Attachment 06.

AER augex model

Assessing the Power and Water Corporation’s augex
forecast

A report to Power and Water Corporation

Confidential final

23 February 2018
Table of contents

Executive summary ............................................................................... 3

Key assessment findings 3

1 Introduction ........................................................................... 5
1.1 Background and scope 5
1.2 Nuttall Consulting experience in this task 6
1.3 Key information sources 6
1.4 Structure 6

2 The AER’s augex model .......................................................... 8

2.1 Overview of augex model 8
2.2 AER augex model form, inputs and output 9
2.2.1 Network specification inputs – network segments and groups 9
2.2.2 Network specification inputs - utilisation profile 9
2.2.3 Network specification inputs – utilisation growth 9
2.2.4 Planning parameters inputs 10
2.2.5 Model outputs 10
2.3 Calibration 11
2.3.1 Augmentation unit cost 11
2.3.2 Volume planning parameters 11

3 PWC augex model development .......................................... 13

3.1 Overview 13
3.2 Augex model development 13
3.2.1 Segmentation 13
3.2.2 Utilisation profiles 14
3.2.3 Load growth 16
3.3 Model calibration 17
3.3.1 Historical calibration period 17
3.3.2 Set up of calibration data 17
3.3.3 Determining planning parameters 18
3.3.3.1 Calculating the unit cost ........................................................................... 19
3.3.3.2 Calculating the utilisation threshold parameters ..................................... 19
3.3.4 Scenario planning parameters 19

4 Augex forecast assessment .................................................. 20

4.1 Model assessment results 20
4.2 Assessment discussion and conclusions 20
4.3 Summary and conclusions 23

Nuttall Consulting does not take responsibility in any way whatsoever to any person or
organisation other than Power and Water Corporation in respect of information set out in
this document, including any errors or omissions therein, arising through negligence or
otherwise.
Nuttall Consulting

Executive summary
Nuttall consulting has been engaged by Power and Water Corporation (PWC) to undertake an
assessment of its augmentation expenditure (augex) forecast. This assessment must use the
predictive model the Australian Energy Regulator (AER) has indicated it will use as part of the
process it will apply to assess expenditure forecasts. This model is called the AER augex model.
To prepare this model, we have used data that PWC will report in its Reset Regulatory Information
Notice. This process has been supported by other data provided by PWC and other comments and
advice provided during the course of various meetings with relevant PWC personnel.
To undertake these assessments, we have applied similar principles to those used when undertaking
an assessment using the AER’s repex model. In this regard, model planning parameters have been
calibrated to reflect the four years of PWC’s augex (2013/14 to 2016/17 inclusive 1). As such, the
model forecast could be considered a type of indicative intra-company (or business-as-usual)
benchmark study 2.

Key assessment findings

Our assessment using the AER’s augex model does not support PWC’s augex forecast. However, a
large portion (56%) of PWC’s augex forecast is due to a single zone substation project. If this project
was excluded from the assessment then the model would most likely provide stronger support for
PWC’s forecast.
The augex model predicts the modellable component of PWC’s augex over the next regulatory
period (2019/20 to 2023/23) to be $32 million (Real 2019) compared to PWC’s forecast for this
component of $50 million, representing a 64% reduction in the augex forecast.
However, as noted above, $28 million of PWC’s augex forecast is due to a single zone substation
project, the Wishart zone substation development. This project results in a large increase in forecast
augex in the second half of the next regulatory period (2021/22 to 2023/24), and it is this increase
that results in the large difference between model forecast and PWC’s.
This is shown in Figure E1 below, which compares the model’s augex forecast to PWC’s. This figure
shows both forecasts relative to PWC’s recent historical augex (as an average per annum) that has
been used to calibrate the augex model. The figure also shows the component of PWC’s forecast
that is due to the Wishart zone substation development.

1
Note, unlike the repex model which uses a five year calibration period, a four year period is used for the augex model
because of how the AER has defined the data that PWC should report in its Reset RIN.
2
Note, unlike with the repex model, the AER has not published benchmark parameters that would allow an inter-company
benchmark study to be produced.

Nuttall Consulting
Augex modelling report Page 3
Nuttall Consulting

18.0

16.0

14.0

12.0
Augex ($ millions)

10.0

8.0

6.0

4.0

2.0

0.0
2013-14 2014-15 2015-16 2016-17 2017-18 2018-19 2019-20 2020-21 2021-22 2022-23 2023-24
PWC forecast (ex Wishart ZSS) PWC forecast - Wishart ZSS PWC historical total (average pa) Augex model total

Figure E1 – Assessment results summary

This figure shows that the model’s forecast is significantly above the component of PWC’s forecast
when the augex associated with this project is excluded, with the model’s forecast over the next
regulatory period of $32 million well above PWC’s forecast for this component of $22 million.
It is important to note that this does not mean that the model assessment supports PWC’s forecast if
the Wishart project is excluded, as this would also require any component of demand growth
associated with this development to be excluded from the model along with any related
augmentation in the calibration period. We have not conducted this analysis. However, assuming
the reduction in demand growth associated with project is significantly less than the relative
reduction in augex then it seem reasonable to conclude that such an assessment would most likely
provide stronger support to PWC’s augex forecast.
It is also important to note that even if the Wishart project could be assessed within the augex
model, there could be valid reasons for the findings discussed above, given the form of business-as-
usual intra-company benchmark provided by the model. The model assessment finding can be
sensitive to individual projects, if their augex dominates the overall forecast. This appears to be the
case for PWC.
In these circumstances, the drivers and solutions for this single project – and its augex - could
incorporate factors that are not allowed for in this type of intra-company benchmark study.
Therefore, these matters would need to be investigated by the AER using its other assessment
approaches.

Nuttall Consulting
Augex modelling report Page 4
Nuttall Consulting

1 Introduction

1.1 Background and scope

Power and Water Corporation (PWC) has engaged us, Nuttall Consulting, to assist in its
preparations for its next regulatory decision by the Australian Energy Regulator (AER).
This decision will cover the five-year period from 1 July 2019 to 30 June 2024.
As part of this engagement, PWC has requested that we:
• develop a model of PWC’s augmentation capex (augex) using the AER’s augex
model
• use the model to assess PWC’s augex forecast, using an approach that could be
applied by the AER
• reconcile the model forecast with PWC’s own augmentation forecast
• prepare an independent report, which can be used as a supporting document to
PWC’s building block proposal to the AER, that sets out the forecast and explains
how we developed the model and forecast.
This document serves as the report to PWC indicated above.
The following definitions are used in this report:
• Augmentation capex (or augex) has the meaning given to it by the AER in its recent
advice on how it will conduct expenditure forecast assessments, which broadly
covers the demand-driven reinforcement, extension or enhancement of the
network, excluding similar activities due specifically to the connection of customers.
• We use the term AER augex model to mean the generic excel workbook that the
AER has advised it will use as an assessment technique in its determinations – and
the AER calls the augex model.
• We use the term PWC augex model to mean the model we have prepared of PWC’s
network using the AER augex model. The PWC augex model is used here to produce
augex forecasts of the PWC network.
• We use the term asset here in a very general sense to reflect the physical unit of
network that is accounted for in the AER augex model. This typically reflects an
individual line or an individual substation 3.
• When discussing the model and providing results in Section 2 and beyond, we will
use the year representation 200x, to represent the regulatory year 200x-1/200x.

3
Note the difference here to an asset in the repex model – or PWC’s systems – which is likely to account for a sub
component of the augex model’s asset.

Nuttall Consulting
Augex modelling report Page 5
Nuttall Consulting

In addition, all expenditure and costs shown in this report represent direct real June 2019
dollars.

1.2 Nuttall Consulting experience in this task

Nuttall Consulting, using Dr Brian Nuttall (the author of this report), developed the Excel
workbook that serves as the basis of the AER’s augex model and advised the AER on its
possible roles and application in regulatory determinations.
Moreover, we were engaged by the AER to provide advice that informed the AER’s past
determinations of the Victorian and Tasmanian Distribution Network Service Providers
(DNSPs). As part of these engagements, Dr Nuttall developed models and forecasts using
the AER’s repex model. Although the augex model is aimed at a different expenditure
activity (network augmentation, rather than asset replacement) it is broadly based upon
similar principles. We have been engaged by a number of DNSPs to assess their augex
forecasts using the augex model and an approach that is similar to that used by the AER
when applying the repex model to assess repex. This is the same approach we have used
here.

1.3 Key information sources

We have used the following key information to develop PWC’s augex model:
• the AER augex model and AER augex model handbook, published on the AER
website
• asset loading and rating data, provided in the format of the asset status tables in
Template 2.4 of the Reset Regulatory Information Notice (RIN); this data covers the
two years 2013/14 (2014) and 2017/18 (2018) 4
• PWC’s historical augex covering the period from 2013/14 to 2016/17, and PWC’s
forecast augex covering the period from 2017/18 (2018) to 2023/24 (2024), as
defined in Table 2.4.6 of the Reset RIN
• PWC’s forecast augmentation capacity added to its network covering the period
from 2017/18 (2018) to 2023/24 (2024), as defined in Table 2.4.6 of the Reset RIN.
We have also held a number of workshops with relevant PWC personnel to clarify data
requirements. Where gaps exist, we have made a number of assumptions based on advice
from PWC to prepare the models. The critical assumptions and their basis will be
discussed in this report.

1.4 Structure
This report is structured as follows:

4
Note, we understand that the data reporting for each of these two years corresponds to the asset loading and rating
entering these years.

Nuttall Consulting
Augex modelling report Page 6
Nuttall Consulting

• In section 2 we provide an overview of the AER augex model, summarising how it

develops a forecast, its inputs and outputs, and how the AER may use it to assess a
DNSP’s augmentation forecast.
• We discuss the methodology we have used to develop the PWC augex models in
Section 3.
• In Section 4 we explain the approach we have used to assess PWC’s augex forecast
using the augex model.
• Section 5 summarises and discusses the results of this assessment.

Nuttall Consulting
Augex modelling report Page 7
Nuttall Consulting

2 The AER’s augex model

Before explaining the development of PWC’s augex model, we first provide an overview of
the AER’s augex model and its application. This should help provide some context to the
results and discussions in the sections that follow.

2.1 Overview of augex model

The AER augex model is an Excel workbook, with a structure, formulas and VBA functions
and macros set up by the AER in order that it can be used by the AER to develop a network
model of a DNSP and use this to prepare augex forecasts.
The DNSP’s network is constructed within the AER augex model as a series of asset
populations. The model uses a probabilistic augmentation algorithm to make predictions
of augmentation needs for each population. The probabilistic augmentation algorithm
assumes that the maximum utilisation that an asset will reach before it must be
augmented (called its utilisation threshold in the model) is normally distributed across any
asset population represented within the model.
From this, the model predicts future augmentation volumes based upon a current
utilisation profile for an asset population represented in the model and forecast growth in
demand (which is used as a proxy for the forecast growth in utilisation).
The AER has indicated that it will use this model to make top-down assessments of a
DNSP’s augex forecast. In this regard, it has indicated that it may use the model in two
ways to develop a benchmark forecast:
1 Intra-company – it will develop a benchmark forecast within the model that reflects
the historical augmentation decisions of the DNSP (this reflects an assumption that
these decisions were prudent and efficient)
2 Inter-company – it will develop a benchmark forecast within the model that reflects
its view of the appropriate augmentation decisions it has determined from the set
of DNSPs (this reflects an assumption that the DNSP’s decisions may not have been
prudent and efficient, and so it has substituted its view on this matter from the
augex models of other DNSPs to test this).

It is important to stress that at this stage the AER has not published any of its analysis of
the above forms of benchmarking. As such, it is unclear how it may approach the
assessment of PWC’s augex forecast.
Importantly, it has not published any inter-company benchmark parameters for this
model. Therefore, we only discuss intra-company benchmarks in this report. As such, the
forecast produced by the model in this report only uses planning parameters that reflect
PWC’s recent historical augmentation and augex decisions, and assumes that these were
prudent and efficient decisions.

Nuttall Consulting
Augex modelling report Page 8
Nuttall Consulting

2.2 AER augex model form, inputs and output

2.2.1 Network specification inputs – network segments and groups
As indicated above, a DNSP’s network is defined as a series of distinct asset categories
within the augex model. These are called network segments in the AER’s documentation
and represent the set of network assets that may have similar planning arrangements i.e.
lines or substations.
To facilitate analysis and reporting, each network segment defined in the model is
assigned to a smaller set of groups. In this way, a model may use a large number of
network segments, to improve the accuracy of the analysis, but a much smaller number of
groups to provide aggregate forecasts for reporting (and benchmarking) purposes.

2.2.2 Network specification inputs - utilisation profile

A utilisation profile must be provided for each network segment used in the model. This
profile represents a snap-shot of the utilisation of the population of assets in that segment
for the initial year of the model. That is, the utilisation profile is essentially a vector that
holds the volume of assets (measured in capacity units e.g. MVA) at one-percentage
increments of utilisation.
The timing of a capacity-related augmentation is typically sensitive to the maximum
demand on an asset. That is, it is the amount of the maximum demand that is above
various capacity limits of an asset that defines the risks and/or service constraints
associated with using the asset. Therefore, within the augex model, the utilisation of any
asset (e.g. the utilisation of a line or substation) is defined as:
- the maximum demand on that asset / the assets capacity limit or rating.
The model itself does not define exactly how the measures of maximum demand or
capacity must be specified. However, the AER has indicated its preference for these
measures in an effort to place all DNSPs on a consistent basis 5, where:
• the maximum demand should be weather corrected to represent a 50% probability
of exceedance condition (and reflect normal network arrangements)
• the capacity of an asset should reflect its thermal rating, assuming a normal load
cycle if applicable (i.e. an asset’s normal cyclic rating).
It is important to note that once the units of capacity in a segment are defined, all
measures of utilisation, capacity being augmented, or capacity needing to be augmented
are reported in the model on that basis.

2.2.3 Network specification inputs – utilisation growth

To predict a network’s augmentation needs, the model must first predict what the
utilisation of the network will be in the future. To do this, the model requires the growth

5
See discussion in Section 5 of AER augex model manual.

Nuttall Consulting
Augex modelling report Page 9
Nuttall Consulting

in utilisation (assuming no augmentation) to be input for each network segment. This is

essentially the growth in maximum demand for each network segment.
The model represents this growth as a single annual compounded growth rate (percentage
growth in one year) that should represent the average annual growth rate over the period
being considered (note here that the model does not hold individual growth rates for each
year of the forecast period).

2.2.4 Planning parameters inputs

The model uses four planning parameters to define the approach it uses to predict future
augmentation needs:
• The utilisation threshold, which is represented as a normal probability distribution,
is defined by two of these parameters:
- the mean utilisation threshold
- the standard deviation of the utilisation threshold.
The utilisation threshold specifies when existing capacity requires augmentation,
and is used to measure this amount from the utilisation profile. In this way, this
parameter defines how the need for augmentation is measured.
• The capacity factor is the third parameter, reflecting the amount of additional
capacity that is added to the network, given the amount of existing capacity that
requires augmentation. It is defined as a proportion of the capacity requiring
augmentation.
For example, if the capacity factor is set at 50%, this means that if the model
calculates that 100 MVA of the existing capacity will require augmentation in the
future then it will assume that 50 MVA of capacity will be added to the network to
address that need.
This parameter relates to the scale, in capacity terms, of the augmentation solution
that is used to address a need.
• The fourth parameter reflects the average augmentation unit cost, where a unit is
specified in terms of the relevant unit of capacity for that network segment (i.e. $ /
kVA of capacity).
Using these parameters, the capacity added to the network (calculated via the utilisation
threshold and capacity factor) multiplied by the augmentation unit cost, produces the
expenditure forecast.

2.2.5 Model outputs

The model produces various outputs. These outputs provide various measures of the
input utilisation profile, such as average utilisation, average threshold, total quantity of
capacity, and total augmentation cost (i.e. quantity x augmentation unit cost).

Nuttall Consulting
Augex modelling report Page 10
Nuttall Consulting

The model also produces forecasts (by year over a 20-year period), including
augmentation capacity volumes, augmentation expenditure and average utilisation.
These outputs are provided at the network segment, segment group and total network
level. When averages are calculated at the network group or network level, the model
uses a weighted average using the augmentation cost of each asset category as the
weighting.

2.3 Calibration
The calibration of a DNSP’s model is the critical process that is applied by the AER (and us
for this assessment) to produce the intra-company benchmark model.
The calibration process concerns deriving the set of model planning parameters that
reflects the actual augmentation outcomes (volumes and expenditure) over the calibration
period.
The following process can be used to calibrate the augex model 6.
This process relies on calculating three parameters for each network segment (or segment
group) from the available data, namely:
• the augex in that segment (or segment group) over the calibration period
• the capacity added (through augmentation) in that segment (or segment group)
over the calibration period
• the capacity that required augmentation in that segment (or segment group) over
the calibration period.

2.3.1 Augmentation unit cost

The augmentation unit cost parameters for each segment is simply the augex divided by
the capacity added to the segment.

2.3.2 Volume planning parameters

The utilisation threshold parameters (mean and standard deviation) and capacity factor
for each segment need to be set to ensure the model reflects the capacity added (through
augmentation) over the calibration period.
However, the calculation of these planning parameters is more complicated because:
• we have three parameters to determine and typically only one variable (the total
capacity added)
• we are looking at history and not predicting into the future.

6
The AER augex model manual does not discuss the calibration process in any detail. However, we understand the AER
will apply a similar process to the one it has indicated it will use to calibrate its repex model. The process we have defined
here should reflect this similar process.

Nuttall Consulting
Augex modelling report Page 11
Nuttall Consulting

Therefore, the calibration of the utilisation threshold parameters is slightly more involved
and involves the following:
• First, in the absence of better information, the need to determine the standard
deviation is removed by making it dependent on the mean. We have assumed that
the standard deviation is the square root of the mean to reflect a similar
assumption the AER has advised it will use for the repex model calibration process.
• Second, the capacity factor is set at a specific value. There are various ways this
could be calculated. Here, PWC has estimated this parameter for our assessment
by analysing a sample of its recent augmentation projects.
• Third, an augex model is developed to reflect the beginning of the calibration
period, with the growth set to represent the growth that occurred over the
calibration period. The mean utilisation is determined within this model to ensure
that the forecast produced by the model over the calibration period equals actual
capacity added due to augmentations during the calibration period.
The above defines the process that will typically be applied.

Nuttall Consulting
Augex modelling report Page 12
Nuttall Consulting

3 PWC augex model development

3.1 Overview
As discussed in Section 2.3, the process to calibrate a model and prepare a forecast
requires the preparation of two augex models:
• The calibration model – This model is developed from the 2013/14 loading and
rating data. The planning parameters are calculated within this model to ensure the
forecast produced by the model to 2018 (i.e. capacity added and augex) matches
what actually occurred.
• The forecast model – This model is developed from the 2017/18 loading and rating
data. This model is used to prepare the forecasts over the next period, using the
planning parameters developed in the calibration model.
The development of these two models, including the parameter calibration process, is
discussed in this section.

3.2 Augex model development

3.2.1 Segmentation
The model produces forecasts for a set of network segments that represent the DNSP’s
network. As such, each segment defined in the model requires its own set of inputs (i.e.
utilisation profile and planning parameters) and the model produces forecasts for each
segment.
Segments have been developed that largely reflect those model categories, defined by the
AER in its Reset RIN, which are relevant to PWC. However, we have combined all HV
feeders in to a single HV feeder segment and all distribution substation types into a single
distribution substation segment to simplify the modelling process.
The table below summarises the groups and segments we have developed for the PWC
augex models.
Table 1 PWC augex model network segments
Network group Network segment
Sub-transmission lines All sub-transmission lines
Transmission substations
Transmission and zone substations
Zone substations
HV feeders All HV feeders
Distribution substations All distribution substations

Nuttall Consulting
Augex modelling report Page 13
Nuttall Consulting

3.2.2 Utilisation profiles

Utilisation definition
In the model, the utilisation of an asset (e.g. an HV feeder or zone substation) is defined
as:
Utilisation (%) = weather corrected peak demand (MVA) / asset rating (MVA).
For each segment, two utilisation profiles have been prepared reflecting the loading in
2013/14 and 2017/18. These profiles use the following asset ratings defined in the asset
status tables of template 2.4 of the Reset RIN.
Table 2 augex model asset rating definitions
Network type asset rating
Sub-transmission lines normal cyclic thermal rating
Transmission and zone substations transformer normal cyclic thermal rating
HV feeders normal thermal rating
Distribution substations normal cyclic thermal rating

It is important to note that any capacities referred to in this report as inputs or outputs of
the PWC augex model are measured on the above basis. This also includes any references
to utilisation and the augmentation unit costs.
Scaling of distribution substation ratings in the augex models
PWC has a material portion of distribution substations with a very high utilisation, which is
near or above the model’s maximum utilisation input limit (150%). Therefore, to ensure
that this limit does not affect our modelling, we have scaled the distribution rating by a
factor of two and performed all calibration and modelling using this scaling.
In our experience, there is nothing unusual in applying this scaling to PWC’s distribution
substations. We have applied similar scaling in the models we have prepared for all other
DNSPs. We do not consider that this scaling should have a material effect on the validity
or accuracy of the model’s forecast.
To avoid confusion, in the tabulated results presented in this report, we show unscaled
values in order that they can be readily interpreted by PWC. However, we also present
the scaled values in brackets in order that they can be reconciled to the model files.
Summary model inputs
The utilisation profiles need to be viewed through the augex model. However, to aid in
the validation of the model, the following table summarises some important parameters
associated with this set of profiles.

Nuttall Consulting
Augex modelling report Page 14
Nuttall Consulting

Table 3 Summary loading, rating and utilisation data in the augex models
Weather correct peak Asset capacity >100%
Asset capacity (MVA) Average utilisation (%)
Segment demand (MVA) utilisation (MVA)
2014 2018 2014 2018 2014 2018 2014 2018
All sub-transmission lines 698 747 2377 2579 29 29 0 0
transmission substations 404 437 945 1090 43 40 0 0
zone substations 337 304 710 755 47 40 0 0
All transmission and zone
741 741 1654 1844 45 40 0 0
substations
All HV feeders 509 571 1439 1533 35 37 25 18
a
All distribution substations 325 409 968 (1936) 1264 (2529) 34 (17) 32 (16) 38 (75) 42 (84)
a – brackets indicate distribution substation parameters, allowing for the rating scaling that is applied in the model

Nuttall Consulting
Augex modelling report Page 15
Nuttall Consulting

3.2.3 Load growth

For each segment, the growth in peak demand is an important input that drives the
forecast. The growth rate assumptions used in the two augex models (noted in the
introduction to this section) have been developed from the non–coincident summated
weather adjusted system annual maximum demand under 50% probability of exceedance
conditions (measured at the zone substation level) as reported in Table 3.4.3 of PWC’s RIN.
Based upon this data:
• for the calibration model, the compound annual growth rate from 2013/14 to
2017/18 is 1.7% per annum
• for the forecast model, the compound annual growth rate from 2017/18 to 2023/24
is 1.03 per annum.
To calculate the growth rate to be applied in the model for each segment, the overall
growth rate in peak demand can be considered to consist of two components:
• demand growth due to new connections associated with that segment that is
driving expenditure that PWC is allocated to the AER’s connections expenditure
category, which is considered to drive expenditure that is not related to the
utilisation of assets in this segment (and so should not be modelled)
• demand growth that is driving expenditure that PWC is allocating to the AER’s
augmentation expenditure category, which is considered to drive expenditure that
is related to the utilisation of assets in this segment (and so can be modelled).
The effects of the first of these components is not assessed through the augex model. The
effects of the second component can be assessed through the augex model, and define
the growth rates that are used in the augex model. We have estimated this component
from data PWC provided on the percentage split of the capacity added for each network
group (see the discussion in Section 3.3.2).
Table 5 below summarises the segment group growth rates used in the PWC augex model,
calculated using the methodology described above.
Table 4 Augex model growth rates
Segment Historical Forecast
All sub-transmission lines 1.70% 1.03%
All transmission and zone
1.70% 1.03%
substations
All HV feeders 1.53% 0.93%
All distribution substations 0.51% 0.31%

In appreciating the differences in growth rates between segments, it is worth noting that
distribution substations (and to a lesser extent, HV feeders) have a lower growth rate
because a large portion of growth in this segment is directly due to new customer
connections, and so is not factored into this model.

Nuttall Consulting
Augex modelling report Page 16
Nuttall Consulting

This should not affect the forecasting in the model because of how the model develops its
forecast, and the inherent assumption that this forecast reflects an intra-company
benchmark referenced back to PWC’s recent history. However, care should be taken in
using these growth rates for any other purpose or comparing them to growth rates
prepared for other purposes.

3.3 Model calibration

3.3.1 Historical calibration period
The historical calibration period reflects the 4-year period prior to the base year, but
inclusive of it. As such, the calibration period covers the commencement of 2013/14 to
the commencement of 2017/18. That is, the model is calibrated to reflect the
augmentations (i.e. the network-initiated capacity added and augex) that occurred from
2013/14 to 2016/17 inclusive.

3.3.2 Set up of calibration data

As discussed in Section 2.3, the initial phase in calibrating the augex model, involves
determining three parameters for each segment. The parameters reflect the
augmentations that have occurred over the calibration period, namely:
• the augex
• the incremental capacity added (because of demand-driven augmentations only)
• the capacity factor.
The table below summarises these parameters for each segment in the PWC augex model.
Table 5 Augex model calibration parameters
capacity added augex Capacity factor

Segment (MVA) $ (millions) (MVA)

All sub-transmission lines 182 6.3 0.50
transmission substations 131 5.7 0.50
zone substations 41 34.4 0.50
All transmission and zone
171 40.1 0.50
substations
All HV feeders 66 16.7 0.50
a
All distribution substations 30 (60) 0.9 0.50
a – brackets indicate distribution substation parameters, allowing for the rating scaling that is applied in the model

These parameters have been calculated using the following methodology and
assumptions.
Capacity added

Nuttall Consulting
Augex modelling report Page 17
Nuttall Consulting

The capacity added in each segment has been determined by first calculating the total
capacity added (for any reason) in each segment. This is calculated as the difference in the
total capacity for each segment from 2013/14 to 2017/18 (see segment totals in Table 4).
At the segment group level, we have assumed the proportion of capacity added for
different drivers, as shown in Table 6 below. The modellable proportions associated with
augex have been used to scale the associated segment total capacity added to produce
the capacity added defined in the table above. These assumptions have been discussed
with PWC.
Table 6 Capacity added to driver assumptions

greenfield/customer
demand growth with asset replacements, with
developments with
expenditure allocated to expenditure allocated to
Category expenditure allocated to
augex (utilisation repex (not utilisation
connections (not
relation and modellable) related)
utilisation related)

Sub transmission
90% 10% 0%
feeders
HV Feeders 70% 20% 10%
Zone substations 90% 10% 0%
DSS 10% 20% 70%

Augex
The augex parameters have been calculated directly from the historical augex defined in
Table 2.4.6 of the reset RIN. Where required, the capacity added in a segment has been
used to apportion augex defined at the segment group level to individual segments.
Capacity factors
To reduce the burden on PWC, we have assumed the capacity factors for use in each
segment group to be 0.5, based upon typical capacity factors we have used for similar
modelling exercises.

3.3.3 Determining planning parameters

The calibration of the planning parameters is performed using the calibration model. This
model is populated using the 2013/14 utilisation profiles and historical load growth, as
explained above. The planning parameters for each segment are calibrated to ensure the
calibration model outputs the parameters set out above (in Table 5).
This calibration process can be consider in two steps:
• calculating the unit cost
• calculating the utilisation threshold parameters.
These two steps are discussed in turn below.

Nuttall Consulting
Augex modelling report Page 18
Nuttall Consulting

3.3.3.1 Calculating the unit cost

We have calculated the augmentation unit costs ($/kVA) for each segment, directly using
the parameters shown in Table 6, whereby the unit cost is simply augex / modellable
capacity added.

3.3.3.2 Calculating the utilisation threshold parameters

The utilisation threshold for each segment is determined through the calibration model by
finding the threshold value that forces the model to forecast the capacity that was known
to have been added over the calibration period.
The following process has been used to apply this approach:
1 Input the unit cost and capacity factor planning parameters into the calibration
model.
2 Assume the standard deviation of the utilisation threshold, for each segment, is the
square root of the mean for that segment.
3 Using the model, determine the mean utilisation threshold parameter that sets the
model’s forecast of capacity added to the network over the calibration period to be
equal to the actual capacity added in the relevant segments. Excel’s goal seek
function is used for this purpose.

3.3.4 Scenario planning parameters

Table 7 below summarises the segment mean utilisation thresholds and unit costs used in
the PWC augex model, calculated using the calibration method described above.
Table 7 Augex model calibrated planning parameters
mean utilisation threshold unit cost
Segment (%) $/kVA
All sub-transmission lines 49.4 34.9
transmission substations 48.2 43.8
zone substations 69.8 848.3

All transmission and zone substations 57.8 234.4

All HV feeders 77.1 253.3

All distribution substationsa 74 (37) 29.3 (14.6)
a – brackets indicate distribution substation parameters, allowing for the rating scaling that is applied in the model

Nuttall Consulting
Augex modelling report Page 19
Nuttall Consulting

4 Augex forecast assessment

In this section we discuss our assessment of PWC’s augex forecast. In keeping with the
AER’s recent approach to the use of its repex model, this assessment is focused on the
aggregate augex forecast over PWC’s next regulatory period, 2019/20 (2020) to 2023/24
(2024).

4.1 Model assessment results

Table 9 summarises the PWC forecast and the comparable augex model forecast. The
results are provided as the total augex forecast over the next 5-year regulatory period.
Table 8 Augex model study results summary
Forecast 2019-20 - 2023-24 ($ millions, June-19)
PWC Augex model
49.9 31.9

18.0

16.0

14.0

12.0
Augex ($ millions)

10.0

8.0

6.0

4.0

2.0

0.0
2013-14 2014-15 2015-16 2016-17 2017-18 2018-19 2019-20 2020-21 2021-22 2022-23 2023-24
PWC forecast total PWC historical total (average pa) Augex model total

Figure 1 intra-company study results

The profile of PWC’s forecast augex compared to the model’s forecasts are shown in
Figure 1. This figure also shows the average per annum historical augex over the
calibration period.

4.2 Assessment discussion and conclusions

Our assessment using the AER’s augex model does not support PWC’s augex forecast.
However, a large portion (56%) of PWC’s augex forecast is due to a single zone substation

Nuttall Consulting
Augex modelling report Page 20
Nuttall Consulting

project. If this project was excluded from the assessment then the model would most
likely provide stronger support for PWC’s forecast.
The augex model predicts the modellable component of PWC’s augex over the next
regulatory period (2019/20 to 2023/23) to be $31.9 million compared to PWC’s forecast of
this component of $49.9 million, representing a 64% reduction in the augex forecast.
However, as noted above, $27.8 million of PWC’s augex forecast is due to a single zone
substation project, the Wishart zone substation development. PWC has advised that this
project is required to cater for the forecast demand in the Berrimah, Wishart and East Arm
areas, which is one of the areas in the NT where significant growth in demand is forecast.
This project results in a large increase in PWC’s forecast augex in the second half of the
next regulatory period (2021/22 to 2023/24), and it is this increase that results in the large
difference between model forecast and PWC’s.
This is shown in Figure 2 and Figure 3 below, which compares the model’s forecast to
PWC’s. These figure shows both forecasts relative to PWC’s recent historical augex (as an
average per annum) that has been used to calibrate the augex model. The figure also
shows the component of PWC’s forecast that is due to the Wishart zone substation
development. Figure 2 is the equivalent chart to Figure 1 above, showing the overall
assessment result. Figure 3 is a similar charts, but showing the results for the model’s
zone substation segment only.
18.0

16.0

14.0

12.0
Augex ($ millions)

10.0

8.0

6.0

4.0

2.0

0.0
2013-14 2014-15 2015-16 2016-17 2017-18 2018-19 2019-20 2020-21 2021-22 2022-23 2023-24
PWC forecast (ex Wishart ZSS) PWC forecast - Wishart ZSS PWC historical total (average pa) Augex model total

Figure 2 intra-company study results – with Wishart project augex

Nuttall Consulting
Augex modelling report Page 21
Nuttall Consulting

Zone substation segment

12.0

10.0

8.0
Augex ($ millions)

6.0

4.0

2.0

0.0
2013-14 2014-15 2015-16 2016-17 2017-18 2018-19 2019-20 2020-21 2021-22 2022-23 2023-24
PWC (ex Wishart ZZS) Wishart ZSS PWC ZSS historical average Model forecast

Figure 3 intra-company study results –zone substation segment only

These figures shows that the model’s forecast is significantly above the component of
PWC’s forecast when the augex associated with the Wishart project is excluded. The
model’s forecast over the next regulatory period of $31.9 million is well above PWC’s
forecast for this component of $22.1 million.
For the zone substation segment only, the difference between the model’s forecast and
PWC’s is more pronounced. With the Wishart project included, the model’s forecast over
the next regulatory period is $18.8 million below PWC’s forecast, with the model
forecasting only $13.1 million compared to PWC’s forecast of $32 million. This difference
largely reflects the overall assessment difference noted above. However, if the Wishart
project augex is excluded, the model forecast for the zone substation segment is
significantly higher than PWC’s forecast, with the model’s forecast remaining at
$13.1 million and PWC’s forecast dropping down to only $4.2 million.
It is important to note that this does not mean that the model assessment supports PWC’s
forecast if the Wishart project is excluded, as this would also require any component of
demand growth associated with this development to be excluded from the model along
with any related augmentation in the calibration period. We have not conducted this
analysis. However, assuming the reduction in demand growth associated with project is
significantly less than the relative reduction in augex then it seem reasonable to conclude
that such an assessment would most likely provide stronger support to PWC’s augex
forecast.
It is also important to note that even if the Wishart project could be assessed within the
augex model, there could be valid reasons for the findings discussed above, given the
form of business-as-usual intra-company benchmark provided by the model. The
assessment finding can be sensitive to individual projects, if their augex dominates the
overall forecast. This appears to be the case for PWC.
In these circumstances, the drivers and solutions for this single project – and its augex -
could incorporate factors that are not allowed for in this type of intra-company

Nuttall Consulting
Augex modelling report Page 22
Nuttall Consulting

benchmark study. This could be the case for PWC as it has advised that it deferred this
project from an originally proposed date of 2015, by constructing a 12 MVA temporary
substation, using the Nomad substation. We have not reviewed the accuracy or
reasonableness of this claim, but it could suggest a valid reason for the difference.
Therefore, these matters would need to be investigated by the AER using its other
assessment approaches, most notably a detailed engineering review.

4.3 Summary and conclusions

Our assessment using the AER’s augex model does not support PWC’s augex forecast, with
the model forecast at 64% of the equivalent PWC augex forecast.
However, the majority of the difference is associated with the modelling of PWC’s zone
substations, where a large portion (56%) of PWC’s augex forecast is due to a single zone
substation project – the Wishart zone substation development. If this project was
excluded from the assessment then the model would most likely provide stronger support
to PWC’s forecast.
The reason for this development and its effect on the assessment results could be due to
matters that are not allowed for in the assessment method. Therefore, these matters
would need to be investigated by the AER using its other assessment approaches (e.g.
detailed engineering review of the project).
We consider this assessment to be a reasonable top-down guide to PWC’s augmentation
needs, of which the results provide a form of independent regulatory challenge to PWC’s
own augex forecast. However, it is important to stress that the assessment used is only a
type of intra-company benchmark, and therefore, it inherently assumes that PWC’s
historical practices and augex were prudent and efficient. We have not tested the validity
of this assumption in our analysis.

Nuttall Consulting
Augex modelling report Page 23