What if—then what?

By Phil Kreveld 

It seems that eventually we’ll end up with a close to 100% solar, wind and battery supported electricity grid. Also apparent is that the mixture of commercial interests, regulatory authorities, the Australian Energy Market Operator (AEMO), and renewed interest of energy ministers to intervene doesn’t augur well for a secure as well as cost-effective future grid.

The Australian Financial Review of February 5 reported that energy ministers have decided on a “sweeping review” of AEMO.

“The review, which will be led by former Treasury official and International Monetary Fund executive Nigel Ray, was agreed at a meeting of ministers in Canberra in December”[.]

“[A]s the energy transition has gathered pace, [AEMO] has acquired significant new powers often at the behest of governments—including a comprehensive grid planning role and the ability to intervene in the market during critical supply shortages.” “[T]he efficacy of how the AEMO receives and responds to instructions/directions from energy ministers to meet their expectations and their responsibilities to the long-term interest of energy consumers [will be scrutinised].”

This is an appropriate introduction into the important discussions of ‘what ifs’, termed ‘contingencies’ and the ‘what thens’ required electrical engineering solutions. They are the reason for AEMO’s headaches, now at risk of becoming migraines because of ‘scrutiny by energy ministers’. This is not to scorn the role of governments in the oversighting of essential services. However, fact is that the ‘morphing’ of renewables with ‘cheap energy’ and grid-supporting batteries with ‘firming’ although politically convenient, leads to ‘contingencies’ actually being realised. The view that we are better positioned to ward off grid failure than Spain (in regard to its blackout of April 28) is not warranted.

In this article, a hopefully clear picture is presented on the limitations imposed by inverter-based generation used by solar, wind and batteries, when replacing synchronous generators mainly associated with coal-fired and gas-fired power stations. Limitations are there to be worked through as opposed to using them as an argument against renewable sources of energy. There are costs associated with engineering solutions to minimising contingencies and these will impact network costs.

Related article: Government announces review of grid operator AEMO

The two critical matters affecting the renewable transition are:

1. Loss of fault (short circuit) current; and
2. Absence of central grid stability control.

It might be surprising that the list is so short, for instance what about inertia, or ‘firming’, and so on. The reason is that we have basically two options only. We can either fashion a grid that apes the one from yesteryear—making everything look like synchronous generators are still with us. Or we can knuckle down to the new game brought about by ‘inverters’. Presently we are sitting on the fence—and, courtesy of the energy ministers scrutinising AEMO, we will remain there.

There are two types (a) credible contingencies and (b) non-credible contingencies. The first type covers generators, or transmission lines developing faults that are foreseeable, etc. The non-credible events are not—they classify as ‘what ifs? Indeed, and ‘what then?’ The repertoire of credible contingencies solely encompasses decades and decades of historical experience. And that is a cause for worry! Because it is all based on ‘synchronous’ history. Right now, there’s sufficient synchronous capacity left in the national grids to use the established distinction between credible and non-credible events—but already as demonstrated—perhaps inconveniently, in the case of Spain, things can go belly-up very quickly.

The single most important distinction between synchronous generators and inverters is their short circuit current capacities.

Synchronous generators: between 3 to 5 times rated current for 3 to 5 seconds.

Inverters: between 1.2 to 2 but at the higher value only for milliseconds.

As will be explained, the need for centralised grid stability control derives from the much shorter time spans for decision making as a result of inverters as generation sources; from seconds to 100 milliseconds, or less.

The easiest way to transition to inverter-based grids is to centralise with a powerful grid forming inverter base, servicing the rest of grid-following inverters. We can’t do this because of the large distances between solar and wind generators, and the consumption (load) centres they service. Long distances between synchronous generators make life difficult enough—and more so with inverters.

Here’s a simple contingency to illustrate the need for rethinking our grid planning. Rather than resorting to graphs and formulas, description of physical processes are used. A generator is supplying a distant load via a transmission line. A short circuit occurs in the transmission line, and as is to be expected a protective circuit breaker should disconnect the line. The distant load might well be supplied by another generator if it is part of a network.

First, we’ll consider the case of a synchronous generator. When the transmission line short circuits, a massive current flows from the generator to the fault, many times larger (as mentioned above) than its normal full-load current. The circuit breaker control acts within a short time span of perhaps 150 milliseconds. Let’s take a grid forming inverter as the nominally equivalent device. The moment the fault occurs, voltage at the fault collapses to zero (it also did so in the first case). A larger current than the normal full load current flows, perhaps initially for a few milliseconds, twice the full load current but then reduces to perhaps 50% or less so as to protect the inverter semiconductor switches. The circuit breaker opens more slowly since opening speed is usually an inverse function (the lower the current, the longer the time to open).

Let’s see how the two types of generators behave before the circuit breaker clears the fault on the transmission line. The synchronous machine speeds up because suddenly its load has ‘disappeared’ while the short circuit is maintained. The very high current that flows causes a strong magnetic reaction on the rotor which has been speeding up and restrains it, if not completely, from losing synchronism. Typically, about 100 milliseconds or more is the time span available for the circuit breaker to act (the critical clearing time) to allow the generator to retain synchronicity.

The grid forming inverter is faced with a very low voltage at its terminals when the short circuit occurs. Its controller response is to require more current so as to build up the voltage but soon hits the maximum allowable current in order to protect its switches. The controller compares the small amount of power being ‘fed into the fault’ against the power reference (note: the synchronous generator also is controlled by comparison toa power reference) rapidly increasing frequency in response. Unlike the synchronous generator there is no physical restraint and it loses synchronism before the circuit breaker has cleared the fault. However, as already mentioned, the short circuit current capability is low and that is a problem for reliable, predictable fault clearing by the circuit breaker.

The synchronous generator although supplying high current into the fault, it is able to maintain voltage at its terminals. That means that the voltage on another unaffected circuit of transmission line is maintained. In contrast, with the grid forming inverter, there is voltage collapse which therefore affects loads connected to the healthy circuit.

As happens when simplifying a reasonably complicated subject, informed readers will point out a couple of things:

1. Inverter current although restricted to protect the switches can be allocated to support voltage in favor of power (and therefore, frequency);
2. A ‘software inserted impedance’ in the control circuit can prevent the collapse of internal voltage; or
3. The grid forming inverter can be switched to grid following mode (this being done before voltage collapse).

Related article: How to bake an electrical mille-feuille

If you dwell on the third point, it will be apparent that an inverter-based grid is truly a new ballgame! To review the matters already raised, attention now has to be paid:

1. Where in the grid will grid forming inverters be situated (recall that the south eastern grid has some 3000 buses, ‘connection points’);
2. Which of these will be switchable to grid following mode;
3. Where will designated grid-following inverters be situated;
4. Where will voltage compensation devices be placed (presumable at strategic points where voltage collapse will cascade throughout the grid);
5. What will the shortest time base be for control of frequency and voltage;
6. What will be the single most important control parameter (probably voltage);
7. What areas of control will require machine learning; and
8. What will human-control degrees of freedom be.

If you’ve read this far, some thoughts will have occurred:

1. What possible useful function could energy ministers play in the transition, except negatively by giving AEMO a very hard time, distracting it from nutting out the above, quite inadequate list of technical matters all part of grid security and voltage and frequency stability;
2. That AEMO is facing a hard choice between being the ‘energy trading floor’ and grid reliability chief;
3. That in the absence of resolving the above question, and leaving it to the energy ministers to make decisions in favour of cheap electrical energy for the punters, the very opposite will be achieved;
4. And finally that the ‘what ifs-then whats’ will play out in islanding, black outs and recriminations to make the legal profession very happy.

It provides no comfort to end this article on such a negative note.