This last week has seen extraordinary events in South Australia’s energy market make front page headlines nation-wide. In an unprecedented move, South Australian business and energy leaders demanded the re-start of a moth-balled power station to provide relief from suging and variable wholesale energy prices… and the Minister complied.
As reported in the Australian Financial Review, prices in the state have been “frequently surging above $1000 a MWh this month and at one point… hitting the $14000MWh maximum price”. The Australian Financial Review reports that average monthly prices have been three to four times higher than in the eastern states during the month of July and new contract prices in South Australia are nearly double the prices in the eastern states.
Image from Energetics.
Let’s be clear: the South Australian electricity supply is the cleanest it has ever been and it is the most vulnerable, volatile and fragile than any time in recent history with no signs of relief in the short-term. As much as many people, including me, want the former (clean power), we are shooting ourselves and our wishes in the head if we keep contributing to the latter. There are few worse advertisements for clean energy than the current market in South Australia. Short-sighted over-development of variable generation without compensatory planning and policy has driven consequences that were entirely foreseeable. Suggestions that the renewable sector is now merely a “scapegoat” for our problems are absurd, stemming from an ideology of nil criticism for some technologies. While those sectors are not alone in the frame as contributing to this problem, the Pollyanna group-think that insists that no line can ever be drawn to the obvious shortcomings of variable generators is starting to positively stink.
Around 12 months ago we published our paper Beyond wind. Since that time I have observed nearly everything we flagged coming true only faster than I anticipated, with the biggest surprise being that we did stand by as reliable generators left the market rather than coughing up to keep them in the game.
I have re-produced an extract of this paper below. As you can see both we and the sources we cite were paying attention to problems in the pipeline. These problems were foreseeable and foreseen. Maybe we just needed more pain to make us pay attention.
Since 2003, the contribution of wind power to electricity generation in South Australia has grown to around 27 % of total annual electricity supplied to the State (Australian Energy Market Operator Ltd. 2014b). This increased wind generation has come mainly at the expense of generation from existing coal and gas generators which are now run less frequently (Australian Energy Market Operator Ltd. 2014b). Yet despite the rapid increase in wind-generated electricity in the State, South Australia still depends on participation in the National Electricity Market for a reliable supply of electricity.
The National Energy Market is spatially the largest electricity grid in the world and serves approximately 9.5 million end-use customers (Australian Energy Regulator 2014). It is a wholesale market for the supply of electricity to retailers and end users in Queensland, New South Wales, the Australian Capital Territory, Victoria, Tasmania and South Australia. Exchange of electricity is facilitated through a pool where the output from all generators in the network is aggregated and scheduled at short (15-minute) intervals, to meet demand across the network. Within the National Electricity Market, electricity is indistinguishable from one generator to another, but network stability concerns mean that there is a need to have generators operating across a wide geographical spread of network nodes. The purpose of the market is to provide efficient and above all, secure electricity supply to meet a dynamically changing electricity demand efficiently (Australian Energy Market Operator Ltd. 2010b).
South Australia’s connection with the National Electricity Market supports both reliability of supply and the efficient use of the wind resource, typically exporting power when the output is high and demand is low (as commonly occurs around 04:00) (Australian Energy Market Operator Ltd. 2013b). Over the entire National Electricity Market, wind contributed 4.4 % of total electricity generation output in 2013-2014, with 74 % coming from coal, and 12 % from gas (Australian Energy Regulator 2014). Despite the ability to sell low-emissions power from wind, South Australia imported 2010 GWh in 2013-2014, six times the quantity exported (338 GWh) (Australian Energy Market Operator Ltd. 2014b).
Trading between adjacent National Energy Market regions relies on high-voltage transmission lines called ‘interconnectors’, which are used to import electricity into a region when demand is higher than can be met by local generators, or when the price of electricity in an adjoining region is low enough to displace the local supply (Australian Energy Market Operator Ltd. 2010b). The efficient use of South Australia’s wind generators relies on two interconnectors to Victoria, as well as substantial transmission infrastructure within South Australia. South Australia’s larger Heywood Interconnector (460 MW) was used at 100 % capacity for 8.7 % of the time in financial year 2012-2013 (Australian Energy Market Operator Ltd. 2013d). A $108-million upgrade of Heywood, to be commissioned in July 2016, aims to accommodate the increase in wind generation that has occurred over the last few years (Electranet 2013). The recently approved development of Australia’s largest wind farm (199 turbines for 600 MW at a cost of ~ $1.3 billion), to be located on the Yorke Peninsula, includes investment in 60 km of undersea cables to transmit the power to load centres, as well as two converter stations (The Ceres Project 2013). In another study, capital costs of > $900 million were identified for the additional transmission requirements to support development of the extensive Eyre Peninsula wind resource, with annual operational and maintenance costs of > $18 million year-1 (Baker & McKenzie, Worley Parsons & Macquarie 2010).
With the benefit of the National Electricity Market ensuring security of supply and efficient export of surplus generation, the wind sector has driven total greenhouse-gas emissions from South Australia’s electricity sector down by one quarter over the last ten years: from just over 8 megatonnes (Mt) CO2-e year-1 to just over 6 Mt CO2-e year-1 (Australian Energy Market Operator Ltd 2014). South Australian electricity now has the second-lowest emissions intensity (> 0.6 kg CO2-e kWh-1) of the Australian states and territories (Figure 4), having diverged sharply from approximate parity with Queensland, New South Wales and the South West Interconnected System from 2005 until today (the South West Interconnected System is a smaller electricity grid that serves the south-west of Western Australia; it is not part of the National Electricity Market). Until recent connection with the National Electricity Market, Tasmanian electricity generation had nearly zero emissions due to a predominant supply from hydro-electric generation. It has retained the lowest-emission electricity of any National Electricity Market region (0.2 kg CO2-e kWh-1), but its relative emissions intensity has risen sharply following the interconnection. Victorian electricity releases approximately 1.2 kg CO2-e kWh-1 due to a dependence on combustion of lignite (brown coal) for electricity supply.
Electricity from wind generation brings challenges related to its variable and intermittent supply. As installed capacity grows, the frequency of sudden changes in wind farm output also increases, rendering the management of power systems and transmissions networks more challenging (Australian Energy Market Operator Ltd. 2013d). A review of the aggregated wind output across three defined geographical regions in South Australia (Mid-North, South-East and Costal Peninsula regions) has found that spatial dispersion of wind generation helps to reduce overall variation in supply, but cannot substantially mitigate it (Australian Energy Market Operator Ltd. 2013d).
The relationship between wind generation and consumer electricity demand , shows “little correlation… between the aggregate wind output and demand in any region” (Australian Energy Market Operator Ltd. 2011a). At times wind supply can be negatively correlated with demand during heat waves (Australian Energy Market Operator Ltd. 2011b). So while the geographic distribution of wind provides some smoothing, the combined variability of wind and consumer demand means that other generation sources are required to respond to rapid changes of supply during periods of low output from wind (Australian Energy Market Operator Ltd. 2013d). For example the largest five-minute change in supply from wind in South Australia was a decrease of 294 MW (Australian Energy Market Operator Ltd. 2013d). To manage this variation, capacity in excess of an entire, large generating unit (280 MW of coal generation from Northern power station) had to be sourced at short notice (Australian Energy Market Operator Ltd. 2013d). Such challenges will increase in size and frequency, and therefore potential economic cost, as wind power supply increases, notwithstanding improving prediction of the availability of electricity from wind (Edis 2014).
The lack of correlation between electricity demand and supply from wind has another long-term impact on overall system costs: the constrained ability to retire other ‘baseload’ (in reality, ‘dispatchable’ (sensu Nicholson, Biegler & Brook 2011), generators from service. This is best illustrated by the poor correlation between supply and peak demand. During periods of peak demand, only a small amount of the total installed wind capacity can be relied on firmly to be providing electricity; the Australian Energy Market Operator currently assumes only 8.6 % for summer and 7.9 % for winter peak demand in South Australia (more precisely, for every MW of wind-generating capacity installed, the Market Operator can only rely on a statistically ‘firm’ 8.6 % of that capacity being available during 85 % of the top 10 % highest demand periods of the year) (Australian Energy Market Operator Ltd. 2013c). During periods of low wind generation, the cost impact is minimal. Pre-existing margins of reserve supply, which insure against the sudden loss of fossil-fuel generators, can also cover the wind variability. As wind-power penetration increases, however, the cost implications become ever more daunting. These subsidised, variable generators supply electricity at low marginal costs (e.g., no fuel requirements, no need for permanent staff at the power plant, etc.). This removes potential generating hours for other (baseload) generators with higher marginal costs to sell power and raise revenue. However, little of this dispatchable generation can permanently exit the market. Most of it must be retained to cover periods of peak demand when wind is generating little electricity. South Australia has 1473 MW of existing and committed registered generation capacity from wind, but the maximum ‘firm’ contribution is only 93 MW (Australian Energy Market Operator Ltd. undated). Just 60 MW of coal has been taken out of service (Australian Energy Market Operator Ltd. 2013a) and the market operator has not been advised of any plant retirements within the 10-year planning outlook (Australian Energy Market Operator Ltd. undated). In the eleven years since wind first entered the South Australian market, registered generation capacity increased 62 % while peak demand grew only 13 % (Figure 5). South Australia has been through a period of system overbuilding (Brook 2010), exemplifying Tainter’s “complexity spiral” whereby societies become more complex as they attempt to solve problems, with increasing costs and diminishing returns as the complexity increases (Tainter 1990 cited in Palmer 2014). Perversely therefore, the addition of variable, low marginal-cost generators gradually places upward pressure on overall system costs, in order to keep all necessary generators in the market (Ueckerdt et al. 2013). There is already evidence of this effect in South Australia.
Initially, the average wholesale price of electricity in South Australia declined from a spot price of > $80 MWh-1 in 2009-2010, to $42 MWh-1 in 2010-2011 (Australian Energy Regulator 2013). The decline in wholesale price was due in part to wind generators sometimes bidding at negative prices because of their ability to earn and sell renewable energy certificates to cover their costs (Australian Energy Regulator 2012). However in 2012-2013, the South Australian wholesale electricity spot price rose by over 70 % (Australian Energy Regulator 2013). The main driver of this rise was a price spike in autumn. This was unusual; autumn is a period of typically subdued demand, and the event occurred against a backdrop of generally lower demand in the National Electricity Market (Australian Energy Regulator 2013). The Australian Energy Regulator attributed the price spike to commercial decisions (i.e., cost control) from non-wind suppliers to take some generating capacity offline, which increased the wholesale price of electricity (Australian Energy Regulator 2013). The Australian Energy Regulator highlighted that the State’s reliance on wind-generated electricity had driven down spot prices, thereby eroding the returns for other generators. During this event, South Australia’s electricity imports were at their highest for six years (Australian Energy Regulator 2013). This illustrates system costs rising perversely from increasing reliance on subsidised, variable renewable energy generators whose output is uncorrelated with demand.
Another reliability issue is the provision of necessary ancillary services to the network to ensure systems stability and power quality, such as frequency-control capability and reactive support (Australian Energy Market Operator Ltd & Electranet 2014). These services are provided by ‘synchronous’ generators, typically traditional coal and gas generation or hydro (in some states), where electricity is generated through turbines spinning in synch at close to 50 Hz. Ancillary services are a physical requirement of any electrical system and have been necessary since the development of reticulated power (Australian Energy Market Operator Ltd. 2010a). However as shown, increased wind participation displaces traditional (non-hydro) synchronous generators from the market. The associated ancillary services reduce or disappear (Australian Energy Market Operator Ltd & Electranet 2014).
The rapid influx of wind generation, combined with proposals for over 3000 MW of additional wind generation (Australian Energy Market Operator Ltd. 2014a) spurred the Australian Energy Market Operator and transmission network operator Electranet to “identify existing limits to secure SA power system operation with high levels of installed wind generation and PV relative to SA electricity demand” (Australian Energy Market Operator Ltd & Electranet 2014). The report stipulates that the asynchronous generation of wind and solar PV “by themselves, are not able to provide the required controls to ensure system security” (Australian Energy Market Operator Ltd & Electranet 2014, p. 2). The report finds that South Australia is able to operate securely with high generation from these sources, even more than 100 % of demand, provided at least one of the following two conditions are met: (a) the Heywood Interconnector linking South Australia and Victoria is operational; and (b) sufficient synchronous generation, such as coal or gas thermal generators, is connected and operating on the South Australia power system (Australian Energy Market Operator Ltd & Electranet 2014, p. 2).
AEMO and Electranet examined the credible event that future market conditions could push the number of synchronous generators in South Australia to zero at any given time, and this coincided with a loss of interconnection. They found:
“Where SA has zero synchronous generation online, and is separated from the rest of the NEM, AEMO is unable to maintain frequency in the islanded SA power system. This would result in state-wide power outage”. (Australian Energy Market Operator Ltd & Electranet 2014, p. 12)
This finding provides insight into how South Australia needs to view variable renewable energy. In electricity terms, South Australia is not, in normal circumstances, an island. The current and future success of integrating variable renewable energy in South Australia hinges on the reliability provided by the rest of the NEM network. In that context, pursuing high penetrations of variable renewables in South Australia, as an end itself, becomes a parochial pursuit more so than a meaningful contribution to decarbonising the National Electricity Market. Proposed solutions to mitigate this risk include payments for minimum synchronous generation to remain online, development of a new market in ancillary services, network augmentation and even curtailing supply from wind and photovoltaics (Australian Energy Market Operator Ltd & Electranet 2014). This again points to system costs that are not represented by technology-specific metrics such as capital cost or levelised cost of electricity of the renewable generator. Such costs would spread nation-wide were other states to follow South Australia’s lead, with each new addition of variable renewable energy eroding the buffer of reliability on which the overall system depends and increasing their implicit operating subsidy.
These phenomena argue strongly that South Australia should plan both for more wind integration, but also on how to move beyond a sole focus on maximising wind capacity. Other forms of low-emissions generation must finish the decarbonisation job that wind has begun, and ultimately meet the role of largest provider. There are no credible plans for decarbonisation of Australian electricity that rely on variable supply alone, so this cannot come from merely a wind-plus-solar photovoltaic combination. Studies that have sought to address this challenge have applied varying combinations of energy storage and dispatchable, synchronous ‘clean’ energy (e.g., burning biomass) to support the variable renewable generators (Australian Energy Market Operator Ltd 2013; Elliston, Diesendorf & MacGill 2012; Seligman 2010; Wright & Hearps 2010). The only real question is just what these constant, dispatchable and synchronous sources of supply should be.