ArticlePDF Available

A comparison of arrangements for increasing self-consumption and maximising the value of distributed photovoltaics on apartment buildings

Authors:

Abstract and Figures

As subsidised feed-in-tariffs for distributed photovoltaic (PV) generation are reduced or abolished in many jurisdictions, there is growing interest in increasing self-consumption to realise greater value from rooftop PV generation. However, deployment of PV on apartment buildings lags behind other residential deployment despite the potential advantages of load aggregation. We present a study of electricity and financial flows in ten 'virtual' Australian apartment buildings under a range of technical implementations and financial arrangements, using real load profiles and simulated generation profiles. Aggregation of diverse household and shared loads, either through an embedded network or 'behind the meter' of individual dwellings, can increase self-sufficiency and self-consumption of on-site generation compared to separate systems supplying common property or individual apartments. While embedded networks can enable access to more beneficial retail arrangements, behind the meter solutions may allow residents to avoid regulatory complexities and additional costs. The relative benefits of each arrangement are dependent on building characteristics and financial settings.
Content may be subject to copyright.
This is a preprint: original submitted version. The published version of the article Mike B. Roberts, Anna Bruce, Iain MacGill, A comparison
of arrangements for increasing self-consumption and maximising the value of distributed photovoltaics on apartment buildings, Solar
Energy, Volume 193, 2019, Pages 372-386, ISSN 0038-092X is available at https://doi.org/10.1016/j.solener.2019.09.067
1
A comparison of arrangements for increasing self-consumption and
1
maximising the value of distributed photovoltaics on apartment buildings
2
3
Mike B Robertsa,b,
1
, Anna Brucea,b and Iain MacGilla,c
4
aSchool of Photovoltaic and Renewable Energy Engineering
5
bCentre for Energy & Environmental Markets
6
cSchool of Electrical Engineering and Telecommunications,
7
University of New South Wales, Sydney 2052, Australia
8
Abstract
9
As subsidised feed-in-tariffs for distributed photovoltaic (PV) generation are reduced or
10
abolished in many jurisdictions, there is growing interest in increasing self-consumption to
11
realise greater value from rooftop PV generation. However, deployment of PV on apartment
12
buildings lags behind other residential deployment despite the potential advantages of load
13
aggregation. We present a study of electricity and financial flows in ten ‘virtual’ Australian
14
apartment buildings under a range of technical implementations and financial arrangements,
15
using real load profiles and simulated generation profiles. Aggregation of diverse household
16
and shared loads, either through an embedded network or ‘behind the meter’ of individual
17
dwellings, can increase self-sufficiency and self-consumption of on-site generation compared
18
to separate systems supplying common property or individual apartments. While embedded
19
networks can enable access to more beneficial retail arrangements, behind the meter solutions
20
may allow residents to avoid regulatory complexities and additional costs. The relative
21
benefits of each arrangement are dependent on building characteristics and financial settings.
22
Keywords
23
Photovoltaics, apartments, embedded network, residential electricity, load aggregation, self-
24
consumption.
25
1. Introduction
26
An estimated 53% of the world’s electricity is used in buildings (IEA, 2017a), half of it (27%)
27
in the residential sector (IEA, 2017b). The International Energy Agency (IEA) has estimated
28
that ‘virtually all’ residential and commercial buildings must achieve net-zero carbon
29
emissions by 2040 (IEA, 2016) if the world is to keep average global temperature rises due to
30
anthropogenic climate change to 1.5°C, in line with the targets in the COP 21 Paris
31
Agreement (United Nations Framework Convention on Climate Change (UN FCCC), 2015).
32
In Australia, although overall carbon emissions continue to rise (Australian Government
33
Department of Environment and Energy, 2018), the electricity industry is undergoing a
34
hesitant transition towards distributed generation and increased renewable energy. This is
35
most notable in the residential sector which leads the world in per-capita deployment of
36
Abbreviations: BAU, Business as Usual; BOM, Bureau of Meteorology; BTM, Behind the Meter; CES,
Community Energy Storage; CP, Common Property; EN, Embedded Network; ENO, Embedded Network
Operator; CREN, Community Renewable Energy Network; FiT, Feed-in Tariff; GST, Goods and Services Tax;
HVAC, Heating, ventilation and Air Conditioning; NEM, National Energy Market; NPV, Net Present Value;
NSW, New South Wales; PV, Photovoltaic; RET, Renewable Energy Target; SC, Self-Consumption; SS, Self-
Sufficiency; SGSC, Smart Grid Smart City; TOU, Time of Use; VB, Virtual Building
1
Corresponding author. E-mail address: m.roberts@unsw.edu.au
This is a preprint: original submitted version. The published version of the article Mike B. Roberts, Anna Bruce, Iain MacGill, A comparison
of arrangements for increasing self-consumption and maximising the value of distributed photovoltaics on apartment buildings, Solar
Energy, Volume 193, 2019, Pages 372-386, ISSN 0038-092X is available at https://doi.org/10.1016/j.solener.2019.09.067
2
rooftop photovoltaics (PV) with over 7GW installed on 1.8 million solar households,
37
representing a 22% penetration of homes nationally (Roberts, Copper, et al., 2018) and
38
exceeding 50% in some areas (International Energy Agency, 2018).
39
Initially, much of this rooftop deployment was supported by generous State Government
40
Feed-in Tariffs (FiTs). These have now been discontinued in most Australian jurisdictions (as
41
in many other countries), although a modest Federal subsidy remains through the Renewable
42
Energy Target (RET)(Australian Government Department of the Environment and Energy,
43
2017). Residential PV in Australia is net metered
2
, with self-consumed PV generation saving
44
the consumer their volumetric retail bundled (network and energy) tariff (21-34 c/kWh
45
(AEMC, 2016)) while PV exports to the grid are generally paid a FiT in the range 5-16
46
c/kWh, close to the wholesale market value of the electricity. A high level of solar resource,
47
rising retail electricity prices and falling PV system costs have made PV systems an attractive
48
investment for many households and kept residential PV deployment rates steady. The
49
significant disparity between the value of self-consumed versus exported PV generation (an
50
outcome of high, mainly volumetric, network costs that comprise almost half of total
51
residential electricity bills), and significant regulatory and pricing barriers to peer-to-peer
52
electricity trading create incentives for consumers to maximise self-consumption (SC) of on-
53
site PV generation (IEA-PVPS, 2016, Masson and Latour, 2012). The issue is also relevant to
54
network businesses, regulators and policy makers, given the technical challenges of high
55
residential PV exports, yet also revenue implications of higher SC. Similar circumstances
56
exist in many other countries, although regulatory approaches vary considerably (Jäger-
57
Waldau, Bucher, et al., 2018).
58
Given the interest in increasing SC by better aligning electricity consumption with PV
59
generation, it might be surprising that a recent review of techno-economic studies of
60
distributed PV (Sommerfeldt and Madani, 2017) found that two thirds of studies did not
61
consider building load at all. Now, however, an increasing number of articles ((Dehler, Keles,
62
et al., 2017, Hiesl, Lorenzi and Silva, 2016, Luthander, Widén, et al., 2015, Luthander,
63
Widén, et al., 2016, Nyholm, Goop, et al., 2016, Vieira, Moura, et al., 2017), for example)
64
explore the use of battery storage and/or demand management to increase SC for detached
65
houses or commercial premises.
66
Meanwhile, a growing population, changing demographics and urban planning strategies are
67
driving a sharp global increase in construction of apartment buildings. In Australia, an
68
increasing proportion of the population (10% in 2016 (ABS, 2016)) now lives in apartments
69
(or units
3
), while a 30% of new dwellings given building approval in 2017 were apartments
70
(ABS, 2018). Yet, a range of technical, organisational and regulatory challenges (Roberts,
71
Bruce, et al.) in many jurisdictions have hitherto largely prevented most apartment residents
72
from participating in the distributed energy transition or enjoying the financial and related
73
benefits of generating their own electricity.
74
Nevertheless, in an increasingly complex retail electricity market, there may be distinct
75
advantages to consumers who can jointly coordinate both their engagement with the market
76
and their use of distributed energy resources, including renewable generation and battery
77
2
In contrast to the use of this term in some other jurisdictions, “net metering” in Australia denotes an
arrangement whereby on-site generation is first applied to self-consumption, with only the excess exported to the
grid and metered, and where different tariffs are applied for consumption versus exports
3
The terms ‘unit’ and ‘apartment’ are used interchangeably in this paper.
This is a preprint: original submitted version. The published version of the article Mike B. Roberts, Anna Bruce, Iain MacGill, A comparison
of arrangements for increasing self-consumption and maximising the value of distributed photovoltaics on apartment buildings, Solar
Energy, Volume 193, 2019, Pages 372-386, ISSN 0038-092X is available at https://doi.org/10.1016/j.solener.2019.09.067
3
storage. As industry capacity for commercial rooftop PV installation develops, the large,
78
collectively owned roofs of apartment buildings provide opportunities for economies of scale,
79
whilst aggregation of diverse and physically proximate loads may create flatter profiles and
80
increase self-consumption (Roberts, Huxham, et al., 2016), as well as affording access to
81
more favourable retail arrangements and increasing independence from electricity retailers
82
and networks (an important driver in uptake of PV and storage (Agnew and Dargusch, 2017,
83
Agnew, Smith, et al., 2018)). A range of options are emerging to assist residents to utilize
84
these opportunities, as explained in Section 2.
85
A number of studies have explored the aggregation of load profiles, PV systems and storage
86
for multiple detached houses. Lopes, Martins, et al. (2016) found that aggregation of building
87
loads from three houses could increase self-sufficiency by 21% and self-consumption by 15%,
88
while Freitas, Reinhart, et al. (2018) explored the benefit of aggregating PV generation from
89
multiple rooftops with different orientations and tilt in reducing battery storage requirements.
90
Additionally, Community Energy Storage (CES) studies (Barbour, Parra, et al., 2018, Parra,
91
Swierczynski, et al., 2017) have examined the technical and economic benefits of applying
92
shared battery storage to multiple household loads.
93
Detailed analysis of SC in apartment buildings has been constrained, in part, by a lack of data
94
about real apartment electricity loads. Models of PV applied to multi-occupancy residential
95
buildings have utilised, for example, aggregated simulated load-profiles for standardised
96
three-bedroom homes (Musolino, Alet, et al., 2014), demand profiles based on multiples of an
97
energy services model for a single house load with a random time offset (Lang, Ammann, et
98
al., 2016) or the synthetic load profile for an average Finnish residential consumer (Van Roy,
99
Leemput, et al., 2014). Such studies, therefore, do not appropriately account for the particular
100
energy characteristics of apartment buildings.
101
Lang, Ammann, et al. (2016) found that PV profitability for four building types, including an
102
apartment building, was dependent on SC, which in turn was dependent on the ratio of PV
103
production to building load and on the shape of the daily load profile. Fina, Fleischhacker, et
104
al. (2018), using ten real 15-minute apartment load profiles (but no shared or common load),
105
found the jurisdictional regulatory environments in Germany and Austria greatly influenced
106
profitability. Sajjad, Manganelli, et al. (2015) found the benefit of net metering (application of
107
PV to the aggregated load) in realising the PV value for a 70-apartment building in Rome to
108
be primarily constrained by regulatory export limits, although the analysis excluded capital
109
costs of the PV. An article on optimal sizing of PV for Swedish apartment buildings
110
(Sommerfeldt and Madani, 2016) recommended sufficient PV capacity to supply 30% - 50%
111
of total load (for buildings with laundries) or 15% - 35% (for those without) to achieve SC of
112
60% - 80%. Also in Sweden, Kozarcanin and Andresen (2018) found that medium voltage
113
grids in multi-apartment areas can accommodate PV generation up to eight times local
114
consumption without overvoltage issues.
115
In this paper, we present results from a study of on-site PV generation and electricity
116
distribution in apartment buildings under a number of different possible technical and
117
commercial arrangements. Annual load profiles for 500 ‘virtual buildings’ (VBs) have been
118
created using real 30-minute load data from a dataset of thousands of apartments, and
119
common property
4
load data for ten apartment buildings of varying size and characteristics in
120
4
Common property (CP) load refers to the shared electricity services in an apartment building
This is a preprint: original submitted version. The published version of the article Mike B. Roberts, Anna Bruce, Iain MacGill, A comparison
of arrangements for increasing self-consumption and maximising the value of distributed photovoltaics on apartment buildings, Solar
Energy, Volume 193, 2019, Pages 372-386, ISSN 0038-092X is available at https://doi.org/10.1016/j.solener.2019.09.067
4
New South Wales (NSW), Australia. These have been combined with generation profiles
121
from PV systems designed to fit the real rooftops of the ten buildings, in order to explore the
122
SC achievable under different arrangements and the financial benefits of retrofitting PV to the
123
buildings under different commercial settings. The study is novel in its application of real
124
apartment and common load data to simulate 500 VBs, allowing a probabilistic assessment of
125
the value of different levels of PV and storage under entirely individual as well as shared
126
arrangements, as well as in its analysis of multiple implementation models. A realistic
127
assessment of PV generation potential based on the actual apartment building footprints that
128
correspond to the common property loads is also unique to this study, and allows us to
129
account for the different rooftop characteristics of a range of apartment building types.
130
The remainder of the paper is set out as follows: Section 2 gives a brief outline of the
131
technical options available for self-consumption of PV on apartment buildings, focusing on
132
Australian regulatory and market arrangements. Section 3 introduces the model, the data and
133
the method used for the study, while Section 4 explains the range of financial settings tested.
134
In Section 5 we present an analysis of the PV self-consumption and electricity self-sufficiency
135
achievable by applying PV to aggregated building loads, and in Section 6 we compare the
136
potential household cost savings under different scenarios. Finally, in Section 7, we present
137
our conclusions and make some suggestions for future investigations.
138
2. Technical Arrangements
139
Apartment building electricity loads combine individual unit loads with shared or common
140
property (CP) loads. The latter are highly dependent on building characteristics and available
141
services which may include lifts, carpark ventilation and lighting, water heating and pumping
142
for pools and centralised HVAC, water heating for apartments and lighting for stairwells,
143
corridors and other common areas. Most Australian apartment buildings are owned under
144
Strata Title, whereby the CP and building structure (including the roof) are owned and
145
managed by a strata body (the Owners’ Corporation or Body Corporate) on behalf of all
146
owners (Randolph and Easthope, 2007, Sherry, 2008).
147
This ownership system could be seen to lend itself to deployment of a shared PV system, on
148
shared property (the roof) to meet shared common property (CP) load (Figure 1(a)), and this
149
is the most common arrangement on Australian apartment buildings to date. Moreover, for
150
many (particularly high-rise) sites with relatively high CP loads, this arrangement is sufficient
151
to fully utilise the available roof area with 100% SC of generated PV. However, as 60% of
152
Australian apartments are in buildings of three storeys or less (ABS, 2016), potential PV
153
generation often exceeds CP demand (Roberts, Huxham, et al., 2016) and there may be
154
opportunities to use PV to also help offset individual apartment loads. Three possible
155
arrangements to facilitate this are shown in Figure 1 and described below.
156
Although installation of independent PV systems behind the meter (BTM) to meet individual
157
apartment loads, and potentially CP, (Figure 1(b)) is relatively straightforward for greenfield
158
sites, the per-Watt installation cost is likely to be higher than for a single larger PV system,
159
and siting individual assets on a shared roof space can create difficulties with permissions and
160
liability. Moreover, while this arrangement avoids the additional infrastructure requirements
161
and regulatory complexity of an embedded network (see below), it risks achieving lower
162
levels of SC and consequently reduced financial benefit for apartment owners and residents.
163
This is a preprint: original submitted version. The published version of the article Mike B. Roberts, Anna Bruce, Iain MacGill, A comparison
of arrangements for increasing self-consumption and maximising the value of distributed photovoltaics on apartment buildings, Solar
Energy, Volume 193, 2019, Pages 372-386, ISSN 0038-092X is available at https://doi.org/10.1016/j.solener.2019.09.067
5
(a) CP Only
(b) Individual BTM
(c) Embedded Network
(d) Shared BTM
Figure 1: Four possible arrangements for PV on apartment buildings:
Conversely, greater SC can be achieved by applying shared PV generation to the aggregated
164
load of a building. Distribution of electricity throughout the building can be organised via an
165
embedded network (EN) (Figure 1(c)), whereby the strata body or other party acting as an
166
‘Embedded Network Operator’ (ENO) purchases electricity from the grid via a ‘parent’ or
167
‘gateway’ meter and on-sells it (along with onsite PV generation), through a ‘child’ meter for
168
each apartment, to residents. This also allows the strata body to leverage the aggregated
169
demand across the building to access advantageous market arrangements (accessing tariffs
170
available to commercial large consumer rather than residential tariffs) for the grid-purchased
171
electricity.
172
At present, however, the Australian regulatory environment relating to energy retailing and
173
embedded network operation imposes a high administrative cost on establishing and operating
174
such an embedded network, while a recent regulatory review (AEMC, 2017) has
175
recommended EN operation be restricted to registered electricity retailers. This barrier, as
176
well as the potentially high capital costs of installation (see Section 4.3), may be avoided
177
through alternate arrangements for utilising a shared PV system behind the meter (BTM). In
178
this scenario (Figure 1(d)), a secondary metering arrangement is used to distribute the on-site
179
generation whilst residents continue to purchase off-site generation directly from an electricity
180
retailer, through their primary meter. This technical arrangement is new to the Australian
181
market and a number of tentative business models are emerging, either involving strata body
182
ownership of the PV and BTM distribution infrastructure, or ownership by a third party,
183
which sells PV generation to apartment residents (and to the strata body for CP load) under a
184
power purchase agreement (PPA).
185
The following section describes the model and dataset used to compare the technical and
186
financial benefits of these arrangements.
187
3. Method: Modelling Electricity Flows
188
3.1. morePVs model
189
This study utilises the Multi-Occupancy Residential Electricity with PV and Storage
190
(morePVs) tool, which models electricity flows and financial transactions in apartment
191
buildings under a range of technical and financial settings. The Python code for morePVs has
192
been made available open source
5
, for transparency and to allow replication of this study and
193
5
The model is available at https://github.com/mike-b-roberts/morePVs
This is a preprint: original submitted version. The published version of the article Mike B. Roberts, Anna Bruce, Iain MacGill, A comparison
of arrangements for increasing self-consumption and maximising the value of distributed photovoltaics on apartment buildings, Solar
Energy, Volume 193, 2019, Pages 372-386, ISSN 0038-092X is available at https://doi.org/10.1016/j.solener.2019.09.067
6
application to other datasets. Further extension of the model to include community energy
194
schemes, microgrids and peer-to-peer trading in a wider context, as well as addition of an
195
accessible graphical user interface, is underway.
196
Some initial results from the model have been presented previously, comparing the value of
197
PV for apartment buildings under a limited set of scenarios (Roberts, Bruce, et al., 2017) and
198
exploring the distribution of costs and benefits of PV in an embedded network (Roberts,
199
Bruce, et al., 2018). This study builds on the first of these publications, exploring a much
200
wider range of technical and financial arrangements.
201
Inputs to the model include building load profiles, comprising half-hourly load data for
202
apartments and common property over the course of a year, and generation profiles from PV
203
systems over the same period. Electricity flows and financial outcomes are modelled for
204
multiple scenarios with identified technical arrangements (as shown in Table 2), internal and
205
external tariffs, and system capital and operational costs.
206
Outputs include financial outcomes for individual households, retailers and network operators
207
under any of the four arrangements, and can be used to assess the distribution of costs and
208
benefits between different stakeholders within the building (Roberts, Bruce, et al., 2018). For
209
this study, however, we focus on the effect of different PV systems on electricity self-
210
consumption and self-sufficiency, and the combined cost savings for all building residents,
211
under a wide range of technical and financial settings.
212
3.2. Virtual building load data
213
In Australia, deployment of interval meters and ‘smart’ meters has been highly variable
214
between jurisdictions (Dickers, 2016, Victoria State Government, 2015) but, even where high
215
penetrations have been reached, there are considerable challenges to collection of 30-minute
216
electricity load data for all the households in an apartment building. Contestable markets for
217
electricity retailing and metering (AEMC, 2015a) mean that data access requires multiple
218
applications, having first obtained the consent of all householders. The challenges are
219
exacerbated by growing concerns about data privacy and general distrust of the electricity
220
sector (Energy Consumers Australia, 2018).
221
This study therefore uses the novel approach of creating ‘virtual buildings’ (VBs) by
222
combining common property load profiles and apartment load profiles from different sources.
223
A dataset of load profiles, including 2000 NSW apartments, collected for the AusGrid Smart
224
Grid Smart City (SGSC) trial in 2012–15, is publicly available. Details of the trial and the
225
resultant dataset are described in the various SGSC reports (Ausgrid, 2014a, Ausgrid, 2014b,
226
Ausgrid, 2014c, AusGrid, ARUP, et al., 2014), while characteristics of the apartment load
227
profiles are described in a forthcoming article (Roberts, Haghdadi, et al., (forthcoming)). For
228
this study, load profiles for the calendar year 2013 were selected and prepared using a method
229
which involves excluding apartments with more than 10% of missing data and filling gaps in
230
the resultant dataset with data from the timestamp in the period which has the most similar
231
load data across the dataset (Roberts, Haghdadi, et al., (forthcoming)).
232
The common property data, collected for apartment building energy audits at sites across
233
Sydney, is the subject of an earlier study (Roberts, Huxham, et al., 2016). Because the time
234
period of this data did not coincide with the apartment load profiles, buildings with any
235
evident temperature-dependent common load (such as HVAC or swimming pools) were
236
excluded from the dataset, leaving ten sites, ranging in height from three to twelve storeys.
237
This is a preprint: original submitted version. The published version of the article Mike B. Roberts, Anna Bruce, Iain MacGill, A comparison
of arrangements for increasing self-consumption and maximising the value of distributed photovoltaics on apartment buildings, Solar
Energy, Volume 193, 2019, Pages 372-386, ISSN 0038-092X is available at https://doi.org/10.1016/j.solener.2019.09.067
7
Virtual apartment buildings were created for each site by combining the actual CP profile for
238
the site with a randomly selected sample of SGSC apartment load profiles, one for each
239
apartment in the actual building. It was not possible to ‘match’ apartment load profiles to
240
buildings as only very minimal information on building characteristics, unit size or household
241
demographics is available for most of the apartments in the SGSC dataset. Therefore, for each
242
modelling scenario, load profiles were randomly selected 50 times to create 50 different VBs
243
for each site, and the average of the energy and financial metrics across all these buildings
244
were calculated.
245
Table 1 shows some characteristics of the ten sites as well as average values (across the 50
246
VBs for each site) for total annual load and the proportion of annual electricity consumption
247
serving the common property load, dubbed the ‘CP Ratio’. It can be seen that, although CP
248
ratio is generally higher in high rise buildings (in part, because of lift usage), there is a large
249
variation of values in both high and low-rise buildings, demonstrating the diversity of the
250
building stock and the range of common facilities available across the buildings.
251
Table 1: Virtual Building Characteristics: note that site tags summarise the number of apartments, floors
252
and proportion of load that is CP.
253
Apartments
Floors
Total Demand (MWh/year)
Average Total Daily Demand
(kWh/unit)
Mean CP Ratio
max_pv Capacity
(kWp)
max_pv Capacity
(kWp / unit)
max_pv PV Ratio %
PV Systems Modelled
(kWp/unit)
Self-Consumption of
max_pv for cp_only
Self-Consumption of
max_pv for btm_i_c
Self-Consumption of max_pv
for en_pv or shared btm
Avoided emissions (max_pv)
(kg CO2-e / unit / year)
208
12
1110
14.6
34%
47.3
0.23
5.8
max_pv
100%
92%
100%
240
161
7
900
15.3
38%
90.3
0.56
13.3
max_pv
89%
84%
100%
590
138
9
870
17.3
44%
42.3
0.31
6.5
max_pv
100%
93%
100%
320
104
8
850
22.4
57%
18.8
0.18
3.0
max_pv
100%
97%
100%
190
52
3
250
13.2
26%
141.5
2.72
77.9
1,1.5,2,2.5, max_pv
14%
36%
47%
2880
48
4
180
10.3
9%
52.5
1.09
36.9
0.5, 1, max_pv
9%
51%
77%
1100
44
4
180
11.2
17%
76.8
1.74
54.3
0.5 1.0,1.5, max_pv
13%
44%
63%
1760
34
4
180
14.5
33%
9.5
0.28
7.2
max_pv
94%
90%
100%
300
26
4
160
16.9
43%
78.5
3.02
67.1
1,1.5,2,2.5, max_pv
28%
45%
54%
3200
20
5
110
15.1
36%
31.5
1.58
41.2
0.5 1.0, max_pv
39%
59%
75%
1730
3.3. PV modelling and allocation to apartment loads
254
For each of the ten sites, a visual analysis of the roof area was carried out using multi-
255
viewpoint aerial imagery (Nearmap Ltd., 2015) to assess roof orientation and inclination, and
256
to identify obstructions and sources of shading. Access routes and areas shaded between 10am
257
and 2pm on the winter solstice were excluded from useable roof area, and the maximum PV
258
array (max_pv) for each site was designed by filling the usable area with panels arranged
259
flush to the roof. Table 1 shows the absolute and per-unit peak capacity of these PV systems.
260
Australian Bureau of Meteorology (BOM) satellite-derived irradiance data (Bureau of
261
Meteorology, 1990-2016) for 2013 at each site, along with temperature and wind speed from
262
This is a preprint: original submitted version. The published version of the article Mike B. Roberts, Anna Bruce, Iain MacGill, A comparison
of arrangements for increasing self-consumption and maximising the value of distributed photovoltaics on apartment buildings, Solar
Energy, Volume 193, 2019, Pages 372-386, ISSN 0038-092X is available at https://doi.org/10.1016/j.solener.2019.09.067
8
the nearest automatic weather station, were used as inputs for the NREL’s System Advisor
263
Model (SAM) (Balir, Gilman, et al., 2013, NREL, 2010) PV Watts module, to give simulated
264
half-hourly PV production for each array over 2013. For each site, the ‘PV Ratio’ was
265
calculated as the ratio of annual generation of the max_pv array to annual building load,
266
averaged across all VBs (see Table 1).
267
An analysis of the impact of PV system size on self-consumption and self-sufficiency under
268
different technical arrangements for each site (Section 5) was carried out by scaling the output
269
of the max_pv system to smaller system sizes in steps of 5kW down to zero. However, for the
270
financial analysis (Sections 4 and 6), for those sites where the max_pv was equivalent to 1kW
271
per apartment or more (See Table 1), smaller systems (in steps of 500W / apartment) were
272
designed by successively excluding roof areas with the lowest insolation, a more realistic
273
approach for a real-world building.
274
For each technical arrangement, PV generation was allocated to apartment and common
275
property loads as shown in Table 2. For the common property only (cp_only) arrangement,
276
the PV was applied only to the common property load, with any excess exported to the grid.
277
For embedded network with PV (en_pv), the PV was treated as a single system connected
278
between the parent and child meters, with the PV energy netted off the aggregate building
279
load. For behind the meter, individual-CP (btm_i_c), a percentage of the total PV capacity
280
equal to the CP Ratio (the percentage contribution of the annual CP load to the annual
281
building load) was first allocated to supply CP, with the remaining capacity allocated equally
282
to separate systems for each apartment, while for btm_i_u (individual-units), the total PV
283
capacity was shared equally between the apartments only. For shared behind the meter
284
arrangements, the PV energy was distributed between all units (btm_s_u and btm_p_u, for
285
strata-owned and PPA arrangements, respectively – Section 4.6) or between apartment and
286
CP loads (btm_s_c and btm_p_c) in proportion to instantaneous demand, with the export of
287
any excess generation allocated in the same proportion.
288
Table 2: PV system configurations for virtual buildings
289
Abbreviation
PV System Configuration
Allocation of PV Generation
To CP
To Units
bau
Business as Usual (no PV)
N/A
cp_only
PV supplying Common Property
only
100%
None
btm_i_u
Individual PV behind the meter
(units only)
None
Equal share of total
capacity
btm_i_c
Individual PV behind the meter
(units + CP)
Proportional to CP Ratio
Equal share of
remaining capacity
btm_s_u
btm_p_u
Shared PV behind the meter
(units only), strata owned or PPA
None
Proportional to
instantaneous load
btm_s_c
btm_p_c
Shared PV behind the meter
(units + CP), strata owned or PPA
Proportional to instantaneous load
en
Embedded network without PV
N/A
en_pv
Embedded network with PV
Proportional to instantaneous load6
6
For en_pv, there is no distinction between energy generated from PV or imported from the grid in assessing
outcomes for the building.
This is a preprint: original submitted version. The published version of the article Mike B. Roberts, Anna Bruce, Iain MacGill, A comparison
of arrangements for increasing self-consumption and maximising the value of distributed photovoltaics on apartment buildings, Solar
Energy, Volume 193, 2019, Pages 372-386, ISSN 0038-092X is available at https://doi.org/10.1016/j.solener.2019.09.067
9
3.4. Energy metrics
290
Luthander, Widén, et al. (2015) define Self-Consumption (SC) and Self-Sufficiency (SS) as
291
the overlapping part of the generation and load profiles as a proportion of the total generation
292
and total load, respectively, as shown in Equation(1) and Equation(2), where L(t) and G(t) are
293
the instantaneous load and generation. This formulation is useful as it is easily adapted to
294
ensure the correct treatment of efficiency losses in a system with a battery energy storage
295
system (BESS) and we include it here for consistency with our own work in this area
296
(Roberts, Bruce, et al., (forthcoming)).
297
!" #$%&'()*+,-.*+,/0+
$.*+,0+
Equation(1)
!! #$%&'()*+,-.*+,/0+
$)*+,0+
Equation(2)
However, where there is no BESS (or significant energy losses), and load and PV are
298
measured in discrete time periods ti, (determined by the temporal resolution of the relevant
299
meters), the measurable SC can be expressed as the proportion of total annual on-site PV
300
generation that is consumed within the building, and self-sufficiency (SS) as the proportion of
301
the total annual building load that is met by on-site generation. In our model, if the total PV
302
generation in the ith half-hourly time interval ti is given by Gtot(ti) = Gsc(ti)+ Gexport(ti), where
303
Gsc(ti) and Gexport(ti) are the self-consumed generation and exported generation respectively,
304
and the total building load is given by Ltot(ti) = Gsc(ti) + Limport(ti), where Gsc(ti) is the building
305
load met by PV generation and Limport(ti) is the load met by grid import, then the self-
306
consumption and self-sufficiency are given by Equation(3) and Equation(4) respectively.
307
!" #1234*56,
789:;
6<7
12=>=*56,
789:;
6<7 ?@AABC #C12=>=*56,D12EFG>H=*56,
789:;
7<7 C
789:;
6<7 12=>=*56,
789:;
6<7 C? @AAB
Equation(3)
!! #1234*56,
789:;
6<7
1I=>=*56,
789:;
6<7 ?@AABC #C1I=>=*56,D1I6JG>H=*56,
789:;
7<7 C
789:;
6<7 1I=>=*56,
789:;
6<7 C? @AAB
Equation(4)
Note that these measurable metrics are likely to be higher than the true SC and SS as ti
308
increases, because any non-simultaneous imports and exports within the half hour time
309
interval are treated as simultaneous (Marshall, Bruce, et al.).
310
The potential emissions reductions from the modelled PV systems were calculated by
311
multiplying the current indirect (Scope 2) emissions factor for consumption of electricity
312
purchased from the grid in New South Wales (0.83 kgCO2-e/kWh (Department of
313
Environment and Energy, 2017)) by the expected annual energy generation from the system,
314
and subtracting the estimated embodied carbon emissions from the manufacture, installation,
315
operation and decommissioning of the PV system (0.045 kgCO2-e/kWh (Hsu, O’Donoughue,
316
et al., 2012)). These avoided emissions for the maximum PV at each site are shown in Table
317
1.
318
4. Method: Financial Modelling
319
4.1. Financial settings and metrics
320
As commonly used in the literature (Sommerfeldt and Madani, 2017), the Net Present Value
321
(NPV) of annual savings compared to business as usual (BAU) was used to compare the
322
benefits of different technical arrangements and PV system sizes, calculated under a range of
323
financial settings. Note that for this study, the modelling was used to assess overall outcomes
324
for the building, without consideration of the distribution of financial benefits between
325
This is a preprint: original submitted version. The published version of the article Mike B. Roberts, Anna Bruce, Iain MacGill, A comparison
of arrangements for increasing self-consumption and maximising the value of distributed photovoltaics on apartment buildings, Solar
Energy, Volume 193, 2019, Pages 372-386, ISSN 0038-092X is available at https://doi.org/10.1016/j.solener.2019.09.067
10
residents, owners and the strata body. For transparency, and to avoid dependence on arbitrary
326
projections of future electricity costs, only a single year of operation was modelled with
327
capital expenditure amortised at a discount rate of 6% to calculate monthly repayments.
328
In general, the Net Present Value (NPV) of an initial investment CO after time T is defined as
329
the sum of cashflows Ft for each time period t and is given by Equation(5) for a discount rate
330
d.
331
KLM #CN O5
*@P0,5
Q
5RS TCUV
Equation(5)
If Et,s and Ot,s are the electricity cost and operating cost in period t for scenario s and
UV-W
is the
332
capital cost for that scenario, the NPV relative to the Business as Usual (BAU) scenario is
333
given by Equation(6).
334
Typical residential PV installations in NSW can currently expect payback within five to seven
335
years (Martin, 2017) and there is evidence (Altmann, 2013, Roberts, Bruce, et al., 2019) that
336
the high turnover of apartment owners (ABS, 2016) means that the cost of a sustainability
337
upgrade to an apartment building must be recouped within a few years if it is to be attractive
338
to owners purely on the basis of energy cost savings. However, as discussed in Section 4.2,
339
the lifetime of a PV system (allowing for inverter replacement) is likely to exceed 25 years,
340
which is also a typical mortgage period for purchasing an apartment. The model was therefore
341
tested for sensitivity to amortisation periods for the capital expenditure of 5, 10, 15, 20 and 25
342
years.
343
4.2. PV capital and operating costs
344
Average installed costs (after Federal government subsidies and Federal goods and services
345
tax (GST) of 10%) for residential and commercial PV installations in NSW of $1.01 to $1.84
7
346
per Watt (Solar Choice, 2018) were used to calculate the capital costs of the PV systems.
347
PV module lifetimes of 25 years have been assumed, with inverter replacement after ten years
348
included at between $0.31/W and $1.10/W. Other operating costs, including replacement of
349
electrical balance-of-system components and occasional cleaning, are likely to be low in
350
comparison to decreases in inverter costs, and have therefore been omitted from this study.
351
Full details are given in Appendix A.
352
4.3. Embedded network capital and operating costs
353
In Australia, the cost of installing electrical supply and metering infrastructure for greenfield
354
apartment buildings is borne by the developer, who then gifts the equipment to the
355
distribution network service provider (DNSP). For an embedded network on a greenfield site,
356
capital costs are similar to those for multiple grid connections, plus installation of a parent or
357
gateway meter at a cost of around $2000 (Roberts, Personal conversations with embedded
358
network operators, 2017) with ownership passed to the strata body. However, retrofitted
359
embedded networks are likely to require the additional cost of replacing child meters, as even
360
where suitable interval meters have already been installed, they are owned by the DNSP
361
7
All costs in this paper are AU$ where AU$1.00 = US$ 0.754 (Bloomberg, 2018)
KLMW#CNX5-YZ[TX5-W T\5-W
*@P 0,5
Q
5RS TCUV-W
Equation(6)
This is a preprint: original submitted version. The published version of the article Mike B. Roberts, Anna Bruce, Iain MacGill, A comparison
of arrangements for increasing self-consumption and maximising the value of distributed photovoltaics on apartment buildings, Solar
Energy, Volume 193, 2019, Pages 372-386, ISSN 0038-092X is available at https://doi.org/10.1016/j.solener.2019.09.067
11
which has no regulatory obligation to transfer them to the ENO and may, indeed, have an
362
incentive to resist EN installation. Similarly, DNSP requirements, as well as building
363
characteristics, will influence the level of additional expenditure that may be required for
364
upgrading outdated wiring, switchboards or meter rooms. For this study, sensitivity to capital
365
costs has been modelled using three settings (capex_low, capex_med and capex_high) with
366
costs of $2,000, $20,000 and $50,000 per site, respectively, in addition to $400 per unit for
367
the child meter installations.
368
The operational costs of an electricity retailer in the energy market have been estimated by the
369
Australian Competition and Consumer Commission (ACCC) to be $230 (15% of the average
370
customer bill) per average NSW residential customer (ACCC, 2017). In an EN, these costs
371
(meter-reading, billing, marketing, customer-assistance and risk management for bad
372
customer debts) are borne by the ENO with an additional cost for maintenance of the EN. As
373
actual EN operating costs are obscured by commercial confidentiality, a value of $250 per
374
customer has been chosen to align with these estimated retailer costs. A rule requiring ENOs
375
to appoint an Embedded Network Manager (ENM) with responsibility for facilitating EN
376
customers wishing to leave the EN and access the retail market (AEMC, 2015b) has recently
377
come into force but, as there is uncertainty in the possible exemptions to this requirement and
378
the costs associated with appointing an ENM, it has not been possible to include this
379
additional OPEX cost in the study.
380
4.4. Customer tariff structure and rates
381
With hundreds of retail tariffs available in Australia’s National Electricity Market (NEM),
382
defining ‘Business as Usual’ is not without its challenges. Retail tariffs represent bundled
383
network tariffs and energy costs, along with various additional costs including environmental
384
levies and retailer costs and margins. Two types of tariffs are available to residential
385
customers of one of the 70 electricity retailers (AER, 2018) operating in the NEM. Standing
386
Offer tariffs, once regulated but now set by retailers, generally offer higher rates but greater
387
price certainty and some additional consumer protections compared to the market retail offers
388
taken by the majority of customers. The total bill of a ‘representative customer’ on a
389
‘representative market offer’ in NSW in 2016 was equivalent to a 15% discount off the
390
standing offer tariff (AEMC, 2016). Therefore, for the Business-as-Usual (BAU) and BTM
391
scenarios in this study, all customers were assumed to be paying their retailer a market tariff
392
equivalent to the 2017 standing offer Time of Use (TOU) tariff from the ‘Retailer of Last
393
Resort’ in the relevant network area for the case study (Energy Australia, 2017), with a 15%
394
discount applied to all fixed and volumetric components.
395
All NSW solar feed-in tariffs (FiTs) are currently flat rate and in the range 6-17 c/kWh
396
depending on the particular retailer and general tariff offering, with the majority below the
397
state regulator’s ‘all time benchmark’ rate which is 12.8 c/kWh for 2017-18 and set to drop to
398
8-9 c/kWh for 2018-19 (IPART, 2018). For this study, the sensitivity of PV value to the
399
current FiT rate was tested using FiTs of 12c, 8c and even zero for all customers.
400
4.5. Tariffs at the parent meter
401
A key driver for embedded networks is that the size of an aggregated building load is likely to
402
trigger access to a commercial ‘large energy consumer’ tariff (comprising a regulated network
403
component and a market retail energy component) at the parent meter, with typical rates lower
404
than residential tariffs. In the relevant network area, the network component, which is
405
This is a preprint: original submitted version. The published version of the article Mike B. Roberts, Anna Bruce, Iain MacGill, A comparison
of arrangements for increasing self-consumption and maximising the value of distributed photovoltaics on apartment buildings, Solar
Energy, Volume 193, 2019, Pages 372-386, ISSN 0038-092X is available at https://doi.org/10.1016/j.solener.2019.09.067
12
dependent on the voltage level, has a relatively high ratio of fixed and capacity to volumetric
406
charges, with the daily capacity charge based on the customer’s peak load in the preceding
407
12-month period. The energy component, determined by negotiation with the retailer and
408
therefore subject to a high degree of variability and to a lack of transparency, is likely to be
409
significantly lower than the estimated 14.63 c/kWh paid by a representative NSW residential
410
retail customer (AEMC, 2018) in 2017/18, and to include a TOU component. We have tested
411
sensitivity to a range of high (‘TOU12’) and low (‘TOU9’) market prices from early 2018
412
(combined with the applicable network tariff) and to FiTs of zero, 8 c/kWh and 12 c/kWh at
413
the parent meter. Full details of these parent tariffs are given in Appendix B.
414
4.6. Costs for PV shared behind the meter
415
Installation of a PV system configured to share generation behind the meter (Figure 1(c)) will
416
also require capital expenditure for distribution and metering infrastructure and will incur
417
operating costs for meter reading, billing, and related services. We modelled two financial
418
arrangements currently being trialled in Australian jurisdictions. For the first (btm_s_u and
419
btm_s_c), capital costs of $4950 per site and $490 per customer are added to the PV capital
420
costs and a monthly metering and billing charge of $4.99 per customer is added to the total
421
bill for the building. These costs could be shared equally between all solar customers or
422
covered by the strata body. In the second arrangement (btm_p_u and btm_p_c), these costs (as
423
well as the PV capital cost) are met by the solar retailer (Allume, 2016) who leases the
424
roofspace from the strata body at no cost. Each customer then enters a rolling power purchase
425
agreement (PPA) with the solar retailer for the purchase of self-consumed generation at
426
around 18c + GST / kWh, and for exported generation at the customer’s FiT rate.
427
4.7. Cost of avoided emissions
428
Although outcomes in the Australian electricity market are currently distorted by the lack of a
429
carbon pricing mechanism, the country is already experiencing the effects of climate change
430
(CSIRO and Bureau of Meteorology, 2015) and future climate mitigation and adaptation
431
represents a real societal cost. We have therefore included the NPV of avoided emissions
432
using a carbon price of AUD 79.53
8
/tCO2, being the mid-point of the range for a 2020 carbon-
433
price consistent with the Paris temperature target as reported by the High Level Commission
434
on Carbon Prices (Stern and Stiglitz, 2017).
435
8
USD 60.00 (Bloomberg, 2018)
This is a preprint: original submitted version. The published version of the article Mike B. Roberts, Anna Bruce, Iain MacGill, A comparison
of arrangements for increasing self-consumption and maximising the value of distributed photovoltaics on apartment buildings, Solar
Energy, Volume 193, 2019, Pages 372-386, ISSN 0038-092X is available at https://doi.org/10.1016/j.solener.2019.09.067
13
5. Results: Electricity Flows
436
5.1. Import and export
437
Figure 2 illustrates the effect of aggregation on PV self-consumption over three summer days
438
for a typical VB (at site a44_f4_cp17 in Table 1), showing building load under BAU as well
439
as for the largest possible PV system under different arrangements. Utilising the PV to meet
440
only CP load (comprising 17% of the total building load) only reduces imports slightly and
441
allows significant export, even on a cloudy day. Connecting 17% of the PV capacity to the CP
442
load and connecting an individual 1.4 kWp system to each apartment significantly reduces
443
import and export (as shown by the blue line) but still results in daily periods of simultaneous
444
import and export. Finally, the generation of a single PV system shared between CP and
445
apartments according to instantaneous load (regardless of whether behind the meter or
446
through an embedded network), substantially reduces imports and exports as PV generation is
447
applied to meet aggregated internal building loads.
448
5.2. Self-consumption and self-sufficiency
449
Recall that Table 1 shows the average SC of generation from the maximum PV system at each
450
site, with the PV applied to CP load only, to CP and unit loads as individual BTM systems, or
451
to the aggregated building load, either through an embedded network or behind the meter.
452
Note that, for buildings with a low PV ratio, while SC of 100% can be achieved through
453
applying the PV to aggregated load, this can also be achieved by supplying CP only, while
454
using the available roof space for individual systems to supply apartments as well (btm_i_c)
455
may reduce overall SC. However, where there is a high PV ratio and low CP ratio (a situation
456
common to many older low-rise apartment buildings), SC can be increased up to eight-fold by
457
applying the PV generation to aggregated load rather than CP only.
458
Figure 3(a) and (b) show the variation of SC and SS with PV system size up to the maximum
459
rooftop capacity of each VB, with the PV applied to the aggregated building load. For sites
460
with sufficient roof space, shared PV systems between 1.5 kWp and 3 kWp per apartment can
461
provide around a third of the annual load. Some of the variation in these metrics between sites
462
Figure 2: Total import and export for site a44_f4_cp17 with Max PV (76.75kWp) over two days
This is a preprint: original submitted version. The published version of the article Mike B. Roberts, Anna Bruce, Iain MacGill, A comparison
of arrangements for increasing self-consumption and maximising the value of distributed photovoltaics on apartment buildings, Solar
Energy, Volume 193, 2019, Pages 372-386, ISSN 0038-092X is available at https://doi.org/10.1016/j.solener.2019.09.067
14
is due to differences in the yield of the PV systems, which has been taken as proportional to
463
the yield of the maximum system for each site and is therefore dependent on the roof form
464
and shading of the building. However, it is also possible to discern a trend of increasing SC
465
with increasing CP Ratio, as CP typically has a higher proportion of daytime load than
466
apartments.
467
Figure 3(c) shows the improvement in both SC and SS that can be achieved through applying
468
a shared PV system (whether BTM or through an EN) compared to individual PV systems for
469
units and common property. Note that for individual systems, SS and SC are highly dependent
470
on CP ratio, with the greatest benefits of shared PV evident for three or four storey buildings
471
with a low CP ratio (for example, a48_f4_cp09 and a44_f4_cp17).
472
(a)
(b)
(c)
Figure 3: PV System (a) self-consumption and (b) self-sufficiency when applied to
aggregated building loads, and (c) compared with application to individual loads for
each of the VBs up to their maximum system size per unit
This is a preprint: original submitted version. The published version of the article Mike B. Roberts, Anna Bruce, Iain MacGill, A comparison
of arrangements for increasing self-consumption and maximising the value of distributed photovoltaics on apartment buildings, Solar
Energy, Volume 193, 2019, Pages 372-386, ISSN 0038-092X is available at https://doi.org/10.1016/j.solener.2019.09.067
15
6. Results & Discussion: Financial
473
6.1. Technical arrangements compared
474
Figure 4 shows the NPV of savings relative to BAU averaged across all households in all VBs
475
with the maximum possible PV system installed at each site, under different technical
476
arrangements and with the financial settings shown.
477
Without consideration of avoided carbon emissions, the addition of pv_max applied to CP
478
load has a positive NPV for all sites, but for sites with low CP ratio / high PV ratio the cost
479
savings are marginal. The addition of Maximum PV as individual systems for units and CP
480
(btm_i_c) is less beneficial than CP only for all sites, and, for sites with low CP ratio / high
481
PV ratio, worse than BAU due to the low SC achieved.
482
Figure 4: NPV of average annual household savings under different
technical arrangements for the VBs (ordered by decreasing number
of apartments) with max_pv
Figure 5: NPV of average annual
household savings with 1.0kWp
/unit for the five VBs that have
sufficient roof space for the PV
It can be seen that for larger sites, particularly those with over 100 apartments, the most
483
significant savings are associated with embedded networks. EN benefits are derived from the
484
reduced energy and network tariffs accessed by aggregating electricity loads at a single point
485
of grid connection, and so are sensitive to the (non-regulated) energy component of the tariff.
486
Conversely, for the smallest building, the capex and opex costs of an EN (in the capex_med
487
scenario) exceed the benefits, even when mitigated by the addition of PV. For the (mostly
488
high- or medium-rise) buildings with low PV ratio / high CP ratio, distribution of shared PV
489
BTM is less beneficial than for low-rise buildings and is worse than BAU for some sites.
490
In Figure 5, which shows the same information for PV systems sized at 1.0 kWp/unit for the
491
five sites that have sufficient roof capacity, the benefits of sharing PV BTM over individual
492
systems are more consistently evident. In all the scenarios shown in both charts, the addition
493
of PV to an EN (as discussed in Section 6.2) increases the savings.
494
This is a preprint: original submitted version. The published version of the article Mike B. Roberts, Anna Bruce, Iain MacGill, A comparison
of arrangements for increasing self-consumption and maximising the value of distributed photovoltaics on apartment buildings, Solar
Energy, Volume 193, 2019, Pages 372-386, ISSN 0038-092X is available at https://doi.org/10.1016/j.solener.2019.09.067
16
If the value of avoided carbon emissions, as described in Section 4.7, are added to the savings,
495
the benefits of adding PV are evident in all scenarios, although insufficient to create a positive
496
NPV for shared BTM systems at the minority of low PV ratio / high CP ratio sites, or for an
497
EN at the smallest site unless maximum PV is applied.
498
6.2. Embedded networks
499
Figure 6(a) shows the dependence of the benefits of an EN (without PV) on the parent tariff
500
and the amortisation period. While large buildings show clear benefit from an EN under most
501
scenarios, short term cost recovery and a high parent tariff can undermine the benefits. Note,
502
however, that the aggregated loads of larger buildings are also more likely to attract
503
favourable retail tariffs at the parent meter than those for smaller buildings. Figure 6(b) shows
504
the range (indicated by the box tails) of average annual household savings across the 50 VBs
505
at each site (but does not include variability between households within each VB). Significant
506
savings can be accessed through retrofitting an EN, even where installation costs are high, if
507
these capex costs can be shared between 100 or more apartments, while for buildings with
508
fewer than 50 apartments the viability of an EN is highly building-specific, with sensitivity to
509
the EN installation costs as well as to the demand characteristics of the building. At these
510
sites, the operating costs, including ENM fees, also become critical to EN viability. However,
511
it is important to note that, for greenfield buildings, the capital costs of an embedded network
512
may not be any higher than for an ‘on market’ arrangement, which would make the EN a
513
favourable option in a wider range of scenarios than for brownfield sites.
514
(a)
(b)
Figure 6: NPV of average annual houshold savings for EN without PV, showing (a) effect of retail
tariff and amortisation period and (b) variability across VBs and between capex scenarios
Adding the maximum possible PV to the EN marginally reduces costs for all sites if capital
515
expenditure is amortised over 20 years (Figure 4), but for the sites with fewer apartments and
516
higher PV ratio, this is dependent on the FiT available at the parent meter, without which the
517
PV system size must be constrained to increase SC and reduce costs. Figure 7 demonstrates
518
that removing the FiT for max_pv systems with low SC can negate EN savings, while
519
increasing it to 12 c/kWh increases benefits. The significance of the negotiated retail tariff is
520
also shown, with the greatest benefits of reduced tariffs accruing to sites with high average
521
demand (see Table 1).
522
This is a preprint: original submitted version. The published version of the article Mike B. Roberts, Anna Bruce, Iain MacGill, A comparison
of arrangements for increasing self-consumption and maximising the value of distributed photovoltaics on apartment buildings, Solar
Energy, Volume 193, 2019, Pages 372-386, ISSN 0038-092X is available at https://doi.org/10.1016/j.solener.2019.09.067
17
Figure 7: NPV of Savings for EN with max PV under different parent tariff and FiT scenarios
523
For the sites with the roof capacity to allow PV systems of 1 kWp/unit or larger, Figure 8
524
shows the financial impact of different PV capacities installed in an EN for different
525
investment periods. For all these sites, adding PV reduces NPV if capital costs must be repaid
526
in five years, but over 10 years or longer, PV systems between 500 W and 1 kW per
527
apartment (likely to achieve 80% self-consumption) can reduce overall costs for all sites, even
528
in the absence of a FiT at the parent meter. With a FiT comparable to the retail component of
529
the import tariff available at the parent meter, larger systems (up to 3 kWp/unit) can be cost
530
neutral or better over 20 years, bearing in mind that, except for the smallest building with 20
531
apartments, these benefits are additional to EN savings.
532
Figure 8: Effect of PV size and financial settings on NPV for EN
6.3. Shared PV behind the meter
533
Figure 9 shows that the benefit of individual PV systems serving all units and CP decreases
534
for systems above 0.5 kWp/unit for these 5 sites in the absence of a FiT (and even with a FiT
535
for some sites). However, a shared BTM arrangement benefits from the reduced capital $/W
536
cost of a larger system (offset against the additional capital cost of the BTM distribution and
537
metering infrastructure) as well as the increased SC afforded by aggregating loads. In the
538
absence of a FiT, the aggregation benefits of a shared system in reducing SC are most
539
pronounced for systems above 1 kWp/unit, but benefits of PV are reduced for both individual
540
and shared arrangements, so systems above 1.0-1.5 kWp/unit (if capex is repaid over ten
541
years) or 1.5-2.0 kWp/unit (if repaid over 20 years) are not financially attractive. However, in
542
This is a preprint: original submitted version. The published version of the article Mike B. Roberts, Anna Bruce, Iain MacGill, A comparison
of arrangements for increasing self-consumption and maximising the value of distributed photovoltaics on apartment buildings, Solar
Energy, Volume 193, 2019, Pages 372-386, ISSN 0038-092X is available at https://doi.org/10.1016/j.solener.2019.09.067
18
the presence of an 8 c/kWh FiT, shared systems up to 2.5 kWp/unit (where there is available
543
roof space) show significant benefit even over 10 years. It is worth noting, however, that if
544
some scale economies can be realised by installing multiple individual PV systems or a large
545
microinverter system with individual panels applied to different apartment loads, the benefits
546
of sharing the PV are reduced.
547
Figure 9 also presents the NPV of average annual savings if purchasing shared PV generation
548
through a Solar PPA (at 18c/kWh for energy consumed, see Section 4.6) which is preferable
549
to the upfront purchase over the short term and in the absence of a FiT (with the solar retailer
550
therefore making a loss), but will sustain a profit for the retailer over the longer term if a FiT
551
is available.
552
As would be expected, considering that CP often has greater daytime load than apartments,
553
savings for BTM PV systems (whether shared or individual) supplying CP as well as
554
apartments are greater in all scenarios than for systems supplying apartments only, except for
555
the site with a CP Ratio of only 9%, where NPV is equal for some settings.
556
Figure 9: NPV of annual savings for shared BTM and individual BTM for different FiTs and investment
terms
7. Conclusions
557
The findings of our study suggest that the increased self-consumption and self-sufficiency
558
brought about by application of a shared PV system to aggregated loads across an apartment
559
building can be translated into financial benefits for households, in the right circumstances.
560
Behind the meter sharing of PV systems sized at 1.0–1.5 kWp/unit offers a means of avoiding
561
the regulatory complexities of embedded networks and can be preferable to individual
562
systems, provided economies of scale reduce capital costs and the strata body or a third-party
563
retailer is able to manage the risks associated with tariff uncertainty and resident turnover
564
over the longer term. However, larger systems are dependent on ongoing availability of export
565
FiTs for profitability.
566
This is a preprint: original submitted version. The published version of the article Mike B. Roberts, Anna Bruce, Iain MacGill, A comparison
of arrangements for increasing self-consumption and maximising the value of distributed photovoltaics on apartment buildings, Solar
Energy, Volume 193, 2019, Pages 372-386, ISSN 0038-092X is available at https://doi.org/10.1016/j.solener.2019.09.067
19
For buildings with between 40 and 100 apartments, the relative benefits of individual, EN and
567
BTM arrangements were shown to be highly sensitive to building-specific factors including
568
installation costs, load profile and financial settings. Although the ten buildings modelled
569
cover a diverse range of apartment numbers, floors and CP load, we cannot draw definitive
570
conclusions for the broader building stock from this limited building set. However, the results
571
suggest that embedded networks, allowing access to commercial retail tariffs, are likely to be
572
the most financially beneficial option for buildings with 100 or more apartments, if the
573
regulatory barriers can be overcome. While the most favourable use of PV for these (and for
574
smaller sites with a low PV ratio) is to apply it to CP load only, adding shared PV sized at
575
1.0-1.5 kWp/unit to an EN also shows increased benefits, even allowing for capex recovery
576
over 10 years if a FiT is available and over 20 years with no FiT.
577
With regulatory changes ongoing in Australia (AEMC, 2017), it is not clear how the
578
additional ENM requirement will affect EN costs, nor whether requirements for ENOs to
579
become authorised electricity retailers will disproportionately benefit the large, incumbent
580
retailers who may have little incentive to support collective RE deployment (Roberts, Passey,
581
et al., 2017).
582
As BTM battery storage becomes more prevalent in the residential market, it would be useful
583
to explore how the addition of shared storage to the PV arrangements discussed above could
584
increase self-consumption and self-sufficiency and provide opportunities to reduce volumetric
585
and/or demand charges, particularly under an EN arrangement, and how this would affect the
586
value of shared PV. Note also that the results presented here only address total benefits across
587
the building, and that these benefits will be shared unevenly between customers with different
588
load characteristics and between the owner occupiers, landlords and tenants of the apartments,
589
while the financial settings will also have implications for the profitability of retailers, ENOs
590
and DNSPs. There is a need for detailed analysis of the distribution of costs and benefits
591
between these different stakeholders under a range of financial settings, in order to fully
592
understand the incentives for consumers under these communal arrangements.
593
Acknowledgements
594
The authors wish to thank Gareth Huxham of EnergySmart Strata for the provision of the
595
common property load data used in the study. We gratefully acknowledge support provided
596
for this research by a grant from Energy Consumers Australia, a studentship from the CRC for
597
Low Carbon Living and an Australian Government Research Training Program scholarship.
598
Declaration of interests: none.
599
References
600
ABS. Census of Population and Housing. 2016; Available from:
601
https://auth.censusdata.abs.gov.au/webapi/jsf/dataCatalogueExplorer.xhtml.
602
ABS. Building Approvals March 2018. 2018 Accessed; Available from:
603
http://www.abs.gov.au/AUSSTATS/abs@.nsf/DetailsPage/8731.0Mar%202018?Open
604
Document.
605
ACCC. ACCC Retail Electricity Pricing Inquiry - Preliminary Report. 2017.
606
AEMC, Final Rule Determination: National Electricity Amendment & National Energy Retail
607
Amendment (Expanding competition in metering and related services) Rule 2015.
608
2015a.
609
AEMC, National Electricity Amendment (Embedded Networks) Rule 2015 No. 15, A.E.M.
610
Commission, Editor. 2015b.
611
This is a preprint: original submitted version. The published version of the article Mike B. Roberts, Anna Bruce, Iain MacGill, A comparison
of arrangements for increasing self-consumption and maximising the value of distributed photovoltaics on apartment buildings, Solar
Energy, Volume 193, 2019, Pages 372-386, ISSN 0038-092X is available at https://doi.org/10.1016/j.solener.2019.09.067
20
AEMC. 2016 Residential Electricity Price Trends (Final Report). 2016.
612
AEMC. Review of Regulatory Arrangements for Embedded Networks - Final Report. 2017.
613
AEMC. 2017 Residential Electricity Price Trends. 2018.
614
AER. Public register of authorised retailers & authorisation applications. 2018 Accessed:
615
4/5/2018; Available from: https://www.aer.gov.au/retail-
616
markets/authorisations/public-register-of-authorised-retailers-authorisation-
617
applications?f%5B0%5D=field_accc_aer_sector%3A4.
618
Agnew, S. and P. Dargusch, Consumer preferences for household-level battery energy
619
storage. Renewable and Sustainable Energy Reviews, 2017. 75: p. 609-617.
620
Agnew, S., C. Smith and P. Dargusch, Causal loop modelling of residential solar and battery
621
adoption dynamics: A case study of Queensland, Australia. Journal of Cleaner
622
Production, 2018. 172: p. 2363-2373.
623
Allume. Allume Energy. 2016 Accessed: 17/10/16; Available from:
624
http://www.allumeenergy.com.au/.
625
Altmann, E., Apartments, Co-ownership and Sustainability: Implementation Barriers for
626
Retrofitting the Built Environment. Journal of Environmental Policy & Planning,
627
2013. 16(4): p. 437-457.
628
Ausgrid. Customer Applications: SGSC Technical Compendium. 2014a.
629
Ausgrid. Smart Metering Infrastructure: SGSC Technical Compendium. 2014b.
630
Ausgrid. Modelling Inputs Compendium: SGSC Technical Compendium. 2014c.
631
AusGrid, ARUP, Energeia, Frontier Economics, Institute for Sustainable Futures and
632
University of Technology Sydney. Smart Grid, Smart City: Shaping Australia's
633
Energy Future - Executive Report. 2014.
634
Ausgrid. ES7 Network Price Guide (version 1.8). 2017a April 2017.
635
Ausgrid. Network Price List 2017-2018. 2017b; Available from:
636
https://www.ausgrid.com.au/Common/Industry/Regulation/Network-prices/Price-lists-
637
and-policy.aspx.
638
Australian Government Department of Environment and Energy. Quarterly Update of
639
Australia's National Greenhouse Gas Inventory: December 2017. 2018 Accessed;
640
Available from: http://www.environment.gov.au/climate-change/climate-science-
641
data/greenhouse-gas-measurement/publications/quarterly-update-australias-national-
642
greenhouse-gas-inventory-dec-2017.
643
Australian Government Department of the Environment and Energy. The Renewable Energy
644
Target (RET) scheme. 2017 Accessed: 29/9/2018; Available from:
645
http://www.environment.gov.au/climate-change/government/renewable-energy-target-
646
scheme.
647
Balir, N.J., P. Gilman and A. Dobos. Comparison of Photovoltaic Models in the System
648
Advisor Model. 2013.
649
Barbour, E., D. Parra, Z. Awwad and M.C. González, Community energy storage: A smart
650
choice for the smart grid? Applied Energy, 2018. 212: p. 489-497.
651
Bloomberg. Market Currencies: 1AUD = 0.754USD. 2018 Accessed: 20/07/2018; Available
652
from: https://www.bloomberg.com/quote/AUDUSD:CUR.
653
Branker, K., M.J.M. Pathak and J.M. Pearce, A review of solar photovoltaic levelized cost of
654
electricity. Renewable and Sustainable Energy Reviews, 2011. 15(9): p. 4470-4482.
655
Bureau of Meteorology, Australian Hourly Solar Irradiance Gridded Data. 1990-2016,
656
BOM.
657
This is a preprint: original submitted version. The published version of the article Mike B. Roberts, Anna Bruce, Iain MacGill, A comparison
of arrangements for increasing self-consumption and maximising the value of distributed photovoltaics on apartment buildings, Solar
Energy, Volume 193, 2019, Pages 372-386, ISSN 0038-092X is available at https://doi.org/10.1016/j.solener.2019.09.067
21
CSIRO and Bureau of Meteorology. Climate Change in Australia Information for Australia’s
658
Natural Resource Management Regions: Technical Report. 2015.
659
Dehler, J., D. Keles, T. Telsnig, B. Fleischer, M. Baumann, D. Fraboulet, A. Faure-Schuyer
660
and W. Fichtner, Self-Consumption of Electricity from Renewable Sources, in
661
Europe's Energy Transition - Insights for Policy Making. 2017. p. 225-236.
662
Department of Environment and Energy. National Greenhouse Accounts Factors July 2017.
663
2017.
664
Dickers, J. Australia’s first market-led smart meter rollout. Utility Magazine, 2016.
665
Energy Australia, Energy Price Fact Sheet ENE391090SR: Basic Home Standing offer Tariff
666
- 5 Day Time of Use. 2017.
667
Energy Consumers Australia. Energy Consumer Sentiment Survey June 2018. 2018.
668
Fina, B., A. Fleischhacker, H. Auer and G. Lettner, Economic Assessment and Business
669
Models of Rooftop Photovoltaic Systems in Multiapartment Buildings: Case Studies
670
for Austria and Germany. Journal of Renewable Energy, 2018. 2018: p. 1-16.
671
Freitas, S., C. Reinhart and M.C. Brito, Minimizing storage needs for large scale
672
photovoltaics in the urban environment. Solar Energy, 2018. 159: p. 375-389.
673
Fronius Australia Pty. Ltd. Fronius Warranties. 2018 Accessed: 2/5/2018; Available from:
674
https://www.fronius.com/en-au/australia/photovoltaics/products/all-
675
products/solutions/fronius-service-solutions/fronius-warranties/fronius-warranties.
676
Heacox, E., Inverter Cost Analysis:The Right Metrics For Improved Paybackload, in Solar
677
Industry Magazine. 2010.
678
Hiesl, A., On the economics of decentralized battery-supported photovoltaic systems.
679
Hsu, D.D., P. O’Donoughue, V. Fthenakis, G.A. Heath, H.C. Kim, P. Sawyer, J.K. Choi and
680
D.E. Turney, Life cycle greenhouse gas emissions of crystalline silicon photovoltaic
681
electricity generation. Journal of Industrial Ecology, 2012. 16(s1): p. S122-S135.
682
IEA. World Energy Outlook. 2016.
683
IEA. World Energy Outlook. 2017a.
684
IEA. Key World Energy Statistics. 2017b.
685
IEA-PVPS. A Methodology for the Analysis of PV Self-Consumption Policies. 2016.
686
International Energy Agency. IEA-PVPS Annual Report 2017. 2018.
687
IPART. Solar Feed-In Tariffs Issue Paper. 2018; Available from:
688
https://www.ipart.nsw.gov.au/files/sharedassets/website/shared-files/pricing-reviews-
689
energy-services-publications-solar-feed-in-tariffs-201819/fact-sheet-solar-feed-in-
690
tariffs-201819-march-2018.pdf.
691
Jäger-Waldau, A., C. Bucher, K.H.B. Frederiksen, R. Guerro-Lemus, G. Massone, B. Mather,
692
C. Mayr, D. Moneta, J. Nikoletatos and M.B. Roberts, Self-consumption of electricity
693
produced from PV systems in apartment buildings Comparison of the situation in
694
Australia, Austria, Denmark, Germany, Greece, Italy, Spain, Switzerland and the
695
USA, in The 7th World Conference on Photovoltaic Power Conversion. 2018:
696
Waikaloa.
697
Kozarcanin, S. and G.B. Andresen, Grid integration of solar PV for multi-apartment
698
buildings. International Journal of Sustainable Energy Planning and Management,
699
2018. 17.
700
Lang, T., D. Ammann and B. Girod, Profitability in absence of subsidies: A techno-economic
701
analysis of rooftop photovoltaic self-consumption in residential and commercial
702
buildings. Renewable Energy, 2016. 87: p. 77-87.
703
This is a preprint: original submitted version. The published version of the article Mike B. Roberts, Anna Bruce, Iain MacGill, A comparison
of arrangements for increasing self-consumption and maximising the value of distributed photovoltaics on apartment buildings, Solar
Energy, Volume 193, 2019, Pages 372-386, ISSN 0038-092X is available at https://doi.org/10.1016/j.solener.2019.09.067
22
Lopes, R.A., J. Martins, D. Aelenei and C.P. Lima, A cooperative net zero energy community
704
to improve load matching. Renewable Energy, 2016. 93: p. 1-13.
705
Lorenzi, G. and C.A.S. Silva, Comparing demand response and battery storage to optimize
706
self-consumption in PV systems. Applied Energy, 2016. 180: p. 524-535.
707
Luthander, R., J. Widén, D. Nilsson and J. Palm, Photovoltaic self-consumption in buildings:
708
A review. Applied Energy, 2015. 142: p. 80-94.
709
Luthander, R., J. Widén, J. Munkhammar and D. Lingfors, Self-consumption enhancement
710
and peak shaving of residential photovoltaics using storage and curtailment. Energy,
711
2016. 112: p. 221-231.
712
Marshall, L., A. Bruce and I. MacGill, Allocation Rules and Meter Timing Issues in Local
713
Energy or Peer to Peer Networks.
714
Martin, J. 10kW solar power systems: Price, power output, and returns. 2017 Accessed:
715
19/5/2017; Available from: https://www.solarchoice.net.au/blog/10kw-solar-power-
716
systems-price-power-output-and-returns/.
717
Masson, G. and M. Latour, Self-consumption as the new Holy Grail of the PV industry: From
718
theory to reality. Photovoltaics International, 2012. 17.
719
Musolino, V., P. Alet, Y. Regamey, A. Hutter, L. Perret-Aebi and C. Ballif, Multi-occupancy
720
buildings as micro-grids: an asset for integrating photovoltaics in power systems, in
721
Renewable Power Generation Conference. 2014.
722
Navigant Consulting Inc. A Review of PV Inverter Technology Cost and Performance
723
Projections. 2006.
724
Nearmap Ltd. Nearmap. 2015 Accessed: 6/11/2017; Available from: http://au.nearmap.com.
725
NREL. System Advisor Model (SAM). 2010 Accessed: 18/10/2017; Available from:
726
https://sam.nrel.gov/.
727
Nyholm, E., J. Goop, M. Odenberger and F. Johnsson, Solar photovoltaic-battery systems in
728
Swedish households – Self-consumption and self-sufficiency. Applied Energy, 2016.
729
183: p. 148-159.
730
Parra, D., M. Swierczynski, D.I. Stroe, S.A. Norman, A. Abdon, J. Worlitschek, T.
731
O’Doherty, L. Rodrigues, M. Gillott, X. Zhang, C. Bauer and M.K. Patel, An
732
interdisciplinary review of energy storage for communities: Challenges and
733
perspectives. Renewable and Sustainable Energy Reviews, 2017. 79: p. 730-749.
734
Randolph, B. and H. Easthope, ‘Governing the compact city’: The governance of strata title
735
developments in Sydney, in International Conference of Sustainable Urban Areas.
736
2007: Rotterdam.
737
Roberts, M.B., G. Huxham, A. Bruce and I. MacGill, Using PV to help meet common
738
property energy demand in residential apartment buildings, in Australian Summer
739
Study in Energy Productivity. 2016: Sydney.
740
Roberts, M.B., A. Bruce and I. MacGill, PV for Apartment Buildings: Which Side of the
741
Meter?, in Asia Pacific Solar Research Conference. 2017: Melbourne.
742
Roberts, M.B., R. Passey, A. Bruce and I. MacGill. APVI / CEEM Submission to AEMC draft
743
report on embedded networks. 2017; Available from:
744
https://www.aemc.gov.au/sites/default/files/content/5928335d-6666-4835-b543-
745
299c593d5672/MarketReview-Submission-RPR0006-CEEM-%28UNSW%29-APVI-
746
171018.PDF.
747
Roberts, M.B., A. Bruce and I. MacGill, Collective Prosumerism - Accessing the Potential of
748
Embedded Networks to Increase the Deployment of Distributed Generation on
749
Australian Apartment Buildings, in EnergyCon. 2018: Cyprus.
750
This is a preprint: original submitted version. The published version of the article Mike B. Roberts, Anna Bruce, Iain MacGill, A comparison
of arrangements for increasing self-consumption and maximising the value of distributed photovoltaics on apartment buildings, Solar
Energy, Volume 193, 2019, Pages 372-386, ISSN 0038-092X is available at https://doi.org/10.1016/j.solener.2019.09.067
23
Roberts, M.B., J. Copper, A. Bruce, T. Barton, N. Haghdadi and R. Hu. Solar Trends Report
751
2018. 2018.
752
Roberts, M.B., A. Bruce and I. MacGill, Opportunities and barriers for photovoltaics on
753
multi-unit residential buildings: Reviewing the Australian experience. Renewable and
754
Sustainable Energy Reviews, 2019. 102.
755
Roberts, M.B., A. Bruce and I. MacGill, The impact of aggregated battery storage on
756
photovoltaic self-consumption and customer value in apartment buildings.
757
(forthcoming).
758
Roberts, M.B., N. Haghdadi, A. Bruce and I. MacGill, Cluster-based characterisation of
759
Australian apartment electricity demand and its implications for low-carbon cities.
760
(forthcoming).
761
Sajjad, I.A., M. Manganelli, L. Martirano, RobertoNapoli, G. Chicco and G. Parise, Net
762
Metering Benefits for Residential Buildings: A Case Study in Italy, in IEEE 15th
763
International Conference on Environment and Electrical Engineering (EEEIC),. 2015.
764
Sherry, C., The Legal Fundamentals of High Rise Buildings and Master Planned Estates.
765
Australian Property Law Journal, 2008. 16(1): p. 1-23.
766
Sol Distribution. 10 Year Warranty Offer on SMA Sunny Boy Inverters. 2017 28/11/2017
767
Accessed: 4/5/2018; Available from: http://blog.sol-distribution.com.au/10-year-
768
warranty-offer-on-sma-sunny-boy-inverters/.
769
Solar Choice. Residential and Commercial Power System Prices: March 2018. 2018
770
Accessed: 2/5/2018; Available from: https://www.solarchoice.net.au/blog/solar-
771
power-system-prices.
772
Solar Juice, Platinum (Wholesale) Price List March 2018. 2018.
773
Sommerfeldt, N. and H. Madani, Solar PV for Swedish Prosumers – a Comprehensive
774
Techno-Economic Analysis, in Proceedings of EuroSun2016. 2016. p. 1-9.
775
Sommerfeldt, N. and H. Madani, Revisiting the techno-economic analysis process for
776
building-mounted, grid-connected solar photovoltaic systems: Part one – Review.
777
Renewable and Sustainable Energy Reviews, 2017. 74: p. 1379-1393.
778
Stern, N. and J.E. Stiglitz. Report of the High-Level Commission on Carbon Prices. 2017.
779
United Nations Framework Convention on Climate Change (UN FCCC).
780
FCCC/CP/2015/L.9/Rev.1 - Adoption of the Paris Agreement. 2015; Available from:
781
http://unfccc.int/resource/docs/2015/cop21/eng/l09r01.pdf.
782
Van Roy, J., N. Leemput, F. Geth, R. Salenbien, J. Buscher and J. Driesen, Apartment
783
Building Electricity System Impact of Operational Electric Vehicle Charging
784
Strategies. IEEE Transactions on Sustainable Energy, 2014. 5(1): p. 264-272.
785
Victoria State Government. Smart meters: End of Rollout. 2015 28/3/2015 Accessed:
786
31/8/2017; Available from: http://www.smartmeters.vic.gov.au/about-smart-
787
meters/end-of-rollout.
788
Vieira, F.M., P.S. Moura and A.T. de Almeida, Energy storage system for self-consumption of
789
photovoltaic energy in residential zero energy buildings. Renewable Energy, 2017.
790
103: p. 308-320.
791
792
This is a preprint: original submitted version. The published version of the article Mike B. Roberts, Anna Bruce, Iain MacGill, A comparison
of arrangements for increasing self-consumption and maximising the value of distributed photovoltaics on apartment buildings, Solar
Energy, Volume 193, 2019, Pages 372-386, ISSN 0038-092X is available at https://doi.org/10.1016/j.solener.2019.09.067
24
Appendix A: PV Capital Costs
793
Average installed costs (after Federal government subsidies and Federal goods and services
794
tax (GST) of 10%) for residential and commercial PV installations in NSW were used to
795
calculate the capital costs of the PV systems (Solar Choice, 2018) and are shown in Table 3
9
.
796
Although PV panels have expected lifetimes exceeding 25 years (Branker, Pathak, et al.,
797
2011) and 25-year warrantees are now common, inverter replacement is likely during the
798
lifetime of the system and constitutes the largest ongoing cost for a PV system (Branker,
799
Pathak, et al., 2011). In line with other studies (Branker, Pathak, et al., 2011, Heacox, 2010),
800
and given the availability of ten-year warrantees from some manufacturers (2018, Sol
801
Distribution, 2017), a lifetime of ten years for inverters has been assumed, with replacement
802
inverter costs based on median wholesale prices (Solar Juice, 2018) in each range plus
803
estimated installation costs, and a typical retail margin of 24% plus GST. It should be noted
804
that although inverter lifetimes significantly greater than ten years are unlikely to be cost
805
effective in the near future, inverter costs are likely to continue to fall as manufacturing
806
volumes increase (Navigant Consulting Inc, 2006). Other operating costs, including
807
replacement of electrical balance-of-system components and occasional cleaning, are likely to
808
be low in comparison to decreases in inverter costs, and have been omitted from this study;
809
nevertheless, the PV system costs used, particularly for 15- and 25-year lifetimes, may be
810
conservative.
811
Table 3: PV system and inverter costs
PV System Size (kWp)
Installed System Cost ($ / Watt)
Inverter Replacement Cost
Min
Max
Before Subsidy
With Subsidy
($ / Watt)
0
1.5
2.49
1.84
1.1
1.5
2
2.35
1.7
1.1
2
3
2.21
1.56
0.95
3
4
1.9
1.25
0.8
4
5
1.75
1.1
0.83
5
7
1.66
1.01
0.65
7
10
1.73
1.08
0.65
10
20
1.85
1.2
0.65
20
30
1.83
1.18
0.42
30
50
1.81
1.16
0.42
50
70
1.77
1.12
0.31
70
100
1.75
1.1
0.31
100
1.73
1.08
0.31
812
Appendix B: Commercial Tariffs at the Parent Meter
813
The tariff paid by the ENO at the parent meter would comprise a regulated network
814
component and a market retail energy component. In the relevant network area of the study,
815
the network component for a low voltage connection would be EA305 or EA310, depending
816
on annual load. These network tariffs have a relatively high ratio of fixed and capacity to
817
volumetric charges, as shown in Table 4, with the daily capacity charge based on the
818
customer’s peak load in the preceding 12-month period.
819
9
All costs in this paper are AU$ where AU$1.00 = US$ 0.754 (Bloomberg, 2018)
This is a preprint: original submitted version. The published version of the article Mike B. Roberts, Anna Bruce, Iain MacGill, A comparison
of arrangements for increasing self-consumption and maximising the value of distributed photovoltaics on apartment buildings, Solar
Energy, Volume 193, 2019, Pages 372-386, ISSN 0038-092X is available at https://doi.org/10.1016/j.solener.2019.09.067
25
The energy and retail component, determined by negotiation with the retailer and therefore
820
subject to a high degree of variability and to a lack of transparency, is likely to be
821
significantly lower than the estimated 14.63 c/kWh paid by a representative NSW retail
822
customer (AEMC, 2018) in 2017/18, and to include a TOU component. We have used a range
823
of high (‘TOU12’) and low (‘TOU9’) market prices from early 2018 plus environmental
824
charges of 1.71 c/kWh and GST.
825
Table 4: Commercial tariffs payable at the parent meter
Component
Name
Annual Energy
Use
(MWh)
Fixed
Charge
(c / day)
Peak
Rate
(c / kWh)
Shoulder
Rate
(c / kWh)
Off-peak
Rate
(c / kWh)
Capacity
Charge
(c/kVA/day)
Network
(ex GST)
(Ausgrid, 2017b)
EA305
160-750 MWh
1905.85
4.95
2.27
1.26
35.74
EA310
> 750 MWh
2403.13
4.40
2.11
1.39
35.74
Retail / Energy
(ex GST)
TOU9
9.00
9.00
6.50
TOU12
12.00
12.00
9.00
Environmental Charges
1.71
1.71
1.71
Combined tariff
(inc GST)
EA305_TOU9
2096.435
17.226
14.278
10.417
39.314
EA305_TOU12
2096.435
20.526
17.578
13.167
39.314
EA310_TOU9
2643.443
16.621
14.102
10.56
39.314
EA310_TOU12
2643.443
19.921
17.402
13.31
39.314
Although avoided transmission use of service costs may be paid for embedded generation
826
where network benefit is demonstrated (Ausgrid, 2017a), it is unusual, though possible, for
827
commercial customers to receive a FiT applied to PV export. However, for this study, we
828
have tested sensitivity to FiTs of 8c/kWh and 12c/kWh at the parent meter, in line with the
829
state regulator’s ‘all time benchmark’ rate for retail FiTs in 2017-18 and 2018-19 respectively
830
(IPART, 2018), as well as to zero payment for exported generation.
831
832
... Thus, they can involve the collaboration of individual consumers within residential buildings, as well as several neighborhoods, for the common purpose of expanding renewable energy and increasing their own share of locally generated renewable electricity. For example, [9] examine how the expansion of residential PV systems affects electricity self-consumption rates. [1] extend this approach by combining a PV system with a storage system, and calculating the achievable annual savings of residents in energy communities. ...
... Approaches to optimizing energy flows within energy communities are also being developed, studied, and tested in scientific literature [12][13][14][15]. Legal frameworks as well as challenges are explored by [9,16]. Indeed, the lack of sufficient legislation to ensure viability is one of the reasons for the delayed further development of energy communities [17,18]. ...
Article
Full-text available
The use of photovoltaic energy (PV) and the involvement of residents within energy communities are becoming increasingly important elements of decentralized energy systems. However, ownership structures are still too complex to empower electricity consumers to become prosumers. We developed a token-based system of the gradual transfer of PV ownership rights, from the initial investor to residential and small-scale commercial consumers. To demonstrate the system, we set up a simulation of a 27-party mixed usage building with different load profiles, ranging from single student apartments to office units with battery electric vehicles, in a German energy community. As a result, we show that the proposed system design is economically viable for all involved stakeholders over the simulation horizon from 2022 to 2036, with a payback time of <5 years, 4 years to distribute 50% of the PV tokens, and an overall self-consumption share of 69%.
... where E PV ,t is the energy generated by PV (kWh) and E Load,t is the electrical energy consumption (kWh). In the current analysis the instantaneous PV generation is given as (Roberts et al., 2019): ...
... The cost of the electrical energy consumed (E C ) by the desired load can be described as Abdulateef et al., 2021;Roberts et al., 2019;Hassan et al., 2021): ...
Article
Full-text available
In this research studies, the on-grid photovoltaic system capacity is sized based on the maximum energy self-consumption supplied to the household. The proposed system is designed and optimised for the house located in Diyala state, Iraq. The investigation is carried out using experimental data for electrical load and metrological data (solar irradiance and ambient temperature) at a high one-minute temporal resolution. The simulation process conducted using MATLAB in order to evaluate optimal system capacity can serve the desired load at maximum energy self-consumption for two setting positionings (i) annual optimum tilt angle, (ii) two-axis tracking system. The results show that based on the measured daily load average 7.42 kWh at the annual positing tilt angle the optimum photovoltaic system capacity approximately 7.15 kWp that showed the annual energy self-consumption dropped from 88.08% to 19.73% and the energy self-sufficiency raised from 12.93 % to 41.93 %. For a two-axis tracking system, the optimum photovoltaic system capacity is approximately 4.4 kWp that showed the annual energy self-consumption dropped from 85.2% to 22.35% and energy self-sufficiency raised from 14.6 % to 41.76 %. Moreover, the obtained economic results show that the energy cost dropped by about 41% (0.117 $/kWh) for the annual positioning tilt angle, and 57% (0.113 $/kWh) for two-axis tracking system on the basis of grid energy cost. The used methodology provides a distinct approach that can be effectively used to size the storage and renewable energy components for future applications.
... Investment costs were calculated using average system costs for each state [29] for the average potential system size for each dwelling type. For apartments, premiums of 25% (for 1-2 storey), 37.5% (for 3 storey) and 50% (for 4 storey or higher) were added to account for the range of additional barriers that apply to these buildings [30] and for the additional infrastructure needed to enable sharing of solar systems between households, whether 'behind the meter' [31] or through an embedded network [32]. These additional costs are based on conversations with solar installers and align with the experience of apartment solar-sharing specialists [31], but are building-specific and, therefore, uncertain. ...
Technical Report
Full-text available
Executive Summary Australia's social and community housing could host as much as 1.8 Gigawatts (GW) of rooftop solar, allowing some of Australia's most vulnerable energy consumers to benefit from the low cost of solar. This is enough to: • provide more than 2,000 skilled jobs for 5 years • save 440,000 households an average of $750 every year for 20 years • deliver bill savings of $328m per year for 20 years, for an initial investment cost of only $360 million a year for five years • generate 2.4 Terrawatt-hours (TWh) of electricity each year • offset 34 Megatonnes of greenhouse gas emissions over its 20-year lifetime,
... Energy communities allow prosumers to exchange the energy produced locally and reduce the need for external energy sources [7]. Shared renewable energy generation can provide greater self-consumption shares to the aggregated load of a building [8]. These are only a few examples of available studies on the energy community modelling. ...
Article
Full-text available
Energy communities are paving the way for new cooperation opportunities related to energy consumption and energy production. Individuals unite in energy communities to reduce the costs related to energy consumption. Although previous work has mainly focused on energy exchange inside the community. This work aims to investigate the Pareto-optimal solutions to the transformation of a historical district into an energy community. For energy efficiency and production measure calculation, a system dynamics model is developed. Multiobjective differential evolution optimization method is employed for the evaluation of energy efficiency and production measures with a focus on net present value, self-sufficiency, annual emission reduction, and specific heat consumption. The optimization target functions can be increased at a cost in net present value. Replacement of household appliances and windows enables significant energy demand reductions while maintaining positive net present value. Electricity production from photovoltaic panels offers an additional pathway to increase selfsufficiency share while maintaining positive net present value.
... Energy for Sustainable Development 68 (2022) 490-500 2020; Jung et al., 2020;Roberts et al., 2019;Schopfer et al., 2018;Solano et al., 2017). The SCR for a specified period is calculated by dividing the energy that is self-consumed by the energy that is generated, while the SSR for a period is calculated by dividing the self-consumed energy by the total energy consumed. ...
Article
Full-text available
The progressive fall in the cost of residential photovoltaic (PV) systems opens the possibility of evaluating systems that do not export surpluses to the grid, avoiding bureaucratic procedures and effects on the grid. The main purpose of this papers is to present a techno-economic model based in the context of Dominican Republic, to evaluate the profitability of residential PV systems, considering a non-incentivized self-consumption scheme, a step tariff of electricity and monthly electricity consumption. The methodology consists of maximizing the net present value (NPV) through the number of PV panels; subsequently, the simple investment recovery time (SPBT), the internal rate of return (IRR), the self-sufficiency rate (SSR), the self-consumption rate (SCR) and others quantities are obtained. In addition, a methodology is proposed to determine the self-consumed energy based on daily power demand profiles, differentiating between regular and irregular loads. For a domestic house with an average monthly energy of 579 kWh, the optimal value of the NPV was 535 USD with a SPBT of 4.83 years and a IRR of 20.8%. The results indicate that non-incentivized self-consumption with PV systems can be profitable in the residential sector of Dominican Republic.
... By considering the current regulatory guidelines, this paper analyzes how an MTB CSC initiative can be implemented in a smart grid environment, contrasting with the current research on CSC in MTB that focuses mostly on the projects economic viability, self-consumption and self-sufficiency. For instance, Lang et al., 2016, Roberts et al., 2019b, Sommerfeldt & Madani, 2016, Syed et al., 2020and Schiera et al., 2019 exploited the role of collective energy generation (andstorage, in Roberts et al., 2019a andSyed et al., 2020) technologies in the overall projects self-sufficiency, self-consumption and peak demand shaving results. These studies utilized different economic and technical approaches to show the correlation between collective energy systems and the energy and economic performance of the overall project: in Roberts et al., 2019a, the exploitation of a combined generation-storage energy system allowed to reach self-consumption and self-sufficiency rates of 19% and 12%, respectively, and a building peak demand shaving of up to 30%; in Syed et al., 2020, the overall building average yearly energy consumption was reduced in about 22%. ...
Article
Collective self-consumption can have an important role contributing to decarbonization and sustainability goals in cities. However, the implementation of such projects is hindered by technical, economic, social and regulatory barriers, which may compromise those goals. Based on the recent guiding principles for collective energy systems established under the scope of the European Union's Clean Energy Package, this work aims to assess how cost minimization and self-consumption maximization collective objectives may influence the economic and energy performance of a shared electricity generation and self-consumption project implemented in a multi-tenancy environment. A multiagent framework is developed to model the building dynamics while optimization algorithms are implemented to exploit individual and collective goals. Our findings show that cost minimization and self-consumption maximization can be conflicting objectives and influence the project energy and economic performance. While the cost minimization objective is more attractive for projects in which cost-driven participants are concerned with recovering investment, the self-consumption maximization objective is more suitable for cost-indifferent participants and projects aimed at energy self-sufficiency. These results raise relevant hints for stakeholders (participants, investors and policymakers), contributing to make better investment decisions and design better policies incentivizing electricity generation and management in multi-tenancy buildings.
Thesis
(In English Below) Obtener un sistema energético que contribuya a asegurar la estabilidad climática del planeta es uno de los desafíos más importantes de la primera mitad del siglo XXI. Con el propósito de contribuir en la búsqueda de vías que permitan superar la crisis climática global, pero desde acciones locales, y apelando a que la tecnología fotovoltaica (FV) cuenta con excelentes características para habilitar la transición energética que se necesita, esta tesis doctoral tiene como principal objetivo analizar, desde un enfoque global y local, el rol que la energía solar FV descentralizada podría jugar en la transición energética sostenible de un país y territorio específico. Para esto, se emplea como caso de estudio a Chile y particularmente, una de las regiones que lo conforma: la región de Aysén. Tanto Chile como la región de Aysén tienen aspectos que son un reflejo de la crisis global del Antropoceno, pero también cuentan con una gran oportunidad para implementar soluciones ejemplares basadas en sus enormes potenciales de energía renovable (ER). Para realizar dicho análisis se han considerado todos los sectores consumidores de energía y se utilizó una herramienta desarrollada por la Lappeenranta University of Technology (LUT), con la que se modelaron escenarios de transición energética hacia un sistema 100 % basado en ER para Chile, desde un enfoque global y local, donde, en el enfoque local se incluyó a la región de Aysén. Los resultados revelan que, tanto en Chile como en la región de Aysén, lograr un sistema energético 100% renovable para el año 2050 es técnicamente factible y económicamente viable. En ese año, dependiendo del enfoque y escala territorial, la contribución a la generación eléctrica por parte de la tecnología FV en su conjunto varía entre 39–86 % y, la contribución de la FV descentralizada varía entre 9–12 %; no obstante, la FV descentralizada aporta entre un 27–52 % de la electricidad final que es mayormente consumida en las ciudades por los sectores eléctrico, térmico y transporte. A su vez, la energía solar FV descentralizada crearía en Chile entre el 9–15 % de los empleos anuales directos durante el periodo de transición. Es decir, entre los años 2020 y 2050, el sector de la FV descentralizada crearía 174.274 empleos directos. Además, los resultados también revelan que Chile puede alcanzar la neutralidad en emisiones de carbono en el año 2030 y, se puede convertir en un país emisor negativo de gases de efecto invernadero a partir del año 2035. Todo esto sería posible utilizando menos del 10 % del potencial tecno-económico de ER disponible en este país. Tras los resultados del trabajo de investigación realizado en esta tesis doctoral, se concluye que la energía solar FV es un elemento vital en la transición energética sostenible, así como también, alcanzar un sistema energético totalmente desfosilizado es más importante que lograr la neutralidad en las emisiones de carbono. Esto último se debe a que una transición a nivel país hacia un sistema energético 100 % renovable implicaría beneficios socio-ambientales y socioeconómicos locales, con impactos globales positivos que se necesitan con urgencia. Si Chile implementara una vía de transición hacia un sistema energético 100 % renovable, no solo podría convertirse en un caso ejemplar en el avance hacia una economía post-combustibles fósiles, si no que también podría contribuir a la transición energética global: a través de la extracción limpia de materias primas clave (como lo son el cobre y el litio), y a través de la producción de combustibles y químicos basados en ER. En resumen, la tecnología FV puede contribuir en la mitigación del cambio climático y la reducción de los niveles de contaminación del aire en las ciudades, al tiempo que impulsa el crecimiento económico local; todo esto, de una manera más descentralizada y participativa. ///////////////////////////////////////// Obtaining an energy system that will help to ensure the climactic stability of the planet is one of the most important challenges of the first half of the 21st century. In order to contribute to the search for ways to overcome the global climate crisis, from local activities, and appealing to the fact that photovoltaic (PV) technology has excellent characteristics which could enable the energy transition that is needed, this doctoral thesis has as its main objective the analysis, from a global and local approach, the role that decentralized solar PV could play in the sustainable energy transition of a specific country and territory. For this purpose, Chile and one of its regions (the Aysén region) are used as a case study. Both Chile and the Aysén region have aspects that reflect the global crisis of the Anthropocene, but they also present a great opportunity to implement exemplary solutions, based on their enormous renewable energy (RE) potentials. To carry out this analysis, all energy-consuming sectors were considered. A tool developed by the Lappeenranta University of Technology (LUT) was used, with which energy transition scenarios were modelled towards a 100% RE-based system for Chile, from a global and local approach. The Aysén region was included in the local approach. The results reveal that, both in Chile and in the Aysén region, achieving a 100% RE system by 2050 is technically feasible and economically viable. In that year, depending on the approach and territorial scale, the contribution to electricity generation by PV technology as a whole would vary between 39–86%. The contribution of decentralized PV would be between 9–12%. However, decentralized PV would contribute 27–52% of the final electricity that is mostly consumed in cities by the power, heat and transport sectors. In turn, decentralized solar PV would create between 9–15% of annual direct jobs in Chile during the transition period. In other words, between 2020 and 2050, the decentralized PV sector would create 174,274 direct jobs. In addition, the results also reveal that Chile could achieve carbon neutrality in 2030 and could become a negative greenhouse gas emitter by 2035. All of this would be possible by using less than 10% of the techno-economic potential of RE available in this country. From the results of the research work carried out in this doctoral thesis, it is concluded that solar PV is a vital element in the sustainable energy transition. We also find that achieving a fully defossilized energy system is more important than achieving carbon neutrality. The latter is due to the fact that a transition at the country level towards a 100% RE system would imply local socio-environmental and socio-economic benefits, with positive urgently needed global impacts. If Chile implements a transition path towards a 100% RE system, it could not only become an exemplary case in moving towards a post-fossil fuel economy, but could also contribute to the global energy transition through the clean extraction of key raw materials (such as copper and lithium), and through the production of RE-based fuels and chemicals. In summary, PV technology can contribute to mitigating climate change and reducing air pollution levels in cities, while boosting local economic growth, doing all of this in a more decentralized and participatory way.
Article
The cost fall of solar photovoltaics (PV) in the last decade has driven remarkable deployment of PV worldwide. A proportion of this deployment has occurred within the distribution network, potentially having a significant economic impact on both private industry stakeholders and society as a whole. Hence, the impacts of distributed PV have raised growing concerns on how to maximize PV societal values while managing potential negative impacts on key stakeholders. However, a review, evaluation, and comparison of the societal PV value and the stakeholder private PV values, together with a thorough analysis of their connections and challenges, are still lacking in the literature. In this article, we first propose some clear definitions underpinning the concepts of societal PV value and private PV value, to then review the literature and assess them for different electricity stakeholders and retail market arrangements. From the study of the Australian National Electricity Market, we found significant misalignments and conflicts between societal and private PV values due to a number of market failures, meaning that there are still not appropriate economic incentives for industry stakeholders to maximise the value of PV for society. Moreover, the total value of PV for residential PV customers falls within the possible range of PV values for society, although without the right price signals.
Article
Full-text available
The reduction in the costs of residential photovoltaic (PV) systems has increased their viability and implementation for self-consumption and export of energy electricity. The implementation of these systems requires feasibility studies, considering the structure of the electricity tariff, the stability in the grid, the incentives and other variables. This work reviews 158 papers on the viability and sizing of residential PV systems, with the purpose of showing a general overview of the subject and that serves as a guide to carry out future research in the residential sector. The results show that the simulation methodology is the most frequent in the techno-economic study of residential PV systems, with a percentage of 45% followed by optimization with 37%; 29 analysis tools were identified in the study of residential PV systems, being MATLAB the most used . The combination of PV systems with batteries or Battery Energy Store System (BESS) has been increasing in the literature, even over PV systems without batteries. In addition, the case of the Dominican Republic is analyzed, identifying three cases to be evaluated, considering the Net metering (NM) program, self-consumption, step tariff and electricity outages. It was determined that in the Dominican Republic, the installed residential PV systems capacity in NM program is approximately 7.83 kW/user .
Conference Paper
Full-text available
Over the last two decades, grid-connected solar photovoltaic systems have increased from a niche market to one of the leading power generation capacity additions annually. In 2018 the total installed photovoltaic electricity generation capacity exceeded 500 GW. Another doubling is forecasted until the early 2020s. Therefore, the further development is coupled with the question at what prices solar photovoltaic electricity can be provided and delivered to the customers. The installation of PV systems for self-consumption is already now an interesting option for many people but in general limited to those who have access to a rooftop they own or can use. A new envelopment to enable residents of multi apartment buildings to commonly use electricity generated by a PV system (collective self-consumption) is a relatively new development and is still facing a lot of administrative and regulatory challenges.
Article
Full-text available
Understanding of residential electricity demand has application in efficient building design, network planning and broader policy and regulation, as well as in planning the deployment of energy efficiency technologies and distributed energy resources with potential emissions reduction benefits and societal and household cost savings. Very few studies have explored the specific demand characteristics of apartments, which house a growing proportion of the global urban population. We present a study of apartment electricity loads, using a dataset containing a year of half-hourly electricity data for 6,000 Australian households, to examine the relationship between dwelling type, demographic characteristics and load profile. The focus on apartments, combined with the size of the data set, and the representative seasonal load profiles obtained through clustering full annual profiles, is unique in the literature. We find that median per-occupant household electricity use is 21% lower for apartments than for houses and that, on average, apartments have lower load factor and higher daily load variability, and show greater diversity in their daily peak times, resulting in a lower coincidence factor for aggregations of apartment loads. Using cluster analysis and classification, we also show the impact of dwelling type on the shape of household electricity load profiles.
Article
Full-text available
Distributed photovoltaics is playing a growing role in electricity industries around the world, while Battery Energy Storage Systems are falling in cost and starting to be deployed by energy consumers with photovoltaics. Apartment buildings offer an opportunity to apply central battery storage and shared solar generation to ag-gregated apartment and common loads through an embedded network or microgrid. We present a study of energy and financial flows in five Australian apartment buildings with photovoltaics and battery storage using real apartment interval-metered load profiles and simulated solar generation profiles, modelled using an open source tool developed for the purpose. Central batteries of 2-3 kWh per apartment can increase solar self-consumption by up to 19% and building self-sufficiency by up to 12%, and shave overall building peak demand by up to 30%. Although the economic case for battery storage applied to apartment building embedded networks is not compelling at current capital prices, with cost thresholds of AU$400-AU$750/kWh compared to AU$750-AU$1000/kWh for individual household systems, there are clear financial benefits to deployment of embedded networks with combined solar and battery storage systems for many sites.
Preprint
Full-text available
This paper reviews opportunities for, and barriers to, increasing photovoltaic (PV) deployment on apartment buildings, with a particular focus on the Australian experience. In 2015, residential loads account for 27% of global electricity use [1], while offsetting these loads with rooftop PV has been significant in developing a commercial global PV market and is critical to achieving COP 21 emissions targets. However, as rapid urbanisation drives increasing housing density, PV penetration in multi-occupancy housing has been limited by comparison with stand-alone housing in many jurisdictions, including in Australia despite its world-leading residential PV penetration. Given the growing commercial attractiveness of residential PV, this also raises equity concerns for apartment households. PV can potentially be installed to supply electricity to common property, to serve individual apartments, or as a resource shared between multiple apartments through embedded networks, local energy trading or ‘behind the meter’ deployment models. Our study undertook a review of the academic literature in this space and of specific Australian arrangements, as well as conducting a series of semi-structured interviews with a range of relevant stakeholders. Barriers identified include the huge variety amongst existing apartment building stock which precludes standardised retrofitting solutions, demographic factors and knowledge issues. However, the Australian regulatory context - including governance of apartment buildings, regulation of the energy market, and electricity tariff policies - also impacts on the options available to apartment residents. New models for deploying PV on apartments are emerging, including initiatives from retailers, developers and community energy organisations. While some issues are specific to the Australian context, or to buildings governed under strata-type arrangements, broader lessons can be taken from the Australian experience, including to inform the design of the regulatory framework required to facilitate widespread PV deployment across all residential housing types.
Conference Paper
Full-text available
Despite potential advantages of load aggregation and scale discounts, few of Australia's 2.3 million apartment residents are amongst the country's 1.8 million solar prosumers. However, embedded networks can be used to distribute rooftop photovoltaic generation to households if split incentives and regulatory barriers are overcome. We present a model of an embedded network with PV in an Australian apartment building. Load data from real apartment households are combined with modelled PV generation to describe energy and cash flows over the course of a year. The distribution of costs and benefits between stakeholders is calculated under a range of financial arrangements. Embedded network benefits are highly sensitive to retail and wholesale energy costs at the parent meter and to site-specific capital costs. The addition of PV can, in some circumstances, increase the financial viability of an embedded network and careful tariff design can help incentivise this investment and ensure customer retention.
Technical Report
Full-text available
This short report aims at providing a brief explanation about self-consumption concept without going in-depth into statistical analysis and current status of PV self-consumption in selected countries all over the world. Based on understanding how self-consumption works and parameters to define self-consumption, the comprehensive analysis of self-consumption in keys countries will be presented in the below full report "Review and Analysis of PV Self-consumption Policies"
Article
Full-text available
This paper analyses the profitability and business models of shared, nonsubsidized PV systems’ usage in multiapartment buildings in Austria in the context of legislative amendments which came into force in July 2017. In addition, it compares the Austrian results with those of Germany, where significantly higher retail electricity prices determine the profitability benchmark. To that end, a multiobjective optimization model is developed for the optimal dimensioning of PV systems and energy storage facilities in keeping with different end user objectives, ranging from minimizing annual electricity costs to maximizing self-consumption. The results show that the profitability of shared use of nonsubsidized PV systems is marginal in Austria. This means that, based on individual apartment load profiles, the profitability gap ranges between 0 and 40 euros per apartment, whereas the consideration of the building as total load leads to a small cost-saving potential of about 90 euros for the whole building in the best case and thus profitability. In contrast, significant profitability of shared PV systems in multiapartment buildings can be achieved in Germany, where the renewable energy surcharge results in high retail electricity prices. At present, different business models, accounting and billing concepts, are being tested in these countries to learn about the best-practice concepts.
Conference Paper
Full-text available
The growing adoption of distributed energy resources (DERs) across Australia may represent the start of a transition of Australia's power system from a centralised generation model towards an interconnected set of embedded microgrid systems. In these systems, local trading of coincident generation and consumption is being explored, with the idea that this would encourage consumer engagement, provide more choice, and incentivise reduced use of networks (via increased balancing of loads and generation locally). However coincidence of generation and load over short time frames, and hence network benefit, can be difficult to determine using existing metering systems. There are many sites undergoing trials to determine how energy flows can be efficiently monitored and accounted for in microgrid systems. At the current time, commercial metering is generally performed on time scales greater than thirty seconds, with most metering systems measuring net flows on half-hourly intervals. This represents a barrier to accurate accounting, since the time at which nodes generate and consume energy within a wide metering period is not known. As a result, end-users may face effective penalties or avoid charges that would accrue to their generation or consumption profile under more accurate accounting. Accurate accounting for the economic benefits of embedded microgrids that result from either reductions in external network use or contributions to improved reliability, rely on sub-second level timing, and cannot be commercially factored (or incentivized) without corresponding metering. In this paper, the coincidence of generation and consumption in a micro-grid setting is examined over different timeframes using a software based simulation, and the impact of different time intervals for accounting is explored using an algorithmic theoretical approach. It is found that there are predictable trends in the way that metering time periods impact the accuracy of 'peer to peer' accounting. For the limited dataset tested, the inaccuracies were found to be small relative to overall energy consumption, however further work is required to determine whether this can be generalized across the majority of schemes.
Article
This paper reviews opportunities for, and barriers to, increasing photovoltaic (PV) deployment on apartment buildings, with a particular focus on the Australian experience. As rapid urbanisation drives increasing housing density, PV penetration in multi-occupancy housing has been limited by comparison with stand-alone housing in many jurisdictions, including in Australia despite its worldleading residential PV penetration. Given the growing commercial attractiveness of residential PV, this also raises equity concerns for apartment households. PV can potentially be installed to supply electricity to common property, to serve individual apartments, or as a resource shared between multiple apartments through embedded networks, local energy trading or ‘behind the meter’ deployment models. Our study undertook a review of the academic literature in this space and of specific Australian regulatory arrangements, as well as conducting a series of semi-structured interviews with a range of relevant stakeholders. Barriers identified include the huge variety amongst existing apartment building stock, demographic factors and knowledge issues. However, the Australian regulatory context - including governance of apartment buildings, regulation of the energy market, and electricity tariff policies - also impacts on the options available to apartment residents. New business models for deploying PV on apartments are emerging, including initiatives from retailers, developers and community energy organisations. While some issues are specific to the Australian context, or to buildings governed under strata-type arrangements, broader lessons can be taken from the Australian experience, including to inform the design of the regulatory framework required to facilitate widespread PV deployment across all residential housing types.
Article
Multiple studies show that large installations of rooftop solar PV in residential low voltage grids, typically with many single family dwellings, may cause overvoltage issues during mid-day when local consumption is low and solar PV electricity generation high. Different so-called smart grid technologies, which often requires additional control, communication and monitoring equipment have been suggested to alleviate these and related problems. In this paper, solar PV integration is studied in the context of multi-apartment buildings where the rooftop potential is significant. To this end, the medium voltage grids of two multi-apartment areas, Bärnstenen and Alabastern in Vaxjö, Sweden are used as study cases. For these areas, it is found that active smart grid control or the introduction of new controllable load is not required. This finding applies to cases with very large solar PV installations corresponding to full coverage of the available rooftops and an annual yield corresponding to about eight times the annual electricity consumption. The conclusion is that multi-apartment residential areas may be ideally suited for large-scale solar PV installations without the need for smart grid infrastructure. These findings are contrary to, but not in disagreement with previous findings for low voltage residential grids.