As the grid moves inevitably towards heavy reliance on renewables and storage, the frequency-setting stability provided by spinning mass will have to be replaced by something smarter.
On the surface, the duck is calm. Below the waterline, it is paddling like mad. And so it is with the National Electricity Market.
The 16 coal-fired power stations that supply about 65% of generation in the NEM spin generators at precise rotational speeds to match the 50Hz frequency in the grid. “If the frequency deviates from that, the generator doesn’t like it,” says Tim Finnigan. “It has to try to stay stuck to that frequency.”
Finnigan, a consultant and engineering physicist, has written a report for the Institute for Energy Economics and Financial Analysis that highlights the obvious: as coal disappears from the grid so will inertia, so we need to stop relying on it.
But it’s not that easy, because inertia in the grid that we know is vital – like a heartbeat.
“If there is a disturbance in the network, say a major transmission line gets taken out or a generator trips out, that creates an immediate imbalance between supply and demand of power in the grid,” Finnigan says, “and all the remaining spinning generators in the grid will immediately deliver extra power to smooth that shortfall.”
They access that power from the stored energy – or inertia – in the heavy spinning mass of their generators. It’s like a speeding truck, he says. Even if no-one’s pushing the throttle, it still packs a lot of energy – if you get in its way.
Tapping that energy will of course result in the rate of spinning to slow. Very quickly, operators will then pull levers at their power plants to get back to speed – to rebalance supply and demand. This provision of “inertia”, Finnigan says, has always been taken for granted. It’s how things are and have always been … and it’s never been monetised.
“In a synchronous system like our NEM it does provide a background form of frequency stability and voltage strength,” he says. “That fleet of large spinning mass provides a very stable background supportive frequency and voltage signal into the grid. Any things causing disturbances tend to be smoothed out by it.”
This lesson in physics describes the tender balance within a fleet of enormous generators. But the world is moving away from all that. It’s easy to guess, then, what will happen as wind and solar – volatile forms of generation without spinning mass – replace thermal generators: inertia and frequency will become less predictable.
“The grid is becoming less stable,” says Finnigan, formerly director of energy at the CSIRO.
As renewables competitively bid into the energy market and shoulder out coal, particularly around midday as rooftop solar exports soar, the grid becomes vulnerable. How can renewables provide inertia to do their bit to help? “They can’t,” he says. “That’s the whole conundrum here.”
Renewables have no inertia, it’s as simple as that. A solar plant includes no spinning parts and the rotors in wind farms turn nowhere near the required rate and are not electromechanically connected to the grid anyway.
Yes, software can be programmed to help renewables behave like they can provide inertia by supplying synthetic inertia, virtual inertia and by emulating synchronous machines. But right now the energy system has its big feet planted firmly in both camps, when it needs to start planning for the inevitable day when there is very little, if not zero, inertia left. “That’s the pathway we need to be thinking more and more about,” Finnigan says.
The same but different
If renewables can’t supply inertia, and renewables will replace fossil-fuelled generation in the grid, what do we call the thing that we once called inertia? Fast frequency response can do sort of the same thing, where energy is dispatched rapidly in response to market signals. The various big batteries around the grid compete to do this, and as more of them are built the market will become way more sophisticated.
“It’s inevitable that the grid is heading towards a very large percentage of renewable supply and very low percentage of synchronous machines … we have to manage that transition. What type of grid and operation design do we want to get there?”
Preserving the underlying philosophy of a synchronous system is like burying your head in the sand.
In the short term, Finnigan says, it may help to establish a market for inertia. But this will mean paying a coal plant, say, for simply having spinning mass on call.
“As the big coal-fired power plants retire, if we just keep loading in renewables the way we are now it’s not going to work – the grid won’t work anymore,” he says. “And putting in synchronous condensers [at renewables plants to provide a type of inertia] doesn’t seem to me to be the answer.”
Grid-forming inverters may provide part of the solution, whereas today’s solar and wind assets are connected via grid-following inverters. When operating in some offshore wind farms around the world, grid-forming inverters have been shown to set the voltage signal in the grid so that passive grid-following devices can use it as a compass. This is similar to the role synchronous machines have in the grid today, he says. “In many ways it can perform better than synchronous machines.”
The technology works similarly in microgrids, “like a master-slave arrangement”.
In a national energy market like Australia, it may take a great many grid-forming inverters to do the job. How many is not known. As the Australian Energy Market Operator works towards rolling out its Integrated System Plan, it will no doubt focus on the role of these inverters.
Until then, the NEM will have to operate as a hybrid system, where synchronous machines and inverter-based renewables muddle along, despite the fact that they operate completely differently. “It’s going to be difficult; they don’t like to work together, they have very different response characteristics. They’re not the best match. But we’re just going to have to work that through.”
Somewhere along the way, a shift to reliance on advanced software, communications and digital systems rather than synchronous spinning machines will occur. Whether we end up with a system of mini-grids strung together with DC links, an ecosystem of microgrids which relies on artificial intelligence or something else, it’s impossible to say.
“Rather than thinking about energy being generated, transmitted, distributed and consumed, it is possible to start imagining a network with generation and consumption more localised and interdependent, with the grid functioning as a tapestry of highly-interconnected regions,” Finnigan writes in the report.
Until then, we might want to get used to the idea that the less inertia we have in the grid the less we need. “We’ve become dependent on it,” he says. “It needs to be there to serve itself; synchronous machines need to be locked together.”
Nothing will go wildly wrong with electrical appliances if frequency deviates beyond the normal operating frequency band 49.85Hz and 50.15Hz. The appliances we use are far more accepting of deviations in frequency than synchronous machines, he says, pointing out that your laptop – which runs on 50Hz at home – will operate just fine in the US where the frequency is 60Hz.
Tomorrow’s grid might be a bit off-beat but it will be much cleaner.