SCR control and OBD

TC Blois – February 2017

Delphi Confidential
Agenda
• Day 1
 AM

• Mechanization
• Feed-forward control
 Principle and Structure
 Calibration procedure
 Calibration guidelines – calibration tools
 PM

• Closed-loop control
 Principle and Structure
 Calibration procedure
 Calibration guidelines – calibration tools

Delphi Confidential 2
Agenda
• Day 2
 AM

• On-Board Diagnostic
 Principle and Structure
 Calibration procedure
 PM

• Case study: Data acquisition and calibration on vehicle

Delphi Confidential 3
Content
• Mechanization
• Feed-forward
 Principle and Structure
 Calibration procedure

• Calibration guidelines – calibration tools

• NOx-sensor based closed-loop

 Principle and Structure
 Calibration procedure

• Calibration guidelines – calibration tools

• OBD
 Principle and Structure
 Calibration procedure

• Calibration guidelines – calibration tools

Delphi Confidential 4
Content
• Mechanization
• Feed-forward
 Principle and Structure
 Calibration procedure

• Calibration guidelines – calibration tools

• NOx-sensor based closed-loop

 Principle and Structure
 Calibration procedure

• Calibration guidelines – calibration tools

• OBD
 Principle and Structure
 Calibration procedure

• Calibration guidelines – calibration tools

Delphi Confidential 5
System mechanizations

Urea
doser T3

SCR DPF DOC

Turbine

NOx T6 T5
Urea
doser

SCR SDPF DOC

NOx T5
Urea
doser

SCR DPF LNT

NOx T6 NOx T5

Delphi Confidential 6
Mechanization and instrumentation for development
LNT + U/F SCR
T° sensor T° sensor Sample probe
/thermocouple
Thermocouple
Urea
NOx sensor
Injector

LNT DPF SCR

NH3 sensor
NOx sensor
T° sensor

Sample probe Sample probe

Probes / analysers Sensors Thermocouples

(1.5 mm max)
- LNT-in - NOx sensor - U/F SCR-bed
HC, CO, …. - DPF-out
- DPF-out - U/F SCR-out
NO - NH3 sensor
NOx - U/F SCR-out
- U/F SCR-out - Temperatures
NOx - LNT-in (or thermocouple)
NH3 - LNT-out
- U/F SCR-in

Delphi Confidential 7
Content
• Mechanization
• Feed-forward
 Principle and Structure
 Calibration procedure

• Calibration guidelines – calibration tools

• NOx-sensor based closed-loop

 Principle and Structure
 Calibration procedure

• Calibration guidelines – calibration tools

• OBD
 Principle and Structure
 Calibration procedure

• Calibration guidelines – calibration tools

Delphi Confidential 8
Fundamentals
Principle (1)

• NOx aftertreatment requires reductants

 Lean NOx Trap uses fuel, in rich “regeneration” phases
 SCR uses NH3 in continuous lean operation

• NH3 is supplied by liquid urea injection (adblue) in the exhaust line

 Adblue

• Is a 32.5% urea solution

• Freezes at -11°C

Urea
(*)
NH 2  CO  NH 2 (aq)  H 2 O(g)  2NH 3 (g)  CO 2 (g)
Delphi urea doser
(*) mass of CO2 produced is roughly
50% of the NOx mass to convert

Delphi Confidential 9
Fundamentals
Principle (2)

• NH3 reduces NOx in the SCR catalyst

4NH 3  4NO  O 2  4N 2  6H 2 O
2NH 3  NO  NO 2  2N 2  3H 2 O Fast reaction

8NH 3  6NO 2  7N 2  12H 2 O

• NOx conversion is optimum in the fast reaction conditions

 NO2/NOx ratio close to 50%
 NO2 is produced in the DOC/CDPF (NO oxidized with O2)

• DOC/CDPF technology adapted to SCR applications

Delphi Confidential 10
Fundamentals
Principle (3)

• NOx reacts with NH3 stored in the SCR catalyst (coverage).

Maximum conversion efficiency [ ] 100%

80% Required stored NH3 for

optimum conversion
60% Required NH3 quantity
for a complete NEDC

40%
180°C

20% 200°C

350°C
0%
0 5000 10000 15000
Stored NH3 [mg]

• Efficiency driven by NH3 coverage

 Control objective is to maintain NH3 coverage

Delphi Confidential 11
Feed forward
Control structure

Operating
NH3
conditions
Coverage
Target

SCR T° model
Urea
+ injection
– controller Urea
SCR-in NOx
sensor &
Injector
model
NH3
Coverage SCR
NO2/NOx
model
Observer cat.
Exhaust flow

SCR model

• The urea is dosed such that the target coverage is maintained

 Filling when the SCR is empty
 Continuous dosing of consumed NH3 (α x NOx)

• SCR model must be characterized

Delphi Confidential 12
Feed forward
Inputs modules – Engine-out NOx

Operating
NH3
conditions
Coverage
Target

SCR T° model
Urea
+ injection
– controller Urea
SCR-in NOx
sensor &
Injector
model
NH3
Coverage SCR
NO2/NOx
model
Observer cat.
Exhaust flow

Delphi Confidential 13
SCR-in NOx sensor and model
• When the NOx sensor is ready (warm and active), SCR-in NOx concentration
is measured
• When the NOx sensor is not available (cold / faulty), SCR-in NOx
concentration is modelled by the LNT model

SCR-in NOx
sensor &
model NOx
LNT model

LNT temperature NOx

storage,
release
Exhaust mass flow & NH3
Conversion SCR-in NOx
Air/Fuel ratio

NOx sensor

Delphi Confidential 14
Feed forward
Inputs modules – SCR bed temperature

Operating
NH3
conditions
Coverage
Target

SCR T° model
Urea
+ injection
– controller Urea
SCR-in NOx
sensor &
Injector
model
NH3
Coverage SCR
NO2/NOx
model
Observer cat.
Exhaust flow

• The SCR bed temperature is modeled with a multi-slice thermal inertia

model without thermal losses (no impact in steady-state)
Outputs:
- TSE_Scr_bed_temp (1 selectable slice)
used for SCR catalyst model
Inputs:
TSE_Scr_in_temp - TSE_Scr_out_temp
ACM_Exh_muffler_flow
Delphi Confidential 15
Feed forward
Inputs modules – SCR bed temperature

• Calibration procedure:
 Measure SCR-in and SCR bed temperatures in transient conditions

• Exhaust flow also required

• A calibration tool is used to match the model to the measurement

 The tool can be provided
 Alternative: Delphi can calibrate based on measurements

Delphi Confidential 16
Feed forward
Inputs modules – NO2/NOx model

Operating
NH3
conditions
Coverage
Target

SCR T° model
Urea
+ injection
– controller Urea
SCR-in NOx
sensor &
Injector
model
NH3
Coverage SCR
NO2/NOx
model
Observer cat.
Exhaust flow

• SCR-in NO2/NOx ratio model depends on the oxidation activity of the LNT
 LNT can increase NO2/NOx ratio compared with engine-out NOx/NOx

• More important when LNT is not regenerated and NOx slips to SCR
 Influencing parameters

• Exhaust flow (SV), LNT temperature

• P_T_DOC_DPF_NO2_CONV_EFF_APM

Delphi Confidential 17
Feed forward
Inputs modules – NO2/NOx model

• Calibration procedure:
 Without LNT regens, transient recording (NEDC, WLTC, …) of

• NO2/NOx downstream of CDPF

• ACM_Turbine_flow
• TSE_Doc_out_temp
 Effect of soot load

NO2/NOx ratio [ ]
higher exhaust flow

100 150 200 250 300 350 400

DOC out T° [°C]

Delphi Confidential 18
Feed forward
SCR catalyst model

Operating
NH3
conditions
Coverage
Target

SCR T° model
Urea
+ injection
– controller Urea
SCR-in NOx
sensor &
Injector
model
NH3
Coverage SCR
NO2/NOx
model
Observer cat.
Exhaust flow

• NH3 coverage observer: 0D chemical model accounting for

 NH3 storage capacity: T°
 NOx conversion efficiency: T°, NO2/NOx, Space Velocity, Stored NH3 (coverage)

Delphi Confidential 19
Feed forward
SCR catalyst model – NH3 / NOx stoichiometry

• Due to the concurrence of several chemical 4NH3  4NO  O2  4N 2  6H 2O

reactions, the NH3/NOx stoichiometry 2NH3  NO  NO2  2N 2  3H 2O

changes with NO2/Nox ratio and temperature 8NH3  6NO2  7N 2  12H 2O

• NH3/NOx stoichiometry is calibrated in P_T_SCR_NOX_NH3_STOICH_APM,

as a function of TSE_Scr_bed_temp and P_T_Scr_no2_nox_ratio
 Not straightforward to identify on engine/vehicle
Catalyst supplier data recommended

•
Default calibration can be used
if catalyst supplier data is not
available. Not recommended
 Key values:

= 1 when only NO is present

for a rear SCR for example

= 1.33 when only NO2 is present

Delphi Confidential 20
Feed forward
SCR catalyst model – Urea to NH3 conversion

• Urea is converted to NH3 in two steps

 Thermolysis
CO(NH 2 ) 2  NH 3  HNCO
 Hydrolysis
HNCO  H 2 O  NH 3  CO 2
 Urea to NH3 conversion starts around 160°C

• The urea to NH3 conversion “efficiency” is calibrated in

P_T_SCR_UREA_TO_NH3_CONV_EFF_APM as a function of
TSE_Scr_in_temp

• Calibration procedure
 Steady State testing: 1 test per SCR temperature
 Calibration tool: Excel tool

• see Steady State procedure and tool described later

Delphi Confidential 21
Feed forward
SCR catalyst model – NH3 capacity

• NH3 storage capacity per catalyst volume [mg/L] is calibrated in

P_T_SCR_NH3_CAPACITY_APM, function of SCR bed temperature
TSE_Scr_bed_temp

• The SCR catalyst volume is calibrated

in
P_T_SCR_CATALYST_VOLUME_APV

• Calibration procedure
 Steady State testing: 1 test per SCR temperature
 Calibration tool: Excel tool

• see Steady State procedure and tool described later

Delphi Confidential 22
Feed forward
SCR catalyst model – NOx conversion efficiency [1]

• The dependency of NOx conversion efficiency to Temperature and stored NH3

is calibrated in P_T_SCR_NOX_CONV_EFF_APM
 Function of TSE_Scr_bed_temp and P_T_Scr_stored_nh3

• This table corresponds to maximum achievable NOx conversion efficiency

 Lowest space velocity
 Optimum NO2/NOx ratio

• Calibration procedure
 Steady State testing: 1 test per SCR temperature
 Calibration tool: Excel tool

• see Steady State procedure and tool described later

Delphi Confidential 23
Feed forward
SCR catalyst model – Steady State Calibration

• One single test procedure can be used to calibrate

 the urea to NH3 conversion P_T_SCR_UREA_TO_NH3_CONV_EFF_APM
 the NH3 storage capacity P_T_SCR_NH3_CAPACITY_APM
 the preliminary maximum NOx conversion efficiency P_T_SCR_NOX_CONV_EFF_APM

• “preliminary” = relative dependency of stored NH3 and temperature

• For a given SCR temperature, run a steady state NH3 filling

 Every 20°C …50°C 50 250

45 225
 At lowest space velocity
40 200
 Measurements: P_T_Scr_urea_mass_flow
35 NOx in 175
• SCR-in NOx

Urea flow [mg/s]

NOx out

Emissions [ppm]
30 150
IN_Nh3_concentration
• Exhaust mass flow 25 125
• NH3 downstream of SCR 20 100

• NOx downstream of SCR 15 75

• SCR Temperature 10 50

• NO2/NOx upstream of SCR 5 25

 Let all emissions stabilize 0 0
0 500 1000 1500 2000
Time [s]

Delphi Confidential 24
Feed forward
SCR catalyst model – Steady State Calibration
100% 100
Measured efficiency vs stored NH3
• An Excel tool is provided to calibrate 90%
Efficiency calibration curve
90
80% 80

NOx conversion efficiency [ ]

 P_T_SCR_UREA_TO_NH3_CONV_EFF_APM
70% 70


Emissions [ppm]
P_T_SCR_NOX_CONV_EFF_APM
60% 60
 P_T_SCR_NH3_CAPACITY_APM 50% 50
40% 40
30% 30
20% 20
10% 10
0% 0
0 500 1000 1500
Stored NH3 [mg]

Delphi Confidential 25
Feed forward
SCR catalyst model – Effect of SV and NO2/NOx ratio on NOx conv efficiency

• NOx conversion efficiency model:

η NOx  F (T, NH3stored , NO 2 /NO x , SV)

• ηNOx_max (T°,NH3stored)
= P_T_SCR_NOX_CONV_EFF_APM
 Preliminary calibration (relative dependency
of stored NH3) comes from the excel tool
 Final calibration with compiled Matlab tool
(see next slides)

• SV, NO2/NOx (and T°) influence on

efficiency
 ηcorr_L(T°,NO2/NOx)
= P_T_SCR_NOX_CONV_EFF_SL_LOW_APM
 ηcorr_H(T°,NO2/NOx)
= P_T_SCR_NOX_CONV_EFF_SL_HIGH_APM
 k(SV)
= P_T_SCR_NOX_CONV_EFF_SL_IPOL_APM
 Calibrated in compiled Matlab tool (see next slides)

Delphi Confidential 26
 Feed forward
SCR catalyst model – Effect of SV and NO2/NOx ratio on NOx conv efficiency

• The exhaust flow (Space Velocity) impacts NOx conversion directly and
indirectly, via NO2 production in the DOC

1000 1
100%

Exhaust flow x10 [g/s] - SCR temperature [°C]

900 0.9
90% Effect of flow on
Maximum conversion efficiency [ ]

80% 800 0.8

SCR efficiency

NOx conversion efficiency

70% 700 0.7
Exhaust mass flow
60% 600 0.6
Effect of flow on SCR bed temperature
50% 500 NOx conversion efficiency 0.5
NO2/NOx ratio
40% 400 0.4
30%
300 0.3
20%
200 0.2
10%
100 0.1
0%
0% 25% 50% 75% 100% 0 0
NO2/NOx ratio [ ] 1590 1600 1610 1620 1630 1640 1650 1660
Time [s]

Delphi Confidential 27
Feed forward
SCR catalyst model – Tuning Calibration Tool

• In steady state, it is difficult to decouple NO2/NOx ratio, from exhaust

temperature and exhaust flow
• Transient data can be used to finalize calibration, with a compiled Matlab tool
 WLTC, RDE, or any transient cycle, with emission measurement

Delphi Confidential 28
Feed forward
NH3 coverage target calibration

• Optimum performance of the SCR catalyst will be reached when filling the
catalyst to its maximum capacity

• Calibration table:P_T_SCR_NH3_STORED_TARGET_APM = Target fraction

of NH3 storage capacity, function of TSE_Scr_bed_temp
 A value of 1 corresponds to a request of P_T_SCR_NH3_CAPACITY_APM

• Typical values for a SCR on Filter:

Delphi Confidential 29
 Feed forward
NH3 coverage target calibration

• To prevent severe NH3 slip, an undefloor SCR cannot be operated too close to
its fully capacity (storage =1)

NH3 storage capacity @ 15ppm NH3 slip

• The target NH3 stored is a
calibration trade-off between
Required NH3 for optimum NOx conversion
and low/no NH3 slip conversion efficiency and
risks of NH3 slip
Stored NH3

Severe NH3 slip

Acceptable NH3 slip

0 100 200 300 400 500

SCR Temperature [°C]

Delphi Confidential 30
LNT and SCR
8 160
• The LNT is the key for low Cum EO NOx
LNT CDPF SCR
7 Cum LNT-out NOx 140
temperature "conversion" Cum SCR-out NOx
6 120
 the SCR is too cold Vehicle speed

5 100
• The LNT cannot be regenerated

Cumulated NOx
4 80
at low temperature
3 60
 At cold start, 2 40
the LNT must be "empty"
1 20

0 0
0 500 1000 1500 2000
Time [s]

• The only way to ensure an "empty" LNT at any time (potential engine shut-
down on RDE) is to always regenerate the LNT < 400...450°C whatever driving
conditions
• Above 400...450°C the NOx storage capacity is low which naturally keeps it "empty"

Delphi Confidential 31
LNT and SCR LNT CDPF SCR

• Combined control of LNT and SCR

 LNT: Low temperature operation and low space velocity (exhaust flow)
 SCR: Medium/high temperature operation and high space velocity
• Trend from in-production LNT + UF SCR application
 LNT used (active regens) wherever possible except in highway conditions (CO2)
• LNT T° ≥ 230°C to achieve effective LNT regens
• LNT T° ≤ 400°C, for NOx storage function
 SCR used whenever possible to treat all NOx coming out of LNT
SCR

LNT

Delphi Confidential 32
Delphi development vehicle
• One of Delphi's internal control development vehicles is based on an
LNT+DPF+SCR architecture
 Honda CRV with 1.6 L engine
• HP + LP EGR
• Single stage turbo
 Close-coupled LNT+CDPF + Underfloor SCR

LNT CDPF SCR

 Delphi FIE, ECU and engine control system

 Delphi HP urea dosing system

Delphi Confidential 33
Engine + Aftertreatment control
• Main control features present on development vehicle
 EURO 6b production control +
• LP EGR control
• Urea dosing control
 including NH3 slip management (no significant NH3 slip at tailpipe)
• Combined control of LNT + SCR
• Emission results and behavior (LNT vs SCR) comparable with other OEM
production vehicle
0.48
0.44 WLTC

0.40 FTP
0.36
warm US06 (RDE)
0.32
0.28
NOx

0.24
0.20
RDE NTE (EURO6 x 2.1)
0.16
0.12
EURO 6
0.08
0.04
0.00
Engine-out LNT-out SCR-out
Delphi Confidential 34
Content
• Mechanization
• Feed-forward
 Principle and Structure
 Calibration procedure

• Calibration guidelines – calibration tools

• NOx-sensor based closed-loop

 Principle and Structure
 Calibration procedure

• Calibration guidelines – calibration tools

• OBD
 Principle and Structure
 Calibration procedure

• Calibration guidelines – calibration tools

Delphi Confidential 35
Background – NOx sensor

Delphi Confidential 36
Background – NOx sensor
• NOx sensor measured NH3 as NOx (with some saturation at high NH3
concentrations)

Delphi Confidential 37
NOx sensor based closed-loop
Control structure

Operating
conditions NH3
Coverage
Target

SCR T° model
Urea
+ injection
– controller Urea
SCR-in NOx
sensor & Injector
model
NH3
Coverage SCR
NO2/NOx
model
Observer cat.

Exhaust flow NOx sensor

Urea dosing
correction NH3 slip
Injection scaler information
&
NOx
efficiency
Catalyst model
calculation
correction
Modeled NOx efficieny scaler

Delphi Confidential 38
Closed Loop prerequisites
• Need for 2 outputs (injection and modeled efficiency control)
 If the catalyst looses efficiency,

• the measured NOx efficiency is low, injection has to be reduced (to avoid NH3 slip)

 If the injector does not deliver enough urea,

• the measured NOx efficiency is low, injection has to be increased (to higher the efficiency)

 For the same measured effect, depending on the root cause of the error, opposite actions have to be
taken

• The closed loop needs 2 outputs and 2 inputs

Delphi Confidential 39
NOx sensor based closed-loop
Control Logic

• Virtual NH3 sensor

 From the NOx sensor signal, a “virtual” sensor generates two separate outputs signals, NOx and NH3,
based on

• NOx sensor value

• Modeled NOx efficiency
• SCR-in NOx concentration

• Corrections are slow (> 100 km for stabilization)

 SCR cat is a big NH3 buffer: No emission impact on cycle if urea is completely shut off
 Instantaneous deviations have no direct impact on emissions

Delphi Confidential 40
Results
400
NH3 sensor

NH3 slip [ppm] - Efficiency [%] - Temperature [degC]

• Road data from durability vehicle 350 NH3 slip Model
SCR temperature

• Temperature range 300

 200 to 350 degC

250

• NH3 slip range 200

 0 to 200 ppm
150

600 100
NOx sensor
NH3 sensor
500 NH3 slip Model 50
SCR temperature

0
400 0 2 4 6 8 10 12 14
[ppm] - [%] - [degC]

Time [h]

300

200

100
Zoom

0
9.4 9.6 9.8 10 10.2 10.4 10.6 10.8 11
Delphi Confidential Time [h] 41
Open-loop behavior

Effect on NOx efficiency Effect on NH3 slip

Too low injection Lower than expected None

Higher than expected

Too high injection High slip
(or None)

Degraded catalyst
Lower than expected High slip
efficiency

Increased catalyst
Higher than expected None
efficiency

Delphi Confidential 42
Fuzzy logic
• Variables can be transformed to fuzzy signals (ie. have values between 0 and
1).
 0 is equivalent to a boolean 0
 1 is equivalent to a boolean 1
 Intermediate values are possible

• Fuzzy signals created :

 NOx efficiency lower than modeled (eff1)
 NOx efficiency same as modeled (eff2)
 NOx efficiency higher than modeled (eff3)
 Low NH3 slip (sip1)
 High NH3 slip (slip2)

Delphi Confidential 43
NOx sensor based closed-loop
Control Principle

NOx eff < NOx eff = Model NOx eff > Model
Model (eff1) (eff2) (eff3)

NH3 slip low (OK)

Injection + No action Efficiency +
(slip1)

NH3 slip high

Efficiency - Injection - Injection -
(slip2)

inj   (eff2  eff3) slip2

inj   eff1 slip1
eff   eff1 slip2
eff   eff3  slip1
Delphi Confidential 44
Fuzzy logic

inj  (eff2  eff3)  slip2 eff1, eff2, eff3

inj   eff1 slip1 1

0.8

Fuzzy value
eff   eff1 slip2 0.6
Efficiency too low
Efficiency OK
0.4 Efficiency too high

eff   eff3  slip1 0.2

0
0.6 0.7 0.8 0.9 1 1.1 1.2 1.3 1.4
Efficiency error

slip1, slip2
• Fuzzification : 1

0.8

Fuzzy value
0.6 NH3 slip OK

 The efficiency error is the error calculated with 0.4

NH3 slip too high

the NOx sensor

0.2
 The NH3 slip is the filtered slip of the “virtual” 0
sensor (filter faster in case of increasing signal) 0 10 20 30 40
NH3 slip [ppm]
50 60 70 80

Delphi Confidential 45
Control equations
• The boolean control equations can be translated to fuzzy logic equations :
 AND is replaced by MIN
 OR is replaced by MAX

• The output of each equation is a fuzzy signal (value between 0 and 1)

which acts on an integrator
 Integrator for injector correction :
• Inj_correction = Inj_correction + Gain*inj+ - Gain*inj-
 Integrator for efficiency model correction
• Eff_correction = Eff_correction + Gain*eff+ - Gain*eff-
 Two separate gain values for injector and efficiency

Delphi Confidential 46
Closed loop overview

Uncorrected NH3 flow

Model inputs Stored NH3 mass

Uncorrected urea flow Desired urea flow
Open loop model Urea command
NOx in Converted NH3 flow

Injector correction
factor
Corrected model
Efficiency Injection action
Injection correction

Efficiency error Fuzzy efficiency error

Signal processing Fuzzification Closed loop control
NOx sensor NH3 slip Fuzzy NH3 slip

Efficiency correction
Efficiency action

Efficiency correction factor

Delphi Confidential 47
Simulation results
1.2

1.1

Injection correction
0.9
Efficiency correction
Efficiency error
0.8 Injection error
50 degC temp shift
0.7
0 1000 2000 3000 4000 5000 6000
Mileage [km]

1
Real NOx conversion efficiency

0.9

0.8

0.7

0.6

0.5
0 1000 2000 3000 4000 5000 6000
Mileage [km]

160

140

120
Real NH3 slip [ppm]

100

80

60

40

20

0
0 1000 2000 3000 4000 5000 6000
Mileage [km]

Delphi Confidential 48
Simulation results (errors indirectly corrected)
1.2

1.1

0.9
Injection correction
0.8 Efficiency correction
Inlet NOx error
0.7
0 500 1000 1500 2000 2500 3000 3500
Mileage [km]

1
Real NOx conversion efficiency

0.9

0.8

0.7

0.6

0.5
0 500 1000 1500 2000 2500 3000 3500
Mileage [km]

160

140

120
Real NH3 slip [ppm]

100

80

60

40

20

0
0 500 1000 1500 2000 2500 3000 3500
Mileage [km]

Delphi Confidential 49
Boolean vs Fuzzy logic
• A boolean control approach could simplify the algorithm
 Replace the fuzzy signals by boolean signals :

1 1

0.8 0.8

Fuzzy value
Fuzzy value

Efficiency too low Efficiency too low

0.6 0.6
Efficiency OK Efficiency OK
0.4 Efficiency too high 0.4 Efficiency too high

0.2 0.2

0 0
0.6 0.7 0.8 0.9 1 1.1 1.2 1.3 1.4 0.6 0.7 0.8 0.9 1 1.1 1.2 1.3 1.4
Efficiency error Efficiency error

1 1

0.8 0.8

Fuzzy value
Fuzzy value

0.6 NH3 slip OK 0.6 NH3 slip OK

NH3 slip too high NH3 slip too high
0.4 0.4

0.2 0.2

0 0
0 10 20 30 40 50 60 70 80 0 10 20 30 40 50 60 70 80
NH3 slip [ppm] NH3 slip [ppm]

 No intermediate speed for the integrators : difficult to find optimum integrator gain values (stability /
precision compromise)

Delphi Confidential 50
Preferred / Corner conditions
• Algorithm does not work in pure steady state condition (constant tailpipe NOx)
 Variation on Passenger Car applications always enable the calculation often enough
 Closed-loop control is stopped under this condition

• Modeled NOx efficiency is a main input to the equation

 The equation considers a constant gain error K on the efficiency during 1s
• If the modeled efficiency gain error changes rapidly, it affects directly the NH3 signal
 Example of the modeled NOx conversion efficiency accuracy on development vehicles on WLTC test :

Delphi Confidential 51
Closed-loop validation results
Nominal system
700

600
Vehicle speed  SCR status
Temp[°C] - VSpeed [km/h]

DOC out
500 SCR bed – FUL (Nominal)
400

300

200  Urea injection

100 – Nominal
0
0 5 10 15 20 25 30 35 40
Time [h]

 Urea quality
– Nominal
2
Inj scaler
NOx eff scaler
1.5
Scalers [ ]

1  Engine-out NOx
– Sensor
0.5

0
0 5 10 15 20 25 30 35 40
Time [h]

300
NH3 slip [ppm] - Efficiency [%]

Measured efficiency [%]

250 NH3 sensor
NH3 slip Model
200

150

100

50

0
0 5 10 15 20 25 30 35 40
Time [h]

Delphi Confidential 52
Closed-loop validation results
Under-dosing
700

600
Vehicle speed  SCR status
Temp[°C] - VSpeed [km/h]

DOC out
500 SCR bed – FUL (Nominal)
400

300

200  Urea injection

100 – – 30%
0
0 5 10 15 20 25 30 35 40 45
Time [h]
 Urea quality
– Nominal
2
Inj scaler
NOx eff scaler
1.5
Scalers [ ]

1
 Engine-out NOx
– Sensor
0.5

0
0 5 10 15 20 25 30 35 40 45
Time [h]

300
NH3 slip [ppm] - Efficiency [%]

Measured efficiency [%]

250 NH3 sensor
NH3 slip Model
200

150

100

50

0
0 5 10 15 20 25 30 35 40 45
Time [h]

Delphi Confidential 53
Closed-loop validation results
Over-dosing
700

600
Vehicle speed  SCR status
Temp[°C] - VSpeed [km/h]

DOC out
500 SCR bed – FUL (Nominal)
400

300

200  Urea injection

100 – + 30%
0
0 5 10 15 20 25
Time [h]
 Urea quality
– Nominal
2
Inj scaler
NOx eff scaler
1.5
Scalers [ ]

1
 Engine-out NOx
– Sensor
0.5

0
0 5 10 15 20 25
Time [h]

300
NH3 slip [ppm] - Efficiency [%]

Measured efficiency [%]

250 NH3 sensor
NH3 slip Model
200

150

100

50

0
0 5 10 15 20 25
Time [h]

Delphi Confidential 54
Closed-loop validation results
BPU (OBD Tier2 Bin5) SCR system
700

600
Vehicle speed  SCR status
Temp[°C] - VSpeed [km/h]

DOC out
500 SCR bed – BPU
400

300

200  Urea injection

100 – Nominal
0
10 20 30 40 50 60 70
Time [h]
 Urea quality
– Nominal
2
Inj scaler
NOx eff scaler
1.5
Scalers [ ]

1
 Engine-out NOx
– Sensor
0.5

0
10 20 30 40 50 60 70
Time [h]

300
NH3 slip [ppm] - Efficiency [%]

Measured efficiency [%]

250 NH3 sensor Some over-estimation of
NH3 slip Model
200 NH3 slip leads to over
150 correction of injection
100

50

0
10 20 30 40 50 60 70
Time [h]

Delphi Confidential 55
Urea Dosing Control
Closed Loop 700

600
Vehicle speed

Temp[°C] - VSpeed [km/h]

DOC out
500 SCR bed

400

• Durability vehicle at TCL : 110 000 km for SCR control validation

300

200

100

• Deviations introduced for performance evaluation

0 5 10 15 20 25 30 35 40
Time [h]
700
Vehicle speed
600

Scalers [- ]VSpeed [km/h]

2 DOC out
500 SCR
Inj bed
scaler
NOx eff scaler
400
1.5
300

Temp[°C]
1
200

 Nominal system 100

0.5
0
0 5 10 15 20 25 30 35 40 45
Time [h]
0
0 5 10 15 20 25 30 35 40
700
Time [h]
2 Vehicle speed
600

[ ] [%] [km/h]
DOC out
Inj scaler
500
300 SCR
NOx bed
eff scaler

- VSpeed
1.5 Measured efficiency [%]
400
250

 Urea injection deviation -30%

NH3 sensor

Efficiency
300 NH3 slip Model

-Scalers
1
200

Temp[°C]
200
150

[km/h] NH3 slip [ppm]

100
0.5
1000
0 5 10 15 20 25
50
0
700 Time [h]
0 5 10 15 20 25 30 35 40 45
Vehicle speed
600
0 Time [h]
DOC out
20 5 10 15 20 25 30 35 40
500 SCR bed Time [h]
Inj scaler
[ -] VSpeed 300
400 NOx eff scaler
1.5
[%]
Measured efficiency [%]
250
300 NH3 sensor
Efficiency

 Urea injection deviation +30%

NH3 slip [ppm] -Temp[°C]
Scalers

NH3 slip Model

1
200
200
100
150
0.5
0
100 10 20 30 40 50 60 70
Time [h]
0
50 0 5 10 15 20 25
Time [h]
02
0 5
Inj scaler 10 15 20 25 30 35 40 45
300 NOx eff scaler Time [h]

 Highly degraded SCR catalyst

1.5
[ ] [%]

Measured efficiency [%]

250 NH3 sensor
Efficiency
NH3 slip [ppm] - Scalers

1 NH3 slip Model

200
(to OBD limit) 150
0.5
100

0
50
10 20 30 40 50 60 70
Time [h]
0
Delphi Confidential 0 5 10 15
Time [h]
20 25
56
300
Urea Dosing Control
Closed Loop – Calibration process

• Data acquisition with offline calibration tool (*)

 NOx phasing
 Virtual NH3 sensor
 NOx efficiency error filtering
 NH3 slip filtering

• Data acquisition and calibration adjustements

 Fuzzification of NOx eff error and NH3 slip
 Control gains

• (*) Offline calibration tool under development

 Tool to be available mid 2017

Delphi Confidential 57
Closed-loop control
Calibration process – NOx phasing

• SCR-in NOx synchronization with SCR-out NOx sensor

 To calculate the online NOx conversion efficiency, the engine-out NOx must be phased with the rear NOx
sensor
 The signal processing consists in

• A volume delay between the SCR-in and SCR-out NOx sensor signal
(P_T_SCR_NOX_CONC_IN_OUT_VOL_APV)
• A filtering of the engine-out NOx (P_T_NOX_CONC_SCR_OUT_FILT_APM and
P_T_NOX_CONC_SCR_OUT_DEC_SCL_APV factor for decreasing signals)

• Calibration procedure
 Drive in transient conditions, with varying exhaust flow and NOx, without urea, to avoid NOx conversion in
the SCR(s)
 Offline calibration tool: adjust the calibration parameters until the filtered and delayed engine-out NOx
(P_T_Nox_concentration_filtered) matches the rear NOx sensor

Delphi Confidential 58
Closed-loop control
Calibration process – Virtual NH3 sensor

• Main calibration parameters for the virtual sensor are a ramp limiter
(P_T_SCR_EST_NH3_SLIP_GRAD_APV) on the raw virtual NH3 and a first
order filter (P_T_SCR_EST_NH3_SLIP_FILT_APV)
 The NH3 sensitivity of the NOx sensor can be calibrated in P_T_SCR_EST_NH3_CORRECTION_APM

Raw virtual NH3

Virtual NH3 signal

• Calibration procedure
 Drive in transient conditions, with varying NH3 slip, and NOx
 Offline calibration tool: adjust the calibration parameters until the virtual sensor (P_T_Scr_nh3_slip_est)
matches the measured NH3 slip

• Pre-requisite: The NOx conversion efficiency model(s) must have been

calibrated

Delphi Confidential 59
Closed-loop control
Calibration process – NOx efficiency error filtering

• Like the virtual NH3 sensor, the NOx efficiency error is calculated from filtered
values with P_T_SCR_EST_NOX_ERR_FILT_APV

"Measured" NOx conversion efficiency

Modelled NOx conversion efficiency

• Calibration procedure
 Same calibration procedure as the virtual sensor: Offline calibration tool: adjust the calibration parameters
to obtain clean (free of noise) NOx efficiency error

Delphi Confidential 60
Closed-loop control
Calibration process – NH3 slip filtering

• Depending on driving conditions, NH3 slip events can occur from time to time.
A heavy filter is then applied on the NH3 slip to "memorize" NH3 slip events
 The filter increases quickly (P_T_SCR_NH3_SLIP_FILT_INC_APV) but decreases slowly
(P_T_SCR_NH3_SLIP_FILT_DEC_APV)

P_T_Scr_nh3_slip
P_T_Scr_nh3_slip_filtered

Time
• Calibration procedure
 Record NH3 slip (NH3 sensor and virtual sensor) in random driving conditions (road)
 Offline calibration tool: adjust the calibration parameters such that NH3 slip events are memorized in a time
range corresponding to the desired control dynamics

Delphi Confidential 61
Closed-loop control
Calibration process – Fuzzification of NOx eff error and NH3 slip

• The Fuzzification tables can be established based on pre-determined criteria

on acceptable NH3 slip and NOx efficiency degradation
1

0.8

Fuzzy value
Efficiency too low
0.6
Efficiency OK
0.4 Efficiency too high

0.2

0
0.6 0.7 0.8 0.9 1 1.1 1.2 1.3 1.4
Efficiency error

0.8

Fuzzy value
0.6 NH3 slip OK
NH3 slip too high
0.4

0.2

0
0 10 20 30 40 50 60 70 80
NH3 slip [ppm]

Delphi Confidential 62
Closed-loop control
Calibration process – Control gains

• The fuzzy control signals act as integrators on

 injection correction P_T_Scr_nh3_inj_scaler_nvv with a gain P_T_SCR_INJ_CORR_GAIN_APV
 efficiency model correction P_T_Scr_nox_eff_scaler_nvv with a gain P_T_SCR_EFF_CORR_GAIN_APV

• Calibration procedure
 Acquire significant quantity of random data (road) with typical system deviations (via modification of
injector calibration and rapid catalyst aging)
 Adjust the gains to obtain required control dynamics: Not faster than the SCR system dynamics to avoid
oscillations

Delphi Confidential 63
Content
• Mechanization
• Feed-forward
 Principle and Structure
 Calibration procedure

• Calibration guidelines – calibration tools

• NOx-sensor based closed-loop

 Principle and Structure
 Calibration procedure

• Calibration guidelines – calibration tools

• OBD
 Principle and Structure
 Calibration procedure

• Calibration guidelines – calibration tools

Delphi Confidential 64
Diagnostics specific to SCR systems
• US OBD has a larger scope than EOBD
 SCR OBD was then developed to first meet US OBD requirements

EOBD US OBD II
•NMHC Conversion Efficiency
•SCR Conversion Efficiency •Misfire detection
•Fuel System • Positive Crank Case
Ventilation
•Exhaust Gas Sensor
• Engine Cooling System
•EGR System
• Idle Speed & Qty
•Boost Pressure Control
• Cold Start Emission
•LNT Conversion Efficiency Reduction Strategy
•PM Filter Efficiency • Permanent Fault
•Comprehensive Components • Similar Conditions
•Other Emission Control • I3C mismatch
System
•Rate Based Monitoring

Delphi Confidential 65
Diagnostics specific to SCR systems
• (f) MONITORING REQUIREMENTS FOR DIESEL/COMPRESSION-IGNITION ENGINES.
 (1) NON-METHANE HYDROCARBON (NMHC) CONVERTING CATALYST MONITORING

• (1.2.3) Other aftertreatment assistance functions:

 The OBD II system shall detect a malfunction when the catalyst is unable to generate the necessary feedgas constituents for proper
SCR system operation.

 (2) OXIDES OF NITROGREN (NOx) CONVERTING CATALYST MONITORING

• (2.2.2) Conversion Efficiency:

 (A) The OBD II system shall detect a NOx catalyst malfunction when the catalyst conversion capability decreases to the point that NOx
or NMHC emissions exceed 1.75 times the applicable FTP standards.
• (2.2.3) Selective Catalytic Reduction (SCR) or Other Active/Intrusive Reductant Injection System Performance:
 (A) Improper reductant delivery leading to exceeding threshold specified in (f)(2.2.2)(A).
 (B) Empty tank
 (C) Improper reductant (the reductant tank is filled with something other than the reductant)
 (D) Feedback control:
• (i) If the system fails to begin feedback control within a manufacturer specified time interval;
• (ii) If a failure or deterioration causes open loop or default operation; or
• (iii) If feedback the control system has used up all of the adjustment allowed by the manufacturer or reached its maximum
authority and cannot achieve the feedback target.

 (5) EXHAUST GAS SENSOR MONITORING

• (5.2.2) NOx sensors:

 (A) Sensor performance faults: The OBD II system shall detect a malfunction prior to any failure or deterioration of the sensor (…) that
would cause a vehicle’s emissions to exceed 1.75 times the applicable FTP NOx standard
Delphi Confidential 66
US Diagnostics specific to SCR systems
Feedgas monitoring

 (1) NON-METHANE HYDROCARBON (NMHC) CONVERTING CATALYST MONITORING

• (1.2.3) Other aftertreatment assistance functions:

 The OBD II system shall detect a malfunction when the catalyst is unable to generate the necessary feedgas constituents
for proper SCR system operation.

• Functional diagnostic only

• No specific diagnostic function required:

 Detection of poor NO2 by

• Poor NOx conversion: To be covered by NOx conversion monitoring

AND
• Poor NMHC conversion: Not specific to SCR systems

• On most applications, data prove that this is not a failure mode (OBD) with
recent SCR technologies (Cu-zeolite)
Delphi Confidential 67
US Diagnostics specific to SCR systems
Empty tank detection

 (2) OXIDES OF NITROGREN (NOx) CONVERTING CATALYST MONITORING

• (2.2.3) Selective Catalytic Reduction (SCR) or Other Active/Intrusive Reductant Injection System Performance:
 (B) Empty tank

• Straightforward based on level sensor embedded in the urea tank module

 In EU, this is part of the "inducement" strategies and not OBD

Delphi Confidential 68
US Diagnostics specific to SCR systems
Feedback control

 (2) OXIDES OF NITROGREN (NOx) CONVERTING CATALYST MONITORING

• (2.2.3) Selective Catalytic Reduction (SCR) or Other Active/Intrusive Reductant Injection System Performance:
 (D) Feedback control:
• (i) If the system fails to begin feedback control within a manufacturer specified time interval;
• (ii) If a failure or deterioration causes open loop or default operation; or
• (iii) If feedback the control system has used up all of the adjustment allowed by the manufacturer or reached its
maximum authority and cannot achieve the feedback target.

• Feedback control = “time to control” = “time to dose” ∆tTTCL

Diagnosis concept
• Estimation of "theoretical start of dosing“ dependent on modeled SCR-
catalyst temperature or modeled heat quantity
• Start timer after "theoretical start of dosing“
• Evaluation if real "start of dosing" is TRUE when maximum allowed
time is elapsed (DtTTCL-max) input for failure detection
Challenge
• Complex SCR-temperature model required CARB requirement: TSCR-
model without any exhaust gas temperature sensor (T6-model based
on T2)

Delphi Confidential 69
US Diagnostics specific to SCR systems
NOx conversion

 (2) OXIDES OF NITROGREN (NOx) CONVERTING CATALYST MONITORING

• (2.2.2) Conversion Efficiency:

 (A) The OBD II system shall detect a NOx catalyst malfunction when the catalyst conversion capability decreases to the
point that NOx or NMHC emissions exceed 1.75 times the applicable FTP standards.
• (2.2.3) Selective Catalytic Reduction (SCR) or Other Active/Intrusive Reductant Injection System Performance:
 (A) Improper reductant delivery leading to exceeding threshold specified in (f)(2.2.2)(A).
 (C) Improper reductant (the reductant tank is filled with something other than the reductant)

• Faults to be detected
 SCR_catalyst_fault
 Urea_solution_low
 Urea_injection_low
 Non OBD

• Urea injection_high
• Urea solution_high

Delphi Confidential 70
OBD principle
Overview
Operating
NH3
• OBD
conditions
Coverage  Urea quality and delivery diagnostic
Target
 Catalyst malfunction diagnostic
SCR T° model
Urea
+ injection
– controller
SCR-in NOx Urea
sensor & model Injector
NH3
Coverage
NO2/NOx
model
Observer SCR
cat.
Modeled NOx
Exhaust flow efficiency
NOx sensor

÷ Urea dosing
NH3 slip
correction
x information
&
OBD Measured NOx NOx
efficiency
Catalyst model efficiency
Function correction calculation

NH3 slip

• Diagnostics based on NOx conversion efficiency monitoring

 Pro: Direct link between NOx conversion efficiency and tailpipe NOx (regulated) on emission cycle

Delphi Confidential 71
OBD principle (2)
NOx efficiency low

• A specific filter (1st order) is applied to the instantaneous measured NOx

conversion efficiency
 The slow adaption closed-loop correction is not fast enough for OBD purposes
 The same fast filter is applied to the modeled efficiency
 The filter is held when not in the appropriate enabling conditions (see further)
 The ratio is used as an efficiency error

P_T_SCR_OBD _NOX _EFF_FILT _APV

2
P_T_Nox _concentration _filtered
K
5
"Measured" NOx conversion efficiency
IN _Nox _tailpipe _concentration
In
Out
3 Init _Out
P_T_Scr_nh 3_slip
R

real _nox _err_diag Simulink_std_lpf

< 1
Modelled4 NOx conversion efficiency K P_T_SCR_OBD _NOX _EFF_THR _APV P_T_Scr_obd _nox _eff _low
P _T_Scr_nox _conv _efficiency
In
Out
mdl _nox _err_diag Init _Out

1 R
p_t_scr_nox _eff _firsttime
Simulink_std_lpf

Delphi Confidential 72
OBD principle (3)
NH3 slip thresholds

• The NH3 slip is compared to two thresholds for pinpointing purposes

 High level
 Medium level
 (NH3 slip below medium level is considered as acceptable)

• At low SCR temperatures (< 350°C) NH3 slip may be very sporadic
 Slip flags are held via turn-off delays

P _T_SCR_OBD_NH3_SLIP _HOLD _T_APV Convert T

Out 1
1 P _T_Scr_obd _nh 3_slip_high
> In
P_T_Scr_nh 3_slip P_T _SCR_OBD_NH3_SLIP _HIGH _APV

Simulink_Turn_Off _Delay

P _T_SCR_OBD_NH3_SLIP _HOLD _T_APV Convert T

Out 2
> P _T_Scr_obd _nh 3_slip_medium
In
P_T_SCR_OBD _NH3_SLIP _MEDIUM _APV

Simulink_Turn_Off _Delay

Delphi Confidential 73
OBD principle (4)
Fault tree

NOx efficiency
low

Yes

medium < NH3 slip NH3 slip

< high > high

Yes No Yes No

Catalyst efficiency Tank recently Doser Tank recently

System OK
too low re-filled self diagnostic re-filled

Yes No Yes No

Urea concentration Urea dosing Urea concentration Urea dosing

too low too low too high too high

Delphi Confidential 74
 OBD principle (5)
Enabling conditions

• Enabling conditions for efficiency and NH3 check

 Inlet NOx concentration > CAL

• NOx sensor accuracies

 Exhaust mass flow > CAL

• Space velocity discrimination effect

 CAL1 < DOC out temperature < CAL2

• NO/NO2 model accuracy

 CAL1 < SCR bed temperature < CAL2

• SCR activity window

 NH3 stored mass > CAL

• Ensures minimum catalyst activity

 No modelled NH3 slip

• When slip is expected (severe T° transients), using the virtual NH3 sensor is too risky

• Enabling conditions for diagnostic

 Minimum time in efficiency and NH3 check conditions

Delphi Confidential 75
Validation results
Nominal – Test cycle
Eff. [%] - NH3 slip [ppm] Temp[°C] - VSpeed [km/h]

200 Vehicle speed

SCR bed
• SCR status
100  FUL
0
0 200 400 600 800 1000 1200 1400
Time [s]

100
Modeled • Urea injection
Measured
50
 Nominal
0
0 200 400 600 800 1000 1200 1400
Time [s]

1.5
•
Efficiency ratio [ ]

Eff. ratio Urea quality

1 Eff. check

0.5
 Nominal
0
0 200 400 600 800 1000 1200 1400
Time [s]

1 • Closed-loop
Slip high
Slip level [ ]

0.5
Slip medium
Slip check  Active

0
0 200 400 600 800 1000 1200 1400
Time [s]

1
Catalyst fault
Faults [ ]

Urea conc low

0.5 Urea inj high
Urea inj low
Urea conc high
0
0 200 400 600 800 1000 1200 1400
Time [s]

Delphi Confidential 76
Validation results
Urea injection too low – Test cycle
Eff. [%] - NH3 slip [ppm] Temp[°C] - VSpeed [km/h]

200 Vehicle speed  SCR status

SCR bed

100 – FUL
0
0 200 400 600 800 1000 1200 1400
Time [s]  Urea injection
100
Modeled – 33% lower
Measured
50

0
 Urea quality
– Nominal
0 200 400 600 800 1000 1200 1400
Time [s]

1.5
Efficiency ratio [ ]

Eff. ratio
1 Eff. check
 Closed-loop
0.5
– Active
0
0 200 400 600 800 1000 1200 1400
Time [s]

1
Slip high
Slip level [ ]

Slip medium
0.5 Slip check

0
0 200 400 600 800 1000 1200 1400
Time [s]

1
Catalyst fault
Faults [ ]

Urea conc low

0.5 Urea inj high
Urea inj low
Urea conc high
0
0 200 400 600 800 1000 1200 1400
Time [s]

Delphi Confidential 77
Validation results
Degraded SCR – Test cycle
Eff. [%] - NH3 slip [ppm] Temp[°C] - VSpeed [km/h]

200 Vehicle speed  SCR status

SCR bed

100 – BPU
0
0 200 400 600 800 1000 1200 1400
Time [s]  Urea injection
100
Modeled – Nominal
Measured
50

0
 Urea quality
– Nominal
0 200 400 600 800 1000 1200 1400
Time [s]

1.5
Efficiency ratio [ ]

Eff. ratio
1 Eff. check
 Closed-loop
0.5
– Active
0
0 200 400 600 800 1000 1200 1400
Time [s]

1
Slip high
Slip level [ ]

Slip medium
0.5 Slip check

0
0 200 400 600 800 1000 1200 1400
Time [s]

1
Catalyst fault
Faults [ ]

Urea conc low

0.5 Urea inj high
Urea inj low
Urea conc high
0
0 200 400 600 800 1000 1200 1400
Time [s]

Delphi Confidential 78
Validation results
Nominal – Road driving (durability)
Eff. [%] - NH3 slip [ppm] Temp[°C] - VSpeed [km/h]

600
Vehicle speed
SCR bed
• SCR status
400

200  FUL
0
0 2 4 6 8 10 12
Time [h]

100
Modeled • Urea injection
Measured
50
 Nominal
0
0 2 4 6 8 10 12
Time [h]

1.5
•
Efficiency ratio [ ]

Eff. ratio Urea quality

1 Eff. check

0.5
 Nominal
0
0 2 4 6 8 10 12
Time [h]

1 • Closed-loop
Slip high
Slip level [ ]

0.5
Slip medium
Slip check  Active

0
0 2 4 6 8 10 12
Time [h]

1
Catalyst fault
Faults [ ]

Urea conc low

0.5 Urea inj high
Urea inj low
Urea conc high
0
0 2 4 6 8 10 12
Time [h]

Delphi Confidential 79
Calibration process overview
• Calibration process
 Open loop :

• (1) Pre-requisite models: SCR bed T° and NOx models

• (2) DOC/CDPF NO2 production model (catalyst related)
4 • (3) SCR catalyst models. Requires data from catalyst supplier
Operating • (4) Target coverage (efficiency) calibration
NH3  Closed loop:
conditions
Coverage
Target • (5) Sensor dynamics and closed-loop response
 OBD
SCR T° model
Urea
+ injection
1 – controller Urea
SCR-in NOx
sensor & Injector
model
NH3
Coverage SCR
NO2/NOx
2 model
Observer cat.

Exhaust flow NOx sensor

3
Urea dosing
correction NH3 slip
Injection scaler information
&
NOx
efficiency
Catalyst model 5
calculation
correction
Delphi Confidential Modeled NOx efficieny scaler 80
81 Delphi Confidential