Over the next few months, with the right wing shift in the federal Coalition government, and state and federal elections afoot, expect to hear a lot about “baseload” and “intermittent” generation.
It will be the core of the conservative push for more fossil fuel generation, particularly coal. They argue that because “coal” is “baseload”, it must therefore be “reliable”. And wind and solar are intermittent, so they cannot be relied upon to keep the lights on.
It’s political rhetoric that belies the reality of the electricity system, the biggest and most complex machine in the country. Australia’s grid has challenges, but they are not necessarily ones that can be solved just by having more “baseload”.
What is really needed – as the Australian Energy Market Operator, chief scientist Alan Finkel, and any number of other independent experts point out – is dispatchable and reliable generation, one that the grid operator can count on, in times of peak demand and heat stress.
And the answer does not lie in traditional “baseload” generation – the more than 100 trips of big fossil fuel plants since December, often at times of soaring heat, underline that point.
But there’s something else, and it’s about the way that various generators – traditional fossil fuel and “synchronous generators”, and inverter based renewables like wind and solar – operate at times of great stress for the grid.
And it is at these times, new analysis shows, that the traditional views of baseload and intermittency are turned upside down.
At a recent seminar in Melbourne, at the Climate and Energy College, one of the country’s leading electrical engineers, Kate Summers, showed these series of graphs that illustrated that – at moments when stability can be won or lost – it has been wind and solar that have held firm, and acted as what one might consider to be “baseload”.
And it has been coal and gas that has proved “intermittent” at the very minutes that stability is needed.
Take last year on February 10 in the heatwave when the grid in NSW was under severe stress. You may remember, this was when two major coal units at Liddell were out of action, output at other coal generators was reduced because of heat stress and the two biggest gas generators also failed in the heat. It very nearly caused a state-wide blackout.
This graph below shows some fascinating insights into what was happening at the critical points.
The dotted red lines are the normal operating frequency bands. Frequency needs to be controlled closely around the middle band of 50Hz, although the National Electricity Market allows it  to range between the top and bottom bands.
The thick red line is where frequency was when the NSW grid was under stress late on that afternoon on February 10, and after the loss of a large gas unit as the temperatures soared above 40°C.
What’s important here are the blue lines, because they signify what the active power from the various generation types within the NEM fleet were doing in response.
In the top graph is the “scheduled generation”, which Summers notes is also what is known as the “synchronous” generation or “baseload”. Here, it is the sum of coal, hydro and whatever gas is in use. In total, there is around 33GW being produced, but the output, as is visible in the graph, is all over the shop.
“When people talk about the intermittency of renewables, lets have a look at these next two charts,” Summers says in her presentation, which you can listen to and watch here.
The semi-scheduled and non-scheduled generation is the sum of all the wind farms and all the solar farms – 650MW and 300MW of semi scheduled and non scheduled respectively. Note the constant output through the crisis.
“Is this intermittent?” Summers asks. “No. They (wind and solar) are doing the job of baseload while the actual baseload band (the synchronous generation) is misbehaving.” Summers noted that the legal framework had made this type of behaviour acceptable.
Summers points to yet another incident in June this year, when the largest synchronous machine in the system, Kogan Creek, suddenly dropped out of the system.
Same thing happens. Frequency goes outside the normal operating bands, the synchronous generators leapt all over the shop, and the wind and solar stayed constant. On both occasions, it took nearly seven minutes before the system could restore stability.
Ironically,a large portion of the ancillary services costs falls to wind and solar generators and customers who are paying the synchronous generators to control the red line.
Unfortunately, this leads to poor control under the current regime. You would normally expect the synchronous generators  to respond – that is to increase or decrease generation to correct changes in frequency, but not continue fluctuating for minutes on end.
What’s going on here?
According to Summers, the Australian grid has become increasingly unstable – not because of the introduction of renewables but because of the introduction of certain markets and the way they are designed.
The market created for Frequency Control and Ancillary Services, is misnamed, Summers says, and has allowed the relaxation of governor controls on the major generators.
The Ancillary Services (called frequency control) are reserve services for recover from an event and time error correction services, neither of which in isolation are in sufficient to control the frequency.
The control of energy and frequency are directly related yet the market has made separate them into separate services which means that market signals are often conflicting within the needs of the grid and deteriorating power quality and control. “We are throwing basic engineering out the door,” she says.
The synchronous machines (scheduled) in the top chart are most capable and designed to control frequency, the red line, and are getting paid hundreds of millions of dollars to do provide that service. Â The semi-scheduled and non-scheduled machines (wind and solar) in the lower charts are the ones (along with customers) paying the money.
It’s doubly ironic because wind and solar farms are now subject to some of the strictest controls on their output and generation, yet the existing fossil fuel fleet is not.
Summers says the control changes made to the synchronous machines weakens the network, and this has been done by the markets, and that is a major concern. The economic dispatch of the system was never intended to become the primary control of the system. This graph below illustrates the steady deterioration of the accuracy of control.
“We are trying to control a very large complex machine. It only works when we get a coherent response from the control systems of all the individual machines,” she says.
“But right now, we don’t know what the control philosophy is. People are setting rules just to have a market without understanding the philosophy of control of the power system.
“The system has gone from centralised planning with distributed control to distributed planning with centralised control. It is inappropriate to think that the power system can be controlled centrally as the communications delays make it impossible to achieve the response required.
“If we think we can control energy dispatch in ignorance of frequency control, we have lost the plot as changing energy changes frequency.”
Summers’ comments highlight a couple of important points. The energy debate is usually dominated by simple political rhetoric – based around emissions or no emissions, cheap prices or expensive ones, baseload versus intermittency.
That just skims over the surface. Behind the scenes, as the clean energy transition continues, debates are raging about good engineering practices and the design of markets. When markets force machines to do things that you don’t want them to do, that’s a major problem.
And that’s what we have in Australia, particularly when those markets have the effect – by intent or accident – to limit or dissuade the introduction of super fast, accurate and versatile technologies like batteries, both big and small.
We’ll leave Summers with one final point, about this idea of “intermittency”.
She says the term was based on the study of micro grids or the performance of single solar panels, or single turbines, it is inappropriate to scale up the output of micro variation, the data from the NEM illustrates that large diverse renewable resources are far more stable in output than singular sources.
She states that a collection of panels and turbines smooth the variations at the connection point and as more are added, they deliver reliable output in the short term. The issue is knowing what will occur in the next hour or day as a matter of accurate forecasting. These graphs based on the actual 4 second data from AEMO prove that.
Batteries, she says, can only do a little bit, particularly if the largest machines in the system are hunting against each other. “We need to fix that (retrieve primary control of large units and sort the market design), and then we can integrate our batteries comfortably” and have good control of the system making it more dependable.
As AEMO has made abundantly clear in its 10 and 20 year forecasts, there is enough generation to meet Australia’s incredibly strict reliability guarantee of 99.998 per cent. Its concern is in those critical moments. What can it depend on?
As AEMO has made abundantly clear in its 10 and 20 year forecasts, there is enough generation to meet Australia’s incredibly strict reliability guarantee of 99.998 per cent. Its concern is in those critical moments. What can it depend on?
