The big switch (off)
A unique set of circumstances has South Australian owners of rooftop solar panels staring down the prospect of having their grid exports curtailed—a scenario that may well lie in all of our futures.
Andrew Reddaway explains.
In recent months, the organisation that operates Australia’s main electricity grid has said that it needs the ability to remotely switch off household solar systems. For many people this is a disturbing prospect since their decision to install solar was partly motivated by a desire to become less beholden to big utilities. No one relishes the idea of Big Brother messing with their personal property, but it’s clear that some way to manage rooftop solar generation becomes necessary when the number of panels statewide approaches the point where they can supply an entire state during sunny periods.
Keeping the lights on requires a minute-by-minute balance between supply and demand. The consequences of under-supply are obvious, but a grid-wide oversupply is also a problem as it would cause over-voltage and also raise the grid’s frequency above a safe level, potentially causing blackouts. Balancing supply and demand is critical, but it becomes an impossible task if the grid’s dominant generator operates without regard to that balance.
South Australia: a vision of the future?
In many parts of Australia, it will be a decade or more before rooftop solar reaches such levels of uptake. The exception is South Australia, where rooftop solar has grown very fast. The state’s sunny climate boosts the output of solar systems and increases their up-front subsidy, which scales by expected generation. Uptake of rooftop solar has also been spurred by high retail electricity prices. South Australia has a long, stringy electricity network that is designed to support air conditioner demand during a heatwave. Most of the time, however, demand is low, since the state has relatively little heavy industry. This “peaky” level of demand translates into an expensive electricity bill, because a portion of consumers’ per-unit energy cost is levied to maintain SA’s extensive network. Energy costs are also influenced by the increasingly expensive gas that is burned to power the state’s remaining fossil-fuel generators when the weather’s not suitable for renewables.
The risk of a blackout
SA’s high uptake of rooftop solar already poses a risk of blacking out the state’s entire grid under some exceptional combinations of conditions, according to the Australian Energy Market Operator (AEMO). Consider a transmission line falling down near Adelaide due to a storm. The cable discharges its electricity into the ground, causing voltage in lines throughout Adelaide to temporarily collapse to a fraction of the normal level, in a manner similar to water pressure dropping when a mains water pipe is ruptured.
In such an event, it’d take a fraction of a second for automatic switches to activate and disconnect the damaged transmission line and nearby sections of the grid. Years ago, an event like this posed little risk of a state-wide blackout, because electricity was supplied predominantly by centralised generators that could keep operating through the brief period of abnormal voltage.
Unfortunately, many current rooftop solar systems are not so robust and may switch off as soon as they experience a severe voltage drop. If such a drop occurs at a time when Adelaide’s solar systems are supplying most of the state’s demand (for example, midday on a mild, sunny Sunday) then their sudden, coordinated disconnection could black out the entire state, because other generators won’t be able to start up quickly enough to take over the load.
Usually AEMO will recognise such a blackout risk ahead of time and avoid it by running centralised generators in addition to rooftop solar, exporting the surplus electricity to Victoria. However, the interconnector between SA and Victoria is not always operational—it’s been out of action six times between 2016 and 2020. At such times SA operates as an electrical island, and AEMO’s options are more limited.
The odds are low of all these conditions—a transmission line falling during a sunny time of low demand while disconnected from Victoria—being met simultaneously. However, the potential consequences are so severe that AEMO has determined that it needs a way to cope with this eventuality.
Several possible avenues are discussed below.
A potential shutdown
Without a solution to the blackout risk noted above, AEMO has said that it will forbid new rooftop solar installations in South Australia, shutting down the state’s rooftop solar industry. This possibility has been reiterated by South Australian energy minister (and Premier) Dan van Holst Pellekaan.
Although SA is a small state, solar installers currently employ over a thousand South Australians. If installations ceased, those jobs would be destroyed, along with many small businesses in the solar sector. Households and businesses would lose a key method to reduce electricity bills, and environmental benefits would not eventuate.
Curtailing rooftop solar
Most generators in the electricity grid are under AEMO’s control, receiving instructions every five minutes to increase or decrease their power. They are prioritised based on the electricity price they offer, as well as network constraints such as congested transmission lines. Rooftop solar is an exception—it always gets first priority over other generators because it’s not managed by AEMO.
After investigating many potential solutions to the blackout risk noted above, AEMO has concluded that it needs the ability to limit rooftop solar generation. This ability would only apply when South Australia is operating as an island, disconnected from Victoria. AEMO would determine a minimum level of operation for large, robust, centralised power plants—for example, 500 megawatts (MW). (By comparison, SA’s average demand is around 1800 MW and its historical maximum is 3300 MW.)
If demand on centralised generators falls toward this level, some rooftop solar generation would be curtailed to avoid any further reduction in demand. When this happens, the rooftop solar that is curtailed would be replaced by centralised generation—which in South Australia would usually be fuelled by fossil gas.
How much curtailment will occur?
In the short term the amount of rooftop solar energy curtailed will be insignificant as a proportion of annual generation, as it will only occur during exceptional situations as noted above. This is what AEMO is currently focused on—they’ve called it a “backstop” facility, to emphasize that it will only be used in such atypical circumstances.
However, Australian rooftop solar is growing strongly. AEMO has modelled several future scenarios with varying levels of renewable supply; two of these predict an output of up to 37,000 MW in 2040, which would represent a dramatic increase on the current level of 9000 MW. In such a future, a majority of households would have solar panels on their roof. This level might result in regular curtailment of rooftop solar on low-demand sunny days, as it exceeds average grid demand which is forecast to remain around its current level of about 20,000 MW.
In future, items such as energy storage, hot water tanks and electric vehicles may soak up a lot of rooftop solar generation (see below). But on sunny days, the batteries may well be fully charged and hot water fully heated soon after lunchtime. After that, curtailment will still be necessary to avoid an oversupply of rooftop solar. This will be the case even if future solar systems are robust enough to avoid blackout risks in scenarios such as that described above.
At times when rooftop solar is being curtailed, solar farms and wind farms will have already been curtailed to zero and switched off to maintain a balance between supply and demand. If such an over-supply eventuates, it seems likely that investors will find ways to use this wasted electricity. Opportunities include refining metal or electrolysing hydrogen for export, or exporting electricity directly to Asia via submarine cable. It’s debatable whether investors could justify the capital cost of such ventures to make use of an energy source that’s only available seasonally. These kinds of developments are likely to reduce the amount of rooftop solar curtailment required, but can’t be relied upon to eliminate it.
It seems likely that in future, rooftop solar curtailment will become common, but will still only amount to a small percentage of annual generation. Such low levels of curtailment would be a good result overall, avoiding blackout risks without significantly undermining the environmental or financial benefits of solar systems.
An off-grid precedent
The future electricity grid’s need to curtail rooftop solar generation has a precedent. Many off-grid systems supplying households and businesses already accept power from solar panels via a standard grid-connect inverter and the normal household switchboard. When the off-grid batteries are full and excess solar has nowhere to go, solar generation must be switched off or at least reduced to a power level that equals consumption. Some off-grid systems achieve this by increasing the electrical frequency to a value which forces the solar inverter to switch off. More sophisticated systems (such as the Australian-made Selectronic SP Pro) can curtail the solar inverter’s output gradually by communicating via a data cable.
A load-shedding precedent
Households are already subject to intervention by distributors during exceptional events. When generation is insufficient to meet demand, suburbs are sometimes intentionally blacked out (“shed”) to avoid a state-wide collapse—as described in our article in Renew issue 147.
Compared to such load-shedding, switching off rooftop solar has a lower impact on people’s lives. On the other hand, the financial impact is quite different. Although load shedding is undesirable, it acts to reduce household electricity bills by reducing the amount of electricity they must pay for. Conversely, solar curtailment will slightly increase electricity bills for solar households.
How would solar be curtailed?
This issue has caught AEMO unprepared. In 2017 AEMO expected that 1000 MW of rooftop solar would be installed in SA by 2022, but actual installations had already exceeded 1200 MW by the start of 2020. AEMO wants a solution in place for spring 2020, but is seemingly still in the early stages of identifying options and consulting with stakeholders.
The overall concept is that occasionally AEMO will identify an amount of solar power (such as 200 MW) that must be curtailed in an area and ask the local electricity distributor to make it happen. Here are the options we’ve seen mentioned by AEMO to curtail rooftop photovoltaic (PV) supply.
It’s possible to install devices in local neighbourhoods that can increase the voltage to approach the maximum value allowed, which is 253 V. This will induce rooftop solar inverters to curtail their panels’ generation, or switch off entirely. This method is fairly crude because even if a device is installed in each street, nearby houses may be curtailed more severely than those further away since they’ll experience different voltages. Also, such high voltages may reduce the lifespan of household appliances.
An advantage of this option is that it would apply curtailment across both existing and future solar systems, minimising the impact on any individual household. Voltage management devices may also be useful in the local distributor’s everyday operations to regulate local street voltages. This problem is described in our discussion paper “Solar Indigestion”, and is the subject of an ongoing Renew research project.
Smart meter switch-off
It’s common for the billing meter to be replaced when a solar system is installed, as many old meters can’t measure exports to the grid. New solar installations could be connected to a special type of smart billing meter that can disconnect the solar circuit when instructed by a signal from the electricity distributor. Under this option, curtailment would occur only in solar systems equipped with such a meter. This isn’t ideal because the impact would be concentrated on owners of recently-installed systems, while previous installations would remain unaffected.
It’s unclear whether such meters are widely available in South Australia, so solar systems may be connected to a smart meter that disconnects the whole property, rather than only the solar. It’s clearly undesirable to black out houses just as a by-product of switching off their solar! Another downside of this option is that the cost of some solar installations would increase as they would need to be connected back to the household billing meter, rather than to a closer switchboard.
Changes in the design or provision of billing meters can be tricky and problematic, as seen in Victoria’s rollout of smart meters. If this option is pursued, it’ll need careful consultation and management.
Dynamic export management
This option would communicate directly with the solar inverter, instructing it to reduce generation via existing built-in functions for demand response. Rather than simply reducing generation, this option could instead focus on reducing exports. If you’re consuming all your solar generation on-site, you wouldn’t be affected. If curtailment is required, this is the fairest way to share it between solar owners because it preserves the highest-value solar generation. Like the previous option, this is probably only sensible for new solar installations, as AEMO is unlikely to proactively send out technicians to every house with an existing solar system, although the required modifications could perhaps be implemented on older systems when their inverter is being replaced. Solar inverters sold in recent years are already equipped with a port to connect a curtailment device called a DRED (Demand Response Enabling Device).
If the inverter is to receive instructions via the internet, AEMO is concerned about reliability. For example, this method may fail if the household wi-fi isn’t working, which is a real possibility if the grid is recovering from a blackout and a household’s solar system manages to restart before its internet router. It’s also conceivable that computer hackers could cause a blackout by coordinating an attack on millions of solar systems.
An advantage of this method is that the same infrastructure could be used by the local distributor to manage local network constraints. In current practice, distributors frequently require solar installations to have a static export limit of 5 kW or 3 kW. This is a “brute force” approach that curtails solar generation whenever solar conditions are good and the solar household isn’t consuming much electricity. Often this curtailment is unnecessary, especially if neighbours are consuming significant power. If distributors could dynamically manage exports, static export limits could be relaxed or eliminated. Only when the local network is actually struggling would solar exports be constrained.
Dynamic export limits are already required in SA for solar systems larger than 200 kW (about 600 panels), perhaps providing the groundwork for smaller solar systems.
Will this affect the grid’s transition to renewable energy?
Our electricity grid is heading away from fossil fuels and toward renewables, especially solar and wind generation. We’ve detailed this transition and its challenges in a series of discussion papers. Happily, AEMO’s moves to curtail rooftop solar won’t undermine this transition. Curtailment will only occur at very sunny times when there’s plenty of solar energy available to power suburbs and factories and charge up energy storage. In a fully renewable grid the real challenge is a cloudy, calm week—curtailment won’t occur during these events.
Householders are leading the way to a fully renewable grid. Of AEMO’s models, the scenario that predicts the highest level of renewable energy sees 94% of the grid’s energy coming from renewable sources, with nearly one-fifth coming from rooftop solar. To reach this level, rooftop solar would need to grow to nearly four times its current size.
Anything that deters rooftop solar may impede—or at least delay—the grid’s renewable transition. Since curtailment probably won’t affect annual bill savings greatly, it seems unlikely that it will deter solar installations to any great extent. For example, rooftop solar was not slowed down when distributors started imposing export limits several years ago.
Rather, curtailment of rooftop solar is necessary if rooftop solar is to quadruple in size. At such a level it would dominate the grid during sunny times of low demand, so to avoid blackout risks it would be essential for AEMO to be able to control solar generation when necessary.
Who will bear the cost of this initiative?
Whichever method is selected to enable solar curtailment, the new electrical hardware and supporting IT systems required won’t come for free. Neither the extent of these costs, nor who will cover them, is known yet. Money spent by local distributors will be recovered from all electricity consumers in their area via electricity bills. Solar owners will pay for any extra time or materials incurred by solar installers. Some items could be covered by AEMO, which is funded by all electricity consumers. Finally, governments could decide to devote taxpayer funds.
We’ll get a better idea as the solutions are better defined. In any case, it’s likely that the costs will be well worth it if they enable the continued uptake of rooftop solar, given its financial and environmental benefits. Methods to allocate network costs are beyond the scope of this article, and are subject to a recent request by consumer advocates to change the National Electricity Rules.
Why aren’t rooftop solar systems more robust?
It’s certainly possible for solar systems to be designed and manufactured to avoid the risks mentioned above. Many existing grid-connect solar inverters are already robust enough to keep generating through a voltage disturbance—the University of NSW found that nearly half the inverters it tested had this capability. It seems this ability also has no significant impact on the inverter’s cost.
The problem lies in the Australian Standard that solar inverters must meet. This standard (AS/NZS 4777.2), developed in 2015, implies that inverters should ride through dips in voltage, but the relevant testing procedure doesn’t require this. This is evidently a severe shortcoming in the standard, and AEMO should probably have addressed it years ago. As well as working toward an improved, new version of AS 4777, AEMO is now rushing through special requirements for solar inverters installed in SA. This may limit SA solar owners’ choice of inverter over the next year or so, as manufacturers of non-compliant inverters modify their production to meet the requirements.
Other potential solutions
Apart from curtailing rooftop solar generation, many other measures exist to mitigate South Australia’s blackout risk. AEMO has determined that although these are helpful, they don’t eliminate the need for solar curtailment in the short term.
AEMO’s analysis included the projected uptake of batteries, including those in electric vehicles. The analysis found that extra batteries would only result in a mild reduction in rooftop solar curtailment, as on critical days the batteries tend to fill up quickly and then have no further effect. However, there’s no indication that AEMO has considered the use of smart battery logic, which would check weather forecast data and, on sunny days, would delay charging until early afternoon. This would spread out battery charging and allow it to soak up solar generation throughout sunny times. Current products do not yet include such logic, but it would be relatively simple to develop for future battery models. Some batteries can even be updated remotely with such functionality by the manufacturer.
Hot water tanks are large electrical appliances whose timing has traditionally been set to turn on at around 1 am. Changing the timer to start in the middle of the day to help soak up solar exports is an obvious idea that should have been implemented years ago. However, AEMO estimates that SA’s hot water load is insufficient to avoid the need for solar curtailment. Swimming pool pumps provide a similar, smaller opportunity.
The new, planned transmission line connecting SA and NSW is very significant. Once this is complete in 2023 or 2024, SA will be connected to two states rather than one, creating an additional outlet for surplus generation and dramatically shrinking the state’s chance of being islanded. AEMO has said that if the transmission line project is abandoned, new SA rooftop PV should be disallowed from 2020, and that existing solar systems may also need to be retrofitted with new capabilities.
If the interconnector does become operational, rooftop solar curtailment may become unnecessary in South Australia for some years. However, such transmission isn’t a permanent solution. Eventually rooftop solar throughout the eastern states will grow large enough to supply Australia’s entire demand for electricity during sunny times of low demand, and the blackout risk will re-emerge.
To help strengthen its grid, SA has also commissioned four large pieces of equipment known as synchronous condensers. These help with several issues related to the energy transition, but not this specific blackout risk.
Could I opt out of solar curtailment?
Options exist for solar owners who are set against the idea of giving someone control over their solar systems. Distributors could offer a choice between a static export limit (of, say, 2 kW) and dynamic export management (i.e. curtailment) with no static limit. Both would ensure that the solar system doesn’t significantly contribute to the blackout risks noted early in this article. Given such a choice, some householders would select the static export limit, but most would probably opt for dynamic export management. Although it gives the distributor control over the householder’s private property, it would deliver superior bill savings and environmental benefits by enabling greater exports over the year.
One clear option is to sever your grid connection and install an off-grid solar system. However, this has several downsides, as explained in our article “Should you quit the grid” in Sanctuary issue 36. Off-grid systems (and their future battery replacements) are expensive and consumer protections are limited. Without a grid connection, excess solar generation has nowhere to go, wasting the environmental benefit and revenue that come with exporting it to the grid.
How can I minimise any curtailment?
As mentioned above, curtailment by AEMO probably won’t amount to a significant percentage of annual generation. We’re advocating for any curtailment to apply only to exports to the grid, rather than also applying to generation used by appliances onsite. If this condition is adopted, you can minimise curtailment by using electricity during sunny times. You can set appliances that use a lot of power—e.g. water heaters, pool pumps, vehicle chargers—to run during the middle of the day. Doing the same with smaller appliances such as washing machines will also make a difference. Smart devices are available to help adjust some appliances’ power on-the-fly to suit solar exports.
Taking these steps has the benefit of enhancing your solar bill savings, because solar is worth much more to you when consumed onsite than it is exported. It would also help avoid the issue of high voltage levels in your local street, which can cause solar inverters to shut down.
Solar is no longer alternative energy—it’s a mainstream, major contributor to our electricity supply. Over the past decade, rooftop solar systems have matured and gained many features to help them integrate with the grid. Curtailment of solar generation (when required) is another such feature, and it’s necessary to enable solar to spread to a majority of household roofs and support the grid’s transition to renewable energy. Curtailment is likely to affect only a very small percentage of annual solar generation.
Several options exist to implement rooftop solar curtailment. Renew is part of AEMO’s consultation process on this issue, and we are advocating for methods that minimise impacts on solar owners while maximising benefits for all consumers. For example, rather than switching off solar systems altogether, it’s better to curtail a portion of their generation. And if curtailment is required, it’s fairer to curtail exports to the grid rather than solar generation supplying household appliances. We’ll keep a close eye on this issue and provide updates as it progresses.
Look for the full paper, including references, at renew.org.au/research.