Can you be made redundant before you’ve started work?

The problem with a rapidly changing electricity system, is sometimes it changes so fast that by the time you have installed a technical solution, it is already redundant. That may be the case with $180 million worth of large spinning machines which are just about to come on line in the live experiment that is the South Australian electricity grid.

System strength is a timely, if very “inside baseball” topic for Australia’s grids. It’s one of the four horsemen of the electricity system power quality apocalypse, along with frequency control, inertia and voltage control. These are the invisible but critical technical factors that must be managed along with having enough electricity as we move to electricity systems with increased renewable generation.

At a regulatory level, the AEMC has just passed a final rule change to set a long-term framework for system strength. In the real world of the South Australian (SA) grid, local transmission company ElectraNet and system operator AEMO are carrying out final tests on four large spinning machines – synchronous condensers – designed to help keep the SA grid stable by providing system strength in key locations of the grid. These are similar to the turbines that generators use, but don’t actually produce any energy, hence they don’t have the fuel costs

AEMO sees system strength as “the ability of the power system to maintain and control the voltage waveform at any given location in the power system, both during steady state operation and following a disturbance”. The traditional way to do this is with large spinning machines, such as coal, gas or hydro generators that are synchronous with the grid they are connected to. Traditional grids had enough sufficient strength inherently provided by these machines that there was nothing to manage or pay for.

This all changed in SA a few years back as the balance of generators shifted from the traditional synchronous machines to newer renewables plant that use inverters to connect to the grid. Inverter connections mean wind and solar generators are not naturally synchronous).

In periods of high renewables, AEMO would direct some gas generators to switch on to provide this rotating mass and keep the system secure, even though the wholesale electricity price was too low to be commercially viable for them. This cost consumers twice: the gas generators are entitled to compensation for being directed on, while some of the renewables had to be displaced to make way for them. Last year the cost of paying gas generators to be directed on was $50m.

Given these costs, $180m for four synchronous condensers (syncons) that are expected to run for decades seemed like a bargain. While it’s likely that they will work out cheaper than the current arrangements, technology is moving fast in this space. Although it’s only two years since the decision to install them was taken, alternative solutions that are much cheaper have been successfully trialled in Queensland and Victoria.

These approaches involve careful choice of large scale inverters and “tuning” the inverters appropriately. While there may be a slight premium involved in selecting the right inverters, the rest is essentially software. The Queensland project cost less than $1m, according to ARENA who co-funded it.

The Queensland trials on inverter based system strength suggests this technology may be location-specific – it may not work in all parts of the grid at all times. It seems likely that as this approach is refined, however, it will become more widely applicable and more straightforward to implement.

Meanwhile, AEMO has made it clear that it is not yet ready to rely entirely on the synchronous condensers. Even once testing is complete it will still require two synchronous generating units to be online at all times. This should cut directions and their associated cost, by perhaps two-thirds, but it is clear that the syncons are not a complete solution. AEMO appears unlikely to review this (despite the promising results in other regions from inverter management) until EnergyConnect is complete in around three years’ time.

There’s no obvious villain here. ElectraNet was obliged by new rules to take responsibility for ensuring there were assets in the SA network that could provide system strength. It went out to tender for generators to promise to turn on when required, and based on the prices it got back, reckoned it was much cheaper on an annualised basis to buy the synchronous condensers. ElectraNet could only make choices based on the technologies available to it, and at the time, EnergyConnect had not been confirmed, while the use of inverter-based resources to support system strength was yet to be explored. Even so it was still a case of comparing apples with oranges.

It was comparing a contract payment that could be renegotiated after a few years when either the requirements or the potential supply options could have materially changed with a regulated cost to consumers that will endure for forty years. The bids may also have been impacted by an expectation that all three tendering parties would need to participate in the contracting option. It’s not clear why a hybrid option of – say – two synchronous condensers and a lower contribution from generators wasn’t considered.

It also seems that AEMO is somewhat conservative in its use of directions. But it’s their job to keep the lights on, and an insecure power system could result in a black system event, such as SA experienced in 2016. Nonetheless there is a potential disconnect in asking TNSPs to procure whatever is necessary to provide system strength over a horizon of several years and then asking AEMO to decide whether it has the resources it needs on a day-by-day basis. This split of responsibilities has been confirmed by the new rules.

Does this matter? In the end, the annualised cost of the SA synchronous condensers is around $15-16m. The potential cost of paying two generating units to be directed on as necessary was estimated by ElectraNet at $12m, though this could fluctuate from year to year. SA consumes about 12TWh of gird electricity each year, so they can expect to pay around $2 or so extra per MWh for system strength. That’s not huge, even if it does turn out that the same results could be achieved more cheaply with the right inverters set up the right way. But all the breathless talk about Australia becoming a “renewable superpower” is predicated on the assumption that we will be able to produce electricity cheaper than other countries in a carbon-constrained world. If this is going to happen, we’ll need a laser-like focus on driving costs down and solving technical challenges in the most efficient way. That’s not happening in SA right now.