is 0,6 Ah which
corresponds to 14% of the average (-6% to +8%) which is reasonably good.
However the break-off point (end of the charge depleting test) is significantly different in both
tests (for a more detailed overview see Figure 44, Figure 45 and Figure 46). This leads to a
difference in the range determination of about 9 km (253.5 km to 263.2 km) or 3,5% with
respect to the average range.
20,000 90 test 1‐1
test 1‐2
18,000 vehicle 58, class 2, version 1.4, L&M&H
80
test 1‐3
test 1‐4
16,000
70
test 1‐5
cumulative discharge energy in As
14,000 test 1‐6
60
test 1‐7
12,000 test 1‐8
50
test 1‐9
10,000
test 1‐10
40
test 1‐11
8,000
test 1‐12
30
6,000 test 1‐13
test 1‐14
20
4,000 test 1‐15
test 1‐16
2,000 10
test 1‐17
test 1‐18
0 0
0 300 600 900 1200 1500 v_set
time in s
Figure 42: Cumulative discharge energy during the CD test 1 for vehicle 58 on the
class 2 cycle (version 1.4)
119
20,000 90 test 2‐1
test 2‐2
18,000 vehicle 58, class 2, version 1.4, L&M&H
80
test 2‐3
test 2‐4
16,000
70
test 2‐5
cumulative discharge energy in As
14,000 test 2‐6
60
test 2‐7
12,000 test 2‐8
50
test 2‐9
10,000
test 2‐10
40
test 2‐11
8,000
test 2‐12
30
6,000 test 2‐13
test 2‐14
20
4,000 test 2‐15
test 2‐16
2,000 10
test 2‐17
test 2‐18
0 0
0 300 600 900 1200 1500 v_set
time in s
Figure 43: Cumulative discharge energy during CD test 2 for vehicle 58 on the class
2 cycle (version 1.4)
140
v_min v_set v_max test 1‐1 test 1‐2 test 1‐3 test 1‐4 test 1‐5
test 1‐6 test 1‐7 test 1‐8 test 1‐9 test 1‐10 test 1‐11 test 1‐12 test 1‐13

120

100 test 2‐12 test 2‐13 test 2‐14 test 2‐15 test 2‐16 test 2‐17 test 2‐18
80
veh 58, test series 1 & 2, WLTC,
class 2, version 1.4, L&M&H
60
40
20
0
0 300 600 900 1200 1500
time in s
Figure 44: Time series of the vehicle speed for CD tests 1 and 2 for vehicle 58
120
140
v_min v_set v_max test 1‐1 test 1‐2 test 1‐3 test 1‐4
test 1‐5 test 1‐6 test 1‐7 test 1‐8 test 1‐9 test 1‐10 test 1‐11

100
test 2‐15 test 2‐16 test 2‐17 test 2‐18
veh 58, test series 1 & 2,
80 WLTC, class 2, version 1.4,
L&M&H
60
40
20
0
600 660 720 780 840 900
time in s
Figure 45: Time series of the vehicle speed for CD test 1 for vehicle 58 at break-off
point
140
v_min v_set v_max test 1‐1 test 1‐2 test 1‐3 test 1‐4 test 1‐5

120

80
60
40
veh 58, test series 1 & 2, WLTC,
20 class 2, version 1.4, L&M&H
0
1200 1260 1320 1380 1440 1500
time in s
Figure 46: Time series of the vehicle speed for CD test 2 for vehicle 58 at break-off
point
121
The driver instruction for the end of a charge depleting test was as follows: If the vehicle
speed falls below the tolerance for a time of 4 seconds or more, the vehicle should be
brought to standstill within the following 15 seconds. As can be seen in Figure 45 and Figure
46, this instruction was not strictly followed. This was also the case for the other vehicles. On
the contrary, Figure 46 shows that the driver was aware that the batteries became fully
depleted but still tried to drive as long as possible with full power so that the actual speed
trace was significantly above the speed trace within a deceleration phase.
In any case, the charge depleting tests especially at the break-off sections were very helpful
for the definition of suitable break-off criteria for the GTR.
Vehicle 59 was also tested by the same lab. But since this vehicle had a 30 minutes
maximum power of 35 kW (55 kW peak power) and a kerb mass of 940 kg, it was classified
as class 3 vehicle (pmr > 34 kW/t). As a consequence it was tested on the class 3 cycle
although the maximum speed was only 124 km/h, which is 6 km/h below the maximum
speed of the cycle.
Another example for a PEV that was tested by the same lab is shown in Figure 47 (vehicle
84). This vehicle had a kerb mass of 1290 kg, a peak power of 56 kW and a 30 minutes
power of 28 kW. The vehicle was originally tested on the class 1 version 2 cycle because the
power to mass ratio is below 22 kW/t, when the 30 minutes power is used as rated power.
But since the vehicle had a maximum speed of 130 km/h, it was also tested on all 4 phases
of the class 2 version 2 cycle and on the first 3 phases (Low, Medium and High speed) of the
class 3 cycle. The 4th phase of the class 3 cycle was skipped, because the vehicle could was
not even able to reach the maximum speed of the extra high speed phase of the class 2
cycle. Figure 48 shows the break-off section for the class 3 cycle of this vehicle.
140
120
100
80 veh 84, test series 3, WLTC,
class 2, version 2,
L&M&H&exH
60
40
v_min v_set v_max 8400501 8400502
20
8400503 8400504 8400505 8400506
0
1500 1560 1620 1680 1740 1800
time in s
Figure 47: Time series of the vehicle speed for CD test 3 for vehicle 84 at break off
section
122
140
v_min v_set v_max 8400601 8400602 8400603 8400604
120
8400605 8400606 8400607 8400608 8400609 8400610

100
veh 84, test series 4, WLTC,
80
class 3, version 5, L&M&H
60
40
20
0
720 780 840 900 960 1020
time in s
Figure 48: Time series of the vehicle speed for CD test 4 for vehicle 84 at break off
section
The remaining PEV’s were all tested on the class 3 cycle.

Vehicle 77 had no problems to drive the Extra-High phase of the class 3 cycle. The break-off
section of this vehicle is unambiguous (see Figure 49).
Vehicle 80 had a kerb mass of 1590 kg and a 30 minutes power of 50 kW and would have
been classified as class 2 vehicle with these values. But it was tested on the class 3 cycle,
once over the whole cycle and once with a second Low phase instead of the Extra-High
phase.
For vehicle 108 the break-off point was reached at a vehicle speed above 110 km/h, which
makes it really challenging to bring the vehicle to a stop within 15 seconds. As a
consequence this time period was extended to 60 seconds in the GTR.
The results of all CD tests for the PEV’s are summarised in Table 22. There is a dependency
of the CD test range and the average speed of the driven cycle but there are of course also
significant differences between the vehicles for a given average speed or a given cycle (see
Figure 50).
123
140
120 v_min v_set v_max 7700101 7700102 7700103 7700104 7700105
100
80
60
veh 77, test series 1, WLTC,
class 3, version 5,
40 L&M&H&exH
20
0
1100 1160 1220 1280 1340 1400
time in s
Figure 49: Time series of the vehicle speed for the CD test for vehicle 77 at break off
section
300
250
200
range in km
150
vehicle 59
vehicle 58
100
vehicle 80
vehicle 84
vehicle 108
50 Linear (vehicle 59) charge depleting tests
Linear (vehicle 84)
Linear (vehicle 108)
0
20 25 30 35 40 45 50
average speed in km/h
Figure 50: Range of the CD tests for the PEVs versus average speed of the cycles
124
vehicle
Test deceleration distance distance to
duration average distance number average speed at
IDveh series Test ID cycle ID description last 15 s in till end of stop last 15 s
in h in h in km of cycles in km end of test
ID m/s² test in m in m
in km/h
58 1 1 20 WLTC, class 2, version 1.4, L&M&H 7.2 253.5 17.3 61.91 ‐1.15 253,401 129.0
7.3 257.8
58 1 2 20 WLTC, class 2, version 1.4, L&M&H 7.3 262.2 17.9 62.74 ‐1.16 262,025 130.7
58 2 3 26 WLTC, class 2, version 1.4, L&M 9.8 269.6 34.4 34.39 ‐0.64 269,515 71.6
9.8 270.7
58 2 4 26 WLTC, class 2, version 1.4, L&M 9.9 271.8 34.7 45.63 ‐0.85 271,725 95.1
59 1 1 14 WLTC, class 3, version 5, L&M&H&L 5.4 166.4 9.2 33.88 ‐0.63 166,362 70.6
59 1 2 14 WLTC, class 3, version 5, L&M&H&L 5.4 5.4 167.7 9.3 167.6 41.08 ‐0.76 167,580 85.6
59 2 4 11 WLTC, class 3, version 5, L&M 6.8 186.6 23.8 59.03 ‐1.09 186,521 123.0
6.8 185.8
59 3 6 1 WLTC, class 3, version 5, L&M&H&exH 2.8 125.7 5.4 89.63 ‐1.66 125,481 186.7
2.8 126.0
84 1 1 31 WLTC, class 1, version 2, L&M&L 7.9 201.2 17.6 59.30 ‐1.10 201,101 123.5
8.0 203.6
84 1 2 31 WLTC, class 1, version 2, L&M&L 8.1 206.0 18.0 35.20 ‐0.65 205,947 73.3
7.0 200.2
84 4 6 12 WLTC, class 3, version 5, L&M&H 3.9 141.5 9.4 69.48 ‐1.29 141,369 144.8
108 1 1 11 WLTC, class 3, version 5, L&M, 1250 kg 5.9 164.5 21.0 40.89 ‐0.76 164,402 85.2
108 2 2 11 WLTC, class 3, version 5, L&M, 1350 kg 5.9 161.5 20.6 50.45 ‐0.93 161,441 105.1
108 3 3 1 WLTC, class 3, version 5, L&M&H&exH, 1250 kg 2.5 112.5 4.8 112.16 ‐2.08 112,290 233.7
108 4 4 1 WLTC, class 3, version 5, L&M&H&exH, 1350 kg 2.4 110.0 4.7 117.28 ‐2.17 109,760 244.3
Table 22: Results of charge depleting tests for the 6 pure electric vehicles
125
In addition to the PEVs, 2 OVC HEVs were tested on the class 3 cycle in CD mode (vehicles
60 and 65). Vehicle 60 had a kerb mass of 1730 kg, a 1.4 litre petrol engine with a rated
power of 63 kW and an electric motor with a peak power of 111 kW. Vehicle 65 had a kerb
mass of 1425 kg, a 1.8 litre petrol engine with a rated power of 73 kW and an electric motor
with 60 kW power, which is presumably the peak power. Both vehicles would be classified as
class 3 vehicles when only the rated power of the ICE is considered. The difference in kerb
mass is due to the fact that vehicle 60 had a much higher traction battery capacity than
vehicle 65.
This resulted in a much higher electrical range for vehicle 60 compared to vehicle 65 (see
Figure 51 to Figure 54). Vehicle 60 was able to drive almost 3 full class 3 cycles (all 4
phases) without assistance of the ICE, while vehicle 60 could only drive the Low, Medium
and High speed phases of one class 3 cycle in full electrical mode (this can be seen from the
comparison of Figure 51 and Figure 53).
Another difference between these vehicles was that the traction battery of vehicle 60
recharged to a certain extent during subsequent CS tests, while this was not the case for
vehicle 65 (this can be seen from the comparison of Figure 52 and Figure 54).
These results built the basis for the prescriptions for charge depleting and charge sustaining
tests in the GTR, especially for the break-off criteria (CD tests) and the determination of the
electric range for PEVs and OVC-HEVs.
But the results also show quite convincingly that the current vehicle classification for PEV
and OVC-HEV in the GTR is not satisfactory. For that reason a downscaling procedure was
developed during phase 1b, as well as a procedure to deal with vehicles that have a capped
maximum speed.
140 vehicle 60, PHEV, CD test, class 3 cycle 4000
3500
120
3000
100
engine speed in min‐1
2500
80
2000
60
1500
40
1000
set_speed
20 act_speed 500
eng_speed
0 0
0 1,000 2,000 3,000 4,000 5,000 6,000 7,000 8,000 9,000
time in s
Figure 51: Charge depleting test for OVC HEV vehicle 60, vehicle speed and engine
speed
126
90000
120
80000
100 70000
sum of CD current in As
60000
80
50000
60
40000
40 30000
set_speed 20000
20 act_speed
10000
current_high_battery_cum
0 0
0 1,000 2,000 3,000 4,000 5,000 6,000 7,000 8,000 9,000
time in s
Figure 52: Charge depleting test for OVC HEV vehicle 60, vehicle speed and current
4000
120
3500
100
3000
engine speed in min‐1
80 2500
60 2000
1500
40
set_speed 1000
20 act_speed
500
eng_speed
0 0
1800 2800 3800 4800 5800 6800
time in s
Figure 53: Charge depleting test for OVC HEV vehicle 65, vehicle speed and engine
speed
127
45000
120
40000
100 35000
sum of CD current in As
30000
80
25000
60
20000
40 15000
set_speed 10000
20 act_speed
5000
current_high_battery_cum
0 0
1800 2800 3800 4800 5800 6800
time in s
Figure 54: Charge depleting test for OVC HEV vehicle 65, vehicle speed and current
128
Appendix 1 – Utility Factors
Subject: Development of a European Utility Factor Curve for OVC-HEVs for WLTP
Authors: A. Eder45, N. Schütze46, A. Rijnders47, I. Riemersma48, H. Steven49
Date: November 2014
Introduction
In contrast to vehicles with combustion engines or NOVC-HEVs (not off vehicle chargeable
hybrid electric vehicles), an OVC-HEV (off vehicle chargeable hybrid electric vehicle) can be
operated in two distinct driving modes:
1.) Charge Depleting mode (electric energy is dissipated from the storage), and
2.) Charge Sustaining mode (electric storage is on a minimum level and only able to support
the driving with regenerated energy; the energy for driving is provided by the combustion
engine, see Figure 55)
The extent to which a vehicle will be driven in either of these modes depends on a
combination of the following factors:
 The capacity of the electric energy storage system;

 The electric energy consumption of the vehicle while driving in charge depleting mode;
 The distance that the vehicle is able to drive in charge depleting mode (resulting from the
first two factors);
 The length and frequency distribution of trips made with the vehicle; and
 The (off-vehicle) charging frequency for the electrical energy storage system.
The share between driving in ‘charge depleting’ and ‘charge sustaining’ mode can be
calculated from these factors, and is expressed as the ‘Utility Factor’ (UF). The UF is defined
as the ratio between the distance driven in ‘charge depleting’ mode divided by the total
driven distance, and can therefore range from 0 (e.g. for a conventional vehicle or for an
HEV) to 1 (for a pure electric vehicle or OVC-HEV that is driven in charge depleting mode
only). Since the fuel and energy consumption, as well as the emissions, are very different
between the two driving modes, Utility Factors are needed in order to calculate weighted
emissions, electric energy consumption, fuel consumption and CO2 values. UFs are based
on driving statistics and the ranges driven in ‘charge-depleting’ and ‘charge-sustaining’ mode
for OVC-HEVs in practical use. From these data, a Utility Factor (UF) curve can be
generated which facilitates a weighting between the measured (emission/electric
consumption/CO2/fuel consumption) values in the two driving modes (‘charge-depleting’ and
45
Dr. Andreas Eder, BMW Group, Germany, Email: andreas.ea.eder@bmw.de
46
Nico Schütze, BMW Group, Email: nico.schuetze@bmw.de
47
Andre Rijnders, RDW, Netherlands, Email: ARijnders@rdw.nl
48
Iddo Riemersma, Sidekick Projects, Netherlands, Email: iddo@sidekickprojects.nl
49
Heinz Steven, Data Analysis and Consultancy, Germany, Heinz.Steven@t-online.de
129
‘charge--sustaining’’) in depen
ndence of the measured range
e that was
s driven in charge
depletin
ng test on th
he WLTC.
Figure 55: Proced

dure with charge
c depleting and sustaining
g test
The currrent versio

on of the WLTP
W Global Technicaal Regulatio
on (GTR) does
d not co
ontain a
uniform UF curve; each con ntracting pa
arty may de evelop its own UF cu urve basedd on the
regional driving sttatistics. Fo
or the purp pose of furrther harmoonization between
b reg
gions, a
methodology for thhe UF calcu ulation and how to annalyse available driving
g data need ds to be
defined, which shoould then be e used for the determination of th
he regional Utility Facttor. This
Techniccal Report describes
d th
he methodo ology that was
w applied to determin ne the UF curve
c for
the Euro
opean Unio on, and is inntended to also
a provide
e a templatee for UF determination in other
regions.
Metho
odology for UF determina
ation
In orderr to develop a method dology for thet UF in such
s a way that the te
est result will
w be as
represe
entative as possible
p forr real-life driiving situations of an average
a OV
VC-HEV flee et, it is in
general important to
t follow thee procedure e as describ bed as follow
ws (Riemerssma [9]):
arch for all available
1. Sea a databases witth vehicle trrips that can
n be used as
a input;
2. Exclude the data that is errroneous or outside the
e scope (in this
t case, thhat might in
nclude
ng out data from vehicles that havve an extrem
takin me high or low
l daily avverage dista
ance);
3. Veriify the balan
nce in UF reelevant characteristics,, such as ro
oad types (c
city, rural,
high
hway), vehic cle types, sh
hare between EU mem mber states,, etc.;
4. Wheere necessa ary and app
propriate, ap
pply weightiing to correct any imba
alance in the
ese
charracteristics;
5. Devvelop a UF curve
c for the
e individual (weighted)) databases
s, and expla
ain differencces
through analysis; and
6. Based on the analysis,
a decide which weight shou
uld be appliied to each database in
n order
each the mo
to re ost represen
ntative UF curve.
c
In SAE J2841 [1] several methodologie es for UF determinatiion are described, wh hich are
nt purposess. Two of th
defined for differen hem in princcipal are su
uitable for th
he above de
escribed
purposee using travel survey daata:
130
 Flee
et UF (FUF F) – Parameters for th
he UF curve e are determ mined at fleeet level, asssuming
the distribution
n within the database is representtative for the target fleet (in this ccase: EU
region); and
 Individual UF (IUF) – Pa
arameters forf the UF curve are first determ mined at inndividual
hicle level, and
veh a are weighted to refflect the disstribution witthin the targ
get fleet.
The Fleeet UF is only

o an addequate me he calculatiion of UF if a repressentative
ethod for th
databasse of OVC-HEVs is avvailable. Fo
or the FUF, the ratio of
o the total electric
e ranges and
the tota
al daily kiilometres travelled
t fo
or all vehicles in thhe databasse are takken into
conside
eration.
This theerefore leadds to vehiclees with high daily travvelled distan

nces receiving a higheer weight
in the UUF calculatioon than thoose with low
wer daily traavelled dista
ances. As a result this method
is liable
e to producce inaccuratte results iff the datab base is not a valid sta
atistical sam
mple, for
example e if it conttains an unnrepresentaative share of vehicles travelling g longer orr shorter
distancees than thatt travelled by
b average OVC-HEV drivers.
d
his effect, SAE J2841 provides

In orderr to avoid th p the
e method fo or calculatin
ng an Individdual UF.
For this method, a distance weighted individual UF for f each veh hicle is dete
ermined. Thhe IUF is
calculatted by the arithmetic
a avverage over all the vehhicles in the
e database and
a therefo ore, each
vehicle’s (individuaal) UF has the same weight. In Fig gure 56, a comparison
c of both me
ethods is
shown.
Figure 56: Compaarison of Individuall UF (IUF F) and Fle eet UF (F FUF) calcu
ulation
method
ds (cs range = driven distance e in chargee sustainin
ng mode). In
I this
examp
ple, the ass
sumed maxx. electric ra
ange of a vehicle
v is 40
4 km.
The dattabases avaailable toda

ay contain a very high share
s onventional vehicles (see Sect.
of co
0). As a result the
e individual UF metho od has been used for the curren nt evaluation of the
Europeaan UF, whe ereas for a recommend ded re-evaluation of th
he UF curve e with a pure OVC-
HEV daatabase, thee fleet UF iss regarded as the morre accuratee calculation
n method. However
H
131
some preparatory work
w would be required
d in order to
o have a representative
e data set available
a
h a re-evaluation (see Sect.
for such S 0).
on necessary for the determinatio

Anotherr assumptio d on of UF is the chargin ng frequenccy of the
OVC-HE EV performmed by the customers.
c As it is currrently not possible
p to evaluate th he future
OVC-HE EV custommer charging g behaviouur, the assu umption of one charg ging event per day
(overnig
ght charge),, according to SAE J28841, is usedd for the furrther analyssis. In the fu
uture this
charging
g frequencyy might be modified
m if more
m accura ate data is available.
a
Datab
base for the
t calcu
ulation of
o Europe
ean Utilitty Factorr
The ma ain influencee on the UF
F curve by using the methodolog
m gy from SAE E J2841 is the
quality and
a resolution of inputt data. In orrder to get the most rep presentativee IUF, regio
onal
differences such as in custom mer behavio our (e.g. utilization, daily driven dis
stance, shaares
of differrent vehicle
e classes) or
o infrastruccture condittions (e.g. density
d andd practicalityy of
opportunities to ch harge the ve
ehicle) cause the need to focus on a specia al weightingg of
the inpuut data to co
orrect imbalances withiin a data baase.
Since O OVC-HEVs area fairly ne ew vehicle types

t on the
e European n market, theere is curre
ently
no repre esentative statistical
s da
ata availablle in Europe e about their practical use. Thereffore
it was d decided to use statisttics about thet use of convention nal vehicless instead. Two
T
comprehensive dattabases are e currently available.
a O is the European
One E W
WLTP database
[2]
that had already been use ed for the developme ent of the WLTC
W speeed profile. The
T
second one is a da atabase that was provid ded by FIAT T [3]. After th
he exclusion of erroneous
data (e.g. implausible start or end datess of recorded trips tha at leads to unrealistic trip
durationns), the dattabases com mbined con ntain in totaal about 1,4 400 conven ntional vehiccles
within thhe Europeaan Union.
Figure 57: Percen

ntage comp
position off average daily
d travelled kilome
eters and vehicle
v
types of
o WLTP an
nd FIAT dattabase
A compparison of both databasses is show wn in Figure

e 57. The EU WLTP da atabase sho ows
a very w
wide distrib
bution of driiving data and
a consists of a high share of diesel
d vehiccles.
Some ofo the vehicles have travelled very v long daily
d distan
nces on avverage (> 100
132
km/day)) which iss – at lea ast from today’s perrspective - considere ed to be not
represeentative for an OVC-HEV use in the near fu uture, as these long-distance drivving
custome ers would have signifficantly loweer total cosst of ownerrship with a conventio onal
diesel vvehicle compared to ann OVC-HEV V. As the absolute number of vehic cles, especiially
in some e segments, is relatively low, the distribution
d of averagee daily trave
elled kilomettres
is very heterogene eous, but onn the other hand, all reelevant vehhicle segme ents for Euroope
are reprresented at least once in the database.
In contrrast, the FIA AT databasse consists of a four times higherr total fleet mileage annd a
high sha are of petrool vehicles, but covers only the mid-to-small vehicle seg gments. Duee to
the high
h number off vehicles in n this databbase and thee limited av
vailable veh
hicle segments,
the distrribution of the daily travvelled kilom
metres is very homogen neous.
Neither database reflects

r the situation in Europe to the full exte ent with res
spect to vehhicle
segmen nts and mileage distrib butions. Thherefore both databases were co onsolidated for
further evaluation n, with thet following steps being taken t to improve the
represeentativenesss of the coombined da atabase. In a first ste ep, the N1 vehicles were
w
excludeed from the EU WLTP database,
d a OVC-HE
as EVs as repla acements fo or conventioonal
N1 veh hicles are not curren ntly in wide e use. In a second step differrent weighting
proceduures based on mileage e and statisstics about new registrrations were e applied. This
T
weightin
ng makes sure that the t distributtion of diffe
erent categgories, e.g. vehicle tyypes
and/or engine typ pes, can be e corrected d to be rep presentative
e of the ve ehicles on the
Europea an market (see
( Sect. 0).
0
The UF curves devveloped on the basis off the databa ases are sh
hown in Figu ure 58.
The redd dashed cuurve is based on a sim mple merge e of both daatabases, however
h in this
curve th
he Fiat data
a is dominatting due to the significcantly higher total fleet mileage in the
Fiat database compared to thee EU WLTP P database.
Figure 58: Compaarison of NEDC UF and WLTP UF scen narios bas

sed on diffferent
weighttings; cf=1
1.0 means s charging frequenc
cy of one (full) ove
ernight
charge
e per day (s
see Sect. 0).
0
133
In orderr to compen
nsate for th
his effect, a 50 % / 50 % weightin
ng approach h was applied.
For eacch electric range the respective e UF was calculated for both databases
d a
and
weighte
ed by 50 % for
f the calcuulation of th
he total UF (dashed
( purple line).
For com mparison, th
he UF curve e from the NEDC
N regulation UN/ECE R101 [4] is also sho own
graph (blackk solid line – “NEDC corrected to WLTP”). This curve was
in this g w adjusted d to
the reduuced electriic range in WLTP (due e to the higher energy demand off the WLTP P) in
order too ensure the comparability of all curves dep picted againnst the elecctric range of
o a
vehicle driving the WLTP drivving cycle. TheT correcte ed NEDC–U UF curve th hat is shown in
Figure 5 58 was dev veloped using the asssumption th hat the elecctric range of a vehiclee in
NEDC is reduced by b about 25 5% due to driving the WLTC,
W similar to the reductions found
from simmulations of
o electrified d vehicles. Hence, the e NEDC-UF F is plotted
d to the WL LTP
electric range, which leads to o a compres ssion in the
e electric range (e.g. th
he UF of 0.55 in
NEDC @ 25 km ele ectric rangee moves back to 18.75 km electricc range in WLTP).
W
Scientific Ana
alysis of Europea
an UF Cu
urves
As menntioned abov ve, the anaalysis of the
e compositio on of the da atabases shhows that b
both
databasses neet to be further complement
c ted in relatio
on to the fo
ollowing asp
pects:
- Thee number off vehicles frrom each co e database, compared to the num
ountry in the mber
of n
new registra
ations in the
e EU,
- Thee average annual
a vehiccle mileage driven withhin each cou
untry,
- Thee percenta age split o of the diffe erent vehiccle classes s within the EU (m mini,
subbcompact, compact,
c miiddle, luxury
y …),
- Thee percentagge of conve entional vehhicles with diesel-
d and petrol combustion eng
gine
with
hin each country,
In orderr to achievee this, statistics from the

t Europeean Environ
nment Agen ncy, as well as
[5],
data froom represenntative instittutes, were applied European Envvironment Agency
A IC
CCT
[6], [7]
and T
Transport & Mobility Leeuven .
Figure 59: Compa arison of registratio

ons of new w passenger cars and amou
unt of
vehicle
es in each country wiithin WLTP
P and FIAT database
d
134
As an example, Figure
F 59 illustrates the
t differen
nces betwe
een the am
mounts of new n
passengger car registrations fo
or each couuntry compared to thee amount off vehicles frrom
each ccountry withhin the da ata base. Three diffe erent appro
oaches weere applied d to
determine the conssolidated Euuropean UF F curve:
w vehicle registration
a) New r n numbers
The firsst step of the process wasw to identify the inteersection of countries in
n the datab
base and
in the European
E Union.
U If there is onlyy data availlable for a number off the countries, the
percentage shares s for countryy specific daata should be normalizzed to receive a total of
o 100%.
Afterwaards each ve ehicle belonngs to a couuntry specifiic sub-datab
base.
The seccond step isi to divide e the vehicles of each sub-database into ca ategories off engine
types (e
e.g. diesel, petrol, etc.). Accordingly countryy and engine type spec
cific UF currves can
be determined in thhe third stepp.
The lasst two calcuulations of the balanccing processs are to consolidate the numbe er of UF
curves to a correccted Europe ean UF curvve. Therefoore the engiine type sp
pecific UF curves
c of
ountry are weighted
each co w acccording to the
t country specific enngine type percentages
p s. Finally
the eng
gine type baalanced, bu ut country specific, UF curves are
e consolidatted by appllying the
country specific pe
ercentages ofo new vehicle registrations.
m of annua
b) Sum al vehicle mileage
m
Consolidation of th
he country-sspecific currves is done
e according
g to the sum
m of annual vehicle
e instead of new vehicle
mileage e registratio
ons.
c) Veh
hicle segm
ments
Before cconsolidatio
on of the UF, the data
abase was separated
s into differen
nt vehicle se
egments
and then weighted according tto the distribution of ve
ehicle types
s in the Euro
opean Union.
The diffferent weigh

htings descrribed above
e result in th
hree UF currves, as sho
own in Figurre 60.
Figure 60: Compa

arison of different UF
F balancing
g approaches
135
All three curves are located between the uncorrected UF that is represented by the dashed
red line (approach that consolidates both data bases without any weighting) and the 50/50
approach represented by the dashed purple line (approach that weights the WLTP UF curve
and the FIAT UF curve each with 50 %).
A further option would have been to apply a combined weighting for each of the above
mentioned criteria. However, there is not enough statistical data available to cover all vehicle
segments and therefore this option was not further evaluated.
From Figure 60 it can be seen that the 50/50 UF curve is a good fit for the trend line which
was weighted to the vehicle type percentages. Based on this analysis, it was decided and
agreed in EU WLTP Meeting in June 2014 that the 50/50 UF curve should be used in Europe
until more representative data are available (see Sect. 0).
The SAE J2841 also provides a method for how to describe the curve in a mathematical way.
Therefore the following exponential approach can be used. A number of coefficients are
provided to fit the curve towards an acceptable accuracy.
The described process to determine the coefficients ensures that several mathematical
requirements and characteristics are fulfilled.
For the calculation of a specific UF for each of the 4 cycle phases in the WLTC, the following
equation is applied:
  k
d 
j  i 1
  
UFi (d i )  1  exp  
 
 C j *  i
 dn



  UF l
  j 1  l 1
Where:
UFi Utility factor for phase i.
di Distance driven from the beginning of the charge depleting test up to the end
of phase i (phase i is the phase for which the delta UF is calculated) in km.
Cj jth coefficient (see Table 23)
dn Normalized distance (see Table 23).
k Amount of terms and coefficients in the exponent (see Table 23).

i Number of considered phase.
j Number of considered term/coefficient.
i 1
UF
l 1
l Sum of calculated utility factors up to phase (i-1).
For the approximated curve, terms and coefficients in the exponent are applied up to the
tenth order. The coefficient values shown in Table 23 are determined according to the
process described in SAE J2841 and fit the 50/50 curve with a maximum error of 0.001
(ΔUFmax = 0.1 %).
136
C1 26.25
C2 ‐38.94
C3 ‐631.05
C4 5964.83
C5 ‐25094.60
C6 60380.21
C7 ‐87517.16
C8 75513.77
C9 ‐35748.77
C10 7154.94
dn[km] 800
k 10
Table 23: Coefficients for the UF calculation equation
Review and recommended application of the European Utility

Factor
As this UF was derived from data based on conventional vehicles it is planned to re-evaluate
UF and charging frequencies by a customer study once a significant number of OVC-HEV
has been placed in the European market, see Figure 61.
Determination of Utility Factor from

heterogeneous databases and assumed
2014 use-cases;
Assumed charging frequency: 1,0 Number of sold
UF method*: Individual PHEV vehicles
in EU
Update of Utility Factor by

e.g. 2020 PHEV customer study
with respective charging frequency
UF method*: Fleet * according to SAE J2841
Figure 61: Schematic representation of re-evaluation
It is recommended that UFs are continuously checked for their robustness concerning the
application of future OVC-HEVs. In order to have a representative UF-study, it is
recommended to use a fundamental robust and scientific approach as described in Sampath
[8]
. An established method that could be used is the stratified sampling. This methodology can
be applied if it is necessary to divide a population into sub-populations.
137
Generally there are two main tasks: The first one is the sampling of vehicle data itself and the
second one the determination of weightings of sub-populations according to important criteria
concerning the evaluation of the UF-curve. The choice of customers that are considered to
represent a sub-population shall fulfil special criteria like a minimum annual mileage and
should be measured continuously during a minimum duration. It is recommended to use a
survey as outlined in Reiser [10] in order to select appropriate customers for the re-evaluation
of real-life UFs.
In addition to each specific OVC-HEV having to be analysed in each specific market
(including the separation of manufacturers, diesel- or petrol-OVC-HEV, different electric
ranges, vehicle-type (from mini to luxury), etc.), the following criteria could also be indicators
for different sub-populations of customers:
 Road category mainly used (highway, A-road, B-road) and home environment (urban,
suburban, rural);
 Driving style (more economic or more sporty); and
 Daily access to public and non-public charging infrastructure.
In order to get representative sub-populations, it is recommended that at least 20 vehicles
are available per survey for the re-evaluation described above. It is also important, that the
driving behaviour is recorded comprehensively for each driving mode (charge depleting and
charge sustaining) for at least 5,000 km per vehicle, in order to ensure that the whole variety
of driving states has been captured (see Reiser [10]).
The main focus of the Utility Factor approach described above is to calculate average values
which are mainly used for fleet monitoring. In contrast to conventional vehicles, the OVC-
HEV customers’ fuel consumption depends not only on driving behaviour and ambient
conditions, but also on driven range and charging frequency.
It is therefore recommended for customer information to provide not just a single fuel
consumption value, but instead to provide, for example, information on consumption
depending on the driven distances.
References
[1] SAE International J2841 “Utility Factor Definitions for Plug-In Hybrid Electric Vehicles
Using Travel Survey Data”, Revision from 2010/09, Issued 2009/03
[2] For access to database, please contact Heinz Steven, heinz.steven@t-online.de
[3] For access to database, please contact Luigi Orofino, Luigi.orofino@fiat.com.
[4] ECE/TRANS/WP.29/343 “Uniform provisions concerning the approval of passenger cars
powered by an internal combustion engine only, or powered by a hybrid electric power train
with regard to the measurement of the emission of carbon dioxide and fuel consumption
and/or the measurement of electric energy consumption and electric range, and of categories
M1 and N1 vehicles powered by an electric power train only with regard to the measurement
of electric energy consumption and electric range”,
http://www.unece.org/trans/main/wp29/wp29wgs/wp29gen/wp29fdocstts.html (2014/07/21)
[5] European Environment Agency “Monitoring CO2 emissions from new passenger cars in
the EU: summary of data for 2013”, 2014/04
[6] ICCT “European vehicle market statistics 2013”, 2013
138
[7] Transport & Mobility Leuven, “TREMOVE - economic transport and emissions model”,
http://www.tmleuven.be/methode/tremove/home.htm (2014/07/21)
[8] S. Sampath “Sampling theory and methods”, 2001, Narosa Publishing House
[9] Riemersma I., “Review of Utility Factor development”, Library of WLTP on CIRCABC -
https://circabc.europa.eu (2014-04-17)
[10] Reiser C.: Kundenverhalten im Fokus der Fahrzeugentwicklung (“Customer behaviour in
the context of vehicle development“), Technical University Dresden, Germany, PhD Thesis,
2010
139
Appendix 2 – Road Load Matrix Family
WLTP IG asked the Annex 4 Task Force to develop the Road Load Matrix Family as an
alternative road load determination option to the coast down, torque meter and wind tunnel
method on the one hand and the calculation method (default road load) on the other hand.
The objective is to deliver realistic road load values for low-volume vehicles, in particular for
large vans, under reduced test burden but without opening a loophole for unwanted
application. The IG indicated a conservative approach as the guiding principle when
developing the Road Load Matrix Family method. Although the Road Load Matrix Family is
built on physical laws, a safety margin should prevent the new method delivers profitable
values compared to standard methods (coast down, torque meter, wind tunnel) and should
provide an incentive to use, if possible, these standard methods.
Principle of the road load matrix family

The basic principle of the Road Load Matrix Family is only one generic road load
measurement, and extrapolation50 of the outcome of this measurement to derive the settings
of vehicle H and L for chassis dynamometer tests. This is in contrast to the standard road
load determination methods, which always use two measurements at the extremes for
vehicle H and L.
Scope
For the Road Load Matrix Family it was decided that the method should not be applicable for
high-volume main stream vehicles. This was achieved by setting an objective criteria by
means of a minimum limit to the technically permissible maximum laden mass of 3000 kg.
The scope of vehicles within the GTR itself is limited to a technically permissible maximum
laden mass of 3,500 kg.
Safety margin
The safety margin implemented in the Road Load Matrix Family method is ensured by the
following two elements:
1. Estimated worst Cd
An important principle of the Road Load Matrix Family method is the selection of a
representative test vehicle. On the one hand the test vehicle should be as representative
as possible for the vehicle family (estimated average mass of optional equipment,
representative body shape) in order to keep the actual average production vehicle as
close as possible to the measured test vehicle. On the other hand the aerodynamic
parameters are not considered in the extrapolation of the road load values to vehicle H
and L, therefore the representative body of the test vehicle should have a configuration
with the estimated worst-case Cd value (e.g by installing external body options such as
spoilers and roofrails, and by selecting the least aerodynamic wheel rims).
50
Strictly speaking this is not an extrapolation, but an extension. The term ‘extrapolation’ is selected as this
was the standard expression used in WLTP meetings. In the gtr-text the use of the term ‘extrapolation was
avoided.
140
2. Correlation factors: The road load values for vehicles H and L are calculated from the
value of the tested vehicle by extrapolation. To establish a safety margin conservative
correlation factors are introduced. In this appendix a correlation factor is defined as the
value to which the dominant vehicle parameters are assumed to correlate with the road
load. The correlation factor has a value between 0 and 1. To ensure a safety margin for
upward and downward extrapolation, the correlation factors are different in both
directions.
Correlation factors
Conservative correlation factors were introduced in order to derive road load values of each
individual vehicle and vehicles H and L that are very likely to be higher than the actual values
if they had been measured. The correlation factors are based on:
a) The best available scientific knowledge of the dependency of road load values to vehicle
parameters. Besides Cd, for which the worst case approach is chosen, the dominant
parameters are test mass (TM), tyre rolling resistance (RR) and frontal area (Af). These
parameters are selected as parameters in the correlation formulas.
b) Observed real world correlations. Only a very limited number of measurements was
available to the Task Force, indicating a direct correlation of typically 85-90% on the
selected vehicle parameters.
c) The determination of the conservative correlation factors was based on the following
assumptions and scientific evidence51:
i. the parameters selected to be included in the correlation are a selection of the
main influences. By assuming they together account for all of the road load
influences, consequently their impact will overrepresented.
ii. total rolling resistance is a combination of tyre and the drivetrain losses. Drivetrain
resistances are only slightly vehicle mass dependent. Typically drivetrain losses
make up for 10%-20% of the total f0 coefficient. The share is typically lower for
vehicle H than for vehicle L.
An EC study yielded 14% of drivetrain losses for a front wheel drive vehicle with
manual transmission. Larger effects for automatic transmissions and all-wheel
drive can be expected.
iii. remaining unexplained effects occur. The separation between rolling resistance
and air drag is not as straightforward as the f0 and f2 equations suggest. In the
standard coast down method this is overcome by the introduction of f1. Yet in the
Road Load Matrix Family method, f1 is set to 0.
d) The outer envelope of observed correlation is considered to be the conservative
approach, implying a higher correlation factor for calculation of road load values towards
vehicle H and a lower correlation factor towards vehicle L. This is shown in Figure 62.
The further away from the measured road load on the test vehicle, the higher the
extrapolated road load will be above the actual road load value.
Based on the evidence listed above, and discussions within the Annex 4 Task Force, a final
decision was made to use a correlation factor of 0.95 for upward extrapolation, and 0.80 for
downward extrapolation. These values should ensure similar safety margins to either sides.
A comparable stringency for upward and downward extrapolation brings an incentive to
select the test vehicle in the middle of the range of CO2 bandwidth.
51
Refer to document WLTP-11-17 at https://www2.unece.org/wiki/display/trans/WLTP+11th+session
141
Figure 62: Upward
d and downward extrrapolation for
f the road load mattrix family
Effect o
of the correla
ation factorr on the safe
ety margin
The saffety marginn for the seelected corrrelation facttors was ca
alculated baased on a (limited)
databasse of road load measu urements off heavy LCV V’s, which was
w provide ed by ACEAA51. This
was done to verifyy if the corrrelation facctors wouldd lead to coomparable safety marrgins for
vehicle L and H. These
T LCV
V’s have CO O2 emission ns in the order of 2600 to 300 g/kkm. The
e and relatiive safety margins
absolute m for the selecteed correlatio
on factors are
a indicate
ed in the
table be
elow for a ty
ypical example vehicle::
Corrrelation facto
or Safety marrgin1
Delta CO2 in
n g/km Relative deelta CO2
Xup2 0.95 2.7 1%
Xdown3 0.80 2.7 1%
1.
Safety marg gin is the ca
alculated rooad load vallues of vehiicle H or L minus
m the
m
measured road
r load va
alues for thee vehicles inn the databbase, expresssed in resu
ulting
d
delta CO2-fifigures.
2.
Xup is the coorrelation fa
actor for the
e calculation
n of the road es of vehicle H
d load value
3.
Xdown is the correlation factor for thhe calculatiion of the ro
oad load vallues of vehiicle L
Table 24: CO2 sa

afety margin for upwa
ard and dow
wnward ex
xtrapolation
n
CO2 callculation
By extraapolation of
o the road load to veh hicle L andd H, the tarrget road looads for me easuring
these vehicles at the
t chassiss dynamometer can be e found. Th he test vehicle is teste
ed at the
vehicle L and the vehicle
v H ro
oad loads, and
a the CO2 results are e used to drraw an interrpolation
line for CO2 again nst cycle energy. For any of the e other veh hicles in thee RLMF, thhe cycle
energy will be calcculated from
m the extrappolated roadd load, and then the COC 2 follows from
f the
ation metho
interpola od. Note tha at the CO2 interpolatio
on line doess not have the
t kinked shape
s of
red line in Figure 62.
6
ect of the sa
The effe afety margin on the ressulting CO2 value by a higher calcculated road
d load is
142
ally represe
graphica ented in Figure 63. For
F the middle vehicle e, the road load is meeasured,
hence tthere is no difference.. However, both vehiccle L and H have receeived a higher CO2
value be
ecause of th
heir higher road loads.
ety margin on CO2

Figure 63: Graphiical representation off the effectt of the safe
Family rrange and extension

e
Wherea as the famiily range of
o the stand y is 35% of the cycle
dard road load family e energy
demand d of vehicle
e HR, no dirrect limitatio
on for the Road
R Load Matrix Fam
mily was prroposed.
This lim
mitation was not deemeed necessarry since:
a) The sco
ope of application is limmited to vehicles with a technicallyy
permisssible maximmum laden mass
m abovee 3 tons
b) A representative ve ehicle (with worst-casee aerodynam mic drag) iss used
as a basis for the road
r load deetermination
n
c) The built in safety margin ensures that thhe differencee to the actual road
load wo
orsens for ve ehicles furth
her away froom the tested vehicle.
The me
ethod of the road load family
f matrix
x is included in the GT
TR in chapte
er 5 of Anne
ex 4.
CO2 inte
erpolation
In addittion to the determinatio
d on of the ro
oad load forr individual vehicles, th
he road loa
ad matrix
family approach
a is also extend
ded to the CO
C 2 determmination in order
o to reduce the test burden
and to a
avoid wind tunnel
t meassurements for
f vehicless falling in th
he scope off the RLMF.
The veh hicles within a RLMF for which CO C 2 interpoolation is ap
pplied havee to fulfill th
he same
criteria as for the Interpolation
I n Family off normal paassenger veehicles. Howwever, the RLMF
R is
not limitted to a 300g/km CO2-range.
- The
e larger the
e range of CO
C 2 the RL LMF, the wider
w the
safety mmargin of the road loadd will be and
d as a conseequence alsso for the CO
C 2.
143
Figure 42: Link be
etween saffety margin
n on Road Load and CO
C 2
The CO O2 measurem ments for veehicle L and H are perrformed on a chassis dyno
d using the road
loads off the RLMF F calculation
n. These me easured CO O2 values are used for the interpoolation of
individual CO2 valu ues based on o the indivvidual cyclee energy, which is an output
o of the RLMF
calculattion. This method
m is similar
s to what
w is alre
eady described in thee GTR on the t CO2
interpolaation for no
ormal passe enger cars, and
a is illustrated in Fig
gure 63.
It is also possible to reducce the saffety margin n by perforrming addiitional coasst down
measurrements bettween the representattive vehicle e and the vehicles
v L and/or H. With
W the
resulting
g cycle eneergies and using
u the COO2 measureements for vehicle
v L an
nd H, the CO
O2-value
for an in
ndividual ve
ehicle can th
hen be calculated more
e precisely.
144
Appendix 3 - Emission legislation
The following emission and fuel consumption legislation was reviewed as a basis for the
GTR:
US-Regulations (EPA and ARB)

CFR-2009-title40-part86-Volume18
CFR-2009-titel40-part1065-Volume32
CFR-2010-titel40-part1065-Volume32
CFR-2010-titel40-part600
California non-methane organic gas test procedures
Compliance guidance letters
Advisory Circulars
US CARB52
UNECE (comparable to EC 715/2007, EC 692 /2008)

ECE-R 83
ECE-R 101
ECE-R 24
GTR no.2 (Two-wheeled motorcycles)
GTR no.4 (Heavy duty vehicles)
Japan
Automobile Type Approval Handbook for Japanese Certification
52
Formaldehyde emissions from light-duty are measured with a methodology based on Federal Test
Procedure as set forth in subpart B, 40 CFR Part Subpart B, 40 CFR Part 86, and modifications located
in “CALIFORNIA EXHAUST EMISSION STANDARDS AND TEST PROCEDURES FOR 2001
AND SUBSEQUENT MODEL PASSENGER CARS, LIGHT-DUTY TRUCKS, AND MEDIUM-
DUTY VEHICLES” page II-1 and II-16 respectively.
The Formaldehyde test method used in CALIFORNIA EXHAUST EMISSION STANDARDS AND
TEST PROCEDURES FOR 2001 AND SUBSEQUENT MODEL PASSENGER CARS, LIGHT-DUTY
TRUCKS, AND MEDIUM-DUTY VEHICLES is the DNPH impinger method or DNPH cartridge.
After collecting Formaldehyde using DNPH impinger or DNPH cartridge, the sample is send to the Lab
to do analysis, such as HPLC.
145
Brazil
ABNT NBR 15598 (Brazilian Standard for Ethanol)
146
Appendix 4 - List of participants to WLTP
Germany
 Stephan Redmann, Ministry of Transport
 Christoph Albus, Ministry of Transport
 Hans Holdik, Ministry of Transport
 Oliver Eberhardt, Ministry of Environment
 Helge Schmidt, TÜV Nord
 Felix Kohler, TÜV Nord
France
 Beatrice Lopez, UTAC
 Celine Vallaude, UTAC
Japan
 Shingo Morita, MLIT
 Jumpei Ueda, MLIT
 Shun Masui, MLIT
 Kazuki Kobayashi, NTSEL
 Hajime Ishii, NTSEL
 Tetsuya Niikuni, NTSEL
 Kazuyuki Narusawa. NTSEL
 Norifumi Mizushima, NTSEL
 Daisuke Kawano, NTSEL
 Nick Ichikawa, JASIC
 Yuichi Aoyama, JASIC
 Toshihisa Yamaguchi, JASIC
 Takashi Fujiwara, JASIC
 Takahiro Haniu, JARI
 Takashi Naono, JASIC
 Masahito Yamashita, JASIC
 Yuki Toba, JASIC
Sweden
 Per Öhlund, Swedish Transport Agency
 Peter Smeds, Swedish Transport Agency
147
India
 H.A. Nakhawa, ARAI
 S. Marathe, ARAI
 Atanu Ganguli, SIAM
 Anoop Bhat, Maruti
 Vikram Khanna, Maruti
 K K Gandhi, SIAM
Poland
 Stanislaw Radzimirski, ITS
Netherlands
 André Rijnders, RDW
 Henk Baarbe, Ministry of Infrastructure & Environment
 Henk Dekker, TNO
 Rob Cuelenaere, TNO
Austria
 Werner Tober, TU Wien
South Korea
 Junhong Park, Ministry of Environment
 Simsoo Park, Korea University
 Hyonwoo Lee, KATRI
 Junho Lee, KATRI
 Hoimyung Choi, AICT
 Cha-Lee Myung, Korea University
 Charyung Kim, KATRI
 Inji Park, KATRI
 Wonwook Jang, Korea University
 Dongsoon Lim, KATRI
USA
 Ed Nam, EPA
 Michael Olechiw, EPA
Switzerland
 Giovanni D’Urbano, Federal Office for the Environment
148
UK
 Chris Parkin, DFT
 Craig Mills, DFT
 Simon Davis DFT
Canada
 Jean-Francois Ferry , Environment Canada
European Commission
 Cova Astorga-llorens, JRC
 Nikolaus Steininger, DG ENTR
 Maciej Szymanski, DG ENTR
 Alessandro Marotta, JRC
 Alois Krasenbrink, JRC
Independent Experts
 Serge Dubuc, Drafting Coordinator
 Heinz Steven, HS Data Analysis and Consultancy
 Iddo Riemersma, Sidekickprojects (expert for Transport & Environment)
 Greg Archer, Transport & Environment
 Christian Vavra, Maha
 Alexander Bergmann, AVL
 Kurt Engeljehringer, AVL
 Werner Hofegger, AVL
 Les Hill, Horiba
 Christian Bach, EMPA
 Rob Gardner, TRL
OICA
 Christophe Sierens, Honda
 Oliver Mörsch, Daimler
 Stefan Klimek, Daimler
 Walter Pütz, Daimler
 Konrad Kolesa, Audi
 Caroline Hosier, Ford
 Wiliam Coleman, Volkswagen
 Wolfgang Thiel, TRT Engineering
 Dirk Bäuchle, Daimler
 Stephan Hartmann, Volkswagen
 Alain Petit, Renault
149
 Samarendra Tripathy, Renault
 Eric Donati, PSA
 Bertrand Mercier, PSA
 Laura Bigi, PSA
 Toshiyasu Miyachi, JAMA Europe
 Thomas Mayer, Ford
 Kamal Charafeddine, Porsche
 Klaus Land, Daimler
 Daniela Leveratto, OICA
 Giovanni Margaria, OICA
 Christoph Lueginger, BMW
 Andreas Eder, BMW
 Markus Bergmann, Audi
 Thorsten Leischner, Daimler
 Thomas Vercammen, Honda
 Christoph Mayer, BMW
 Arjan Dijkhuizen, Toyota
 Paul Greening, ACEA
 Jakob Seiler, VDA
 Wouter, Vandermeulen, Daimler
 Claudia Walawski, Daimler
 Ernst-Peter Weidmann, Daimler
 Thomas Adam, Audi
 Pedro Casels, BMW
 Annette Feucht, Audi
 Winfried Hartung, Opel
 Thomas Johansson, Volvo
 Christoph Luenginger, BMW
 Bungo Kawaguchi, Toyota
 Matthias Nägeli, VW
 Raymond Petrovan, Opel
 Daniel Scherret, Opel
 Thomas Vogel, Opel
 Volko Rohde, VW
 Olle Berg, Volvo
 Takakuza Fukoka, Toyota
 Andreas Obieglo, BMW
 Ljubica Radic, BMW
 Ingo Scholz, VW
 Nico Schütze, BMW
 Marisa Faith, Ford
 Mark Guenther, Ford
 Anthony Smith, Ford
 Darren Crisp, Ford
150
AECC
 Dirk Bosteels
 John May
 Cecile Favre
ICCT
 Peter Mock
CLEPA
 Matthias Tappe, Bosch
 Danitza Fedeli, Delphi
 Pierre Laurent, CLEPA
 Peter Flanker, CLEPA
 Dimitris Vartholomaios, DENSO
151

WLTP - GRPE 72 02 Rev.1

Uploaded by

Copyright:

Available Formats

WLTP - GRPE 72 02 Rev.1

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

WLTP - GRPE 72 02 Rev.1

Uploaded by

Copyright:

Available Formats

Submitted by the Chair of the IWG on WLTP Informal document GRPE-72-02-Rev.

Technical Report on the development of a

Informal document no. GRPE-72-02-Rev.1

(sponsored by the European Commission)

3 Organisation, structure of the project and contributions of the different subgroups to

3.3.1 Laboratory procedures for electrified vehicles (LabProcEV) ..................................... 13

3.3.2 Particulate mass/Particulate number (PM/PN) .......................................................... 14

3.3.3 Additional pollutants (AP) .......................................................................................... 15

3.3.4 Reference fuel (RF) ................................................................................................... 15

3.4 WLTP phase 1b ................................................................................................................ 17

3.4.1 Drafting GTR ............................................................................................................. 18

3.4.2 EV subgroup .............................................................................................................. 19

3.4.3 AP Taskforce ............................................................................................................. 19

3.4.4 Round Robin testing .................................................................................................. 20

3.4.5 Taskforces on open issues ........................................................................................ 22

3.4.5.1 Reference Fuels ................................................................................................. 24

3.4.5.2 Definitions .......................................................................................................... 25

3.4.5.3 Normalization ..................................................................................................... 27

3.4.5.4 Number of tests .................................................................................................. 30

3.4.5.5 Review of coastdown tolerances ........................................................................ 32

3.4.5.6 Fuel consumption calculation ............................................................................. 32

3.4.5.7 Speed trace tolerance / drive trace index ........................................................... 33

3.4.5.8 Utility Factors ..................................................................................................... 36

3.4.5.9 Additional pollutants ........................................................................................... 37

3.4.5.10 Mode selection and predominant mode ............................................................. 39

3.4.5.11 Other taskforces ................................................................................................. 41

4.4.1 Interpolation method .................................................................................................. 46

4.4.2 Vehicle selection ....................................................................................................... 48

4.4.3 Interpolation/extrapolation range ............................................................................... 49

4.4.4 Vehicle test mass ...................................................................................................... 49

4.4.5 Vehicle coastdown mode and dynamometer operation mode................................... 50

4.4.6 Tyres ......................................................................................................................... 51

4.4.7 On-board anemometry .............................................................................................. 51

4.4.8 Default road load factors ........................................................................................... 52

4.4.9 Road load matrix family ............................................................................................. 52

4.4.10 Torque meter method ................................................................................................ 54

4.4.11 Wind tunnel method .................................................................................................. 56

4.4.12 Alternative delta Cd.A determination.......................................................................... 62

4.4.13 Road load family ........................................................................................................ 63

4.4.14 Manufacturer’s responsibility on road load ................................................................ 66

4.4.15 Alternative vehicle warm-up procedure ..................................................................... 67

4.4.17 Electrified Vehicles .................................................................................................... 68

4.4.18 RCB correction for OVC-HEVs, NOVC-HEVs and NOVC-FCHVs ............................ 69

4.4.19 Shortened test procedure for PEV range test ........................................................... 73

4.4.20 Phase-specific values for EVs ................................................................................... 77

4.4.21 Interpolation method for electrified vehicles .............................................................. 80

4.4.22 End of PEV range criteria .......................................................................................... 82

4.4.23 FCV test procedure ................................................................................................... 84

4.4.24 WLTP post-processing .............................................................................................. 85

4.5.1 Annex 3 – Reference fuels ........................................................................................ 87

4.5.2 Annex 4 - Road and dynamometer load .................................................................... 87

4.5.3 Annex 5 – Test equipment and calibrations .............................................................. 95

4.5.4 Annex 6 – Type 1 test procedure and test conditions ............................................... 96

4.5.5 Annex 7 – Calculations .............................................................................................. 98

4.5.7 Annex 9 – System equivalency ............................................................................... 104

5 Validation of the test procedure ......................................................................................... 105

5.1.1 Participants and vehicles, measured parameter ..................................................... 105

5.1.2 Evaluation issues .................................................................................................... 111

5.2 Results ............................................................................................................................. 112