Residential sector demand response – Is it worth chasing?

Share this article:

With the lion’s share of recent demand response efforts focusing on the commercial and industrial sectors, are opportunities being missed in the residential sector? Or is convincing Australian households to switch off just not worth it? In this article, MHC summarises results from several trials from Australia and overseas. Using key learnings from these trials, we have conducted a desktop study to gauge how much demand could be shaved from load peaks under various scenarios. Is the size of the prize is worth pursuing? 

While demand response (DR) isn’t new to the Australian electricity market, it has experienced somewhat of a renewed focus in recent months. Faced with further increasing prices and a coming summer of uncertain supply security, initiatives such as ARENA and AEMO’s recent demand response pilot program[1] are looking to demand response as a key ingredient to solving the grid’s current challenges. The focus to date has understandably been on the commercial and industrial (C&I) sector, largely due to the ability of sector to provide large loads for management and the clearer profit motive of its customers.

However, with the residential sector accounting for over a quarter of all electricity consumed in Australia[2], households may well be the next frontier for DR opportunities. Nevertheless, with the clear differences between C&I and residential customer – particularly in terms of load profiles and customer behaviours – it is worth subjecting it to a critical eye.


Several notable residential demand response trials have been conducted both in Australia and overseas – these are summarised below.

Australia – Mojo Power

Key result = 40% take-up rate achieved

Prior to the February 2017 heatwave that hit NSW, energy retailer Mojo Power conducted a demand management trial that involved offering financial savings to consumers if they voluntarily reduced demand between peak hours 4-6pm on the 10th February. Out of the 500 customers targeted, 200 households responded (i.e. 40% take-up rate)[3]. Participants were required to turn down/off air conditioners and pool pumps during these hours, and received between $25 and $140 in bill credits as incentives.

Results from the trial (submitted by Mojo Power to the Finkel Review) showed that the top 10% of participants reduced their consumption to almost zero, as shown in the graph below:

Figure 1: Demand response by Mojo customers on 10 February 2017[4]

Australia – United Energy

Key result = 30% load reduction for peak events

In Victoria, a residential electricity demand response trial conducted by United Energy (UE) resulted in a more than 30% load reduction for peak events occurring during the 2016 summer season[5].

Like the Mojo Power trial, the UE trial targeted peak usage from 4pm-7pm over four events during the summer. The UE example however was conducted via an energy management program and app that provided households with a personal energy use goal for the peak event period and real-time updates on how their usage tracked against this. Customers were incentivised with $25 for each ‘event day’ they met their usage goal up to $100 over the season.

Australia – Energy Queensland

Key result = 30% take-up rate achieved

The merger of Ergon Energy and Energex has the largest residential DR program in Australia with up to 770,000 residential customers (approximately 30% of the Queensland residential customers) connected to load controlled tariffs targeting electric hot water systems, pool pumps and PeakSmart air-conditioners.

Both electric hot water and pool pump tariff incentives are utilised to implement Energy Queensland’s DM program; the Pool Rewards Program entitles households to a $200 financial incentive when pool pumps are connected to the Economy Tariff 31 or 33 whereas the Hot Water Rewards scheme provides rebates of $100 or $200 to eligible South-East Queensland households that connect an electric hot water system to an economy electricity tariff.

According to Energex’s Demand Management Plan 2017-18[6], load control of solely hot water supplies can cut peak demand by 6.71%. An alternate analysis targeting PeakSmart air-conditioning systems led to a 25% demand reduction (8kW) between 4:30pm to 5:30pm.

Australia – AGL Energy

Key result = 16% of air conditioning system load under management

AGL Energy is working with 68 households in Carrum Downs, Melbourne to examine how energy management technology, as well as solar and storage, can be used to balance spikes in electricity demand during hot weather, and ultimately reduce consumer energy costs. By targeting demand drawn by air-conditioning systems in a four-month trial (December 2015 – March 2016) the aim was to prove capability in reducing demand by 25kW for each hot weather event[7]. This equates to 368W per household, or approximately 16% load reduction of a medium sized 2300W air conditioning system[8].

Australia – AusGrid

Key result = 6% winter peak and 2% summer peak load reduction targeting hot water systems only

In NSW, AusGrid carry out load control schemes using ‘ripple’ signals sent along power lines to remotely switch off hot water systems.

With 500,000 residential customers on controlled load tariffs, across AusGrid’s network area peaks can be reduced by 6% in winter and 2% in summer. Subsequently, hot water load control trials carried out in 2012/13 saw on average a 274W demand reduction in summer and 407W reduction in winter[9].

The results suggest that customers will tolerate the occasional control of small hot water systems for several hours on peak days for a modest financial incentive.

United States

Key result = 12% peak load reduction targeting air conditioning systems only

A study held in Wichita, Kansas[10] examined the possible demand reduction due to solely targeting residential air conditioning systems when using a 20-minute load cycling method. The direct load control programs took place during peak demand periods between 2-4pm and consequently saw a reduction of 12%. This figure aligns to that provided by the Federal Energy Regulatory Commission’s 2009 National Assessment which concluded that residential air conditioning systems can contribute to a 10% reduction in demand[11] – which in turn align to MHC’s calculations for targeting only air conditioning systems.


MHC has applied key results from these trials to conduct a high-level technical analysis to estimate how much load could be saved under various scenarios if residential DR was rolled-out NEM-wide.

Scenarios considered the main in-home devices/appliances that could be managed by a DR aggregator. The key assumptions are shown below in Table 1, showing the total controllable load for the devices listed during peak events:

Device(s) / Appliance(s) Controlled[12] Prop’n of households with electric- powered device (%) No. households in NEM (#) Total appliances in Australia (#) Ave. power rating per device (kW)[13] Prop’n of load control-lable by DR (%)[14] Prob. of Device in Use During Peaks (%) Total controllable load during peak event (GW)
Air Conditioning 74%[15] 8.1 million [16],[17] 6.0 m 2.3 kW 16%[18] 80% 1.8 GW
Water Heating 56%[19] 4.6 m 1.3 kW 100% 40% 2.4 GW
Pool Pumps 12%[20] 1.0 m 1.0 kW 100% 40% 0.4 GW
Dishwasher 55%[21] 4.5 m 2.4 kW 100% 10% 1.1 GW
Washing Machine 97%[22] 7.9 m 0.9 kW 100% 10% 0.7 GW
Tumble Dryer 55%[23] 4.5 m 2.4 kW 100% 10% 1.1 GW

Table 1: MHC desktop analysis – Key assumptions.


Given that we would not expect 100% take-up of DR in the residential market, MHC then applied a range of DR market penetration levels (Low=5%, Med=15% and High=30%) to estimate the amount of load reduction that could be called upon during a peak event. The summary results are shown below in Table 2:

Device(s) / Appliance(s) Controlled Available load under management (GW)
Low take-up (5%) Med take-up (15%) High take-up (30%)
Air Conditioning 0.09 GW 0.27 GW 0.53 GW
Water Heating 0.12 GW 0.36 GW 0.71 GW
Pool Pumps 0.02 GW 0.06 GW 0.12 GW
Dishwasher 0.05 GW 0.16 GW 0.32 GW
Washing Machine 0.04 GW 0.11 GW 0.21 GW
Tumble Dryer 0.05 GW 0.16 GW 0.32 GW
TOTAL: 0.37 GW 1.1 GW 2.2 GW
% peak reduction[24]: 1.1% 3.2% 6.4%

Table 2: Available residential load under management (GW) – Scenario-based, NEM-wide.


While noting that different states will likely experience peak demand on different days of the year, the results of the ‘high’ penetration scenario suggest that 2.2 GW of demand could be reduced during a peak event by managing those devices included above. Considering the maximum demand across the NEM in FY17 of 34.4 GW[25], a 2.2 GW reduction would equate to a 6.4% reduction in peak demand. These figures do however reduce to only 0.37 GW and 1.1% in a ‘low’ take-up scenario.

These results suggest that while opportunity for significant peak reductions through residential DR exists, achieving wide-scale take up amongst residential consumers will be key. This in turn will require the offer of strong incentives to residential customers. There would also likely be hesitation amongst customers to cede full control of their devices to a DR aggregator, meaning an ‘opt-out’ would be required to encourage significant take-up rates.


While our desktop analysis (supported by results of various trials) suggests that a not insignificant peak reduction of 2.2 GW could be achieved in a ‘high’ scenario, the next question to be answered (and a topic for further investigation) is to consider how much chasing this peak reduction would cost.

Costs would largely involve the cost of incentives paid to residential consumers to participate in DR programs – Particularly given that our results suggest that significant take-up rates would be required for any meaningful peak reduction. This would in turn require an understanding of customer take-up rates at different incentive levels. While a small portion of ‘early adopter’ customers may be incentivised at relatively low costs, achieving 30% penetration as modelled in our ‘high’ scenario would be expected to require significantly higher incentives to encourage residential participation.

Costs would also include the expense of rolling-out of hardware and systems required to operate residential DR. While a combination of indirect ‘passive’ load management (i.e. requiring the customer to switch-off devices) or expensive ripple control technologies have been largely used to date, the increasing number of internet-connected household devices now means communicating via IoT (“Internet of Things”) could enable DR aggregators to simply communicate on/off signals via the internet or other IoT protocols such as ZigBee or Z-Wave. This could enable aggregators to convince customers to give up their involvement in the process as part of a bundle of energy services, in return for financial benefit.

Once the costs of achieving the peak reductions identified here are better known, comparing these to the costs of the alternative – that is, purchasing the required caps and other hedging instruments from the wholesale energy market – will enable us to answer the “Is residential DR worth it?” question. With wholesale energy costs rising so significantly in recent months, and forecast to continue to do so, the bar could be an easy one to jump over.

The size of the available load is definitely enough to warrant undertaking this next stage of the analysis – which we will write about in the next edition of QSI.


[1] ARENA “Demand Response Competitive Round” [Accessed online]

[2] Climate Change Authority (prepared by Vivid Economics), “Analysis of electricity consumption, electricity generation emissions intensity and economy-wide emissions.” P. 28. [Accessed online]

[3] Potter. B, Financial Review (2017, March) Sydney family paid to turn off power on 40-degree day. [Accessed online]

[4] Finkel. A, Department of the Environment and Energy (2017, June) Independent Review into the Future Security of the National Electricity Market

[5] Vorrath, (2016, May) S. Australian demand response trial delivers 30% peak load reduction [Accessed online]

[6] Energex, (2017, May) Demand Management Plan 2017-18

[7] AGL Energy Limited (2016, March) AGL trials impacts of emerging technologies on the grid and energy bills [Accessed online]

[8] Essential Energy, “Energy Answers Appliance Guide”, [Accessed online]

[9] AusGrid Demand Management (2016, August) Hot Water Load Control Trials

[10] Jewell. W, Wichita State University (2014, September) The Effects of Residential Energy Efficiency on Electric Demand Response Programs. [Accessed online]

[11] Op. cit. P. 2

[12] Refrigerators omitted from study due to low customer acceptance to offer control of this appliance.

[13] Essential Energy, “Energy Answers Appliance Guide”, [Accessed online]

[14] MHC assumptions (unless references otherwise)

[15] Australian Bureau of Statistics, “Environmental Issues: Energy Use and Conservation, Mar 2014”. [Accessed online]

[16] Australian Bureau of Statistics, “Household and Family Projections, Australia, 2011 to 2036”. (Average of 2016 Series I, II and II applied.) [Accessed online]

[17] Australian Bureau of Statistics, “Australian Demographic Statistics, Dec 2016”, [Accessed online]

[18] As per AGL trial results

[19] Australian Bureau of Statistics, “Environmental Issues: Energy Use and Conservation, Mar 2014”. [Accessed online]

[20] Australian Bureau of Statistics, “Environmental Issues: People’s Views and Practices, Mar 2007”. [Accessed online]

[21] Australian Bureau of Statistics, “Environmental Issues: Energy Use and Conservation, Mar 2014”. [Accessed online]

[22] Australian Bureau of Statistics, “Environmental Issues: Energy Use and Conservation, Mar 2008”. [Accessed online]

[23] Australian Bureau of Statistics, “Environmental Issues: Energy Use and Conservation, Mar 2014”. [Accessed online]

[24] Based on 34.4 GW peak NEM demand 2016/17 from Australian Energy Market Operator (AEMO), “NEM Fact Sheet”. [Accessed online]

[25] Op. Cit.