Major blackouts prompt rethink about frequency control on the grid

Source: Unsplash

The wind industry – or more specifically the software settings on many wind turbines – copped much of the blame for the state-wide blackout in South Australia in September 216. But it was the response, or the lack of it, from some of the state’s biggest gas generators that really worried many energy experts.

Their fears were confirmed in August last year when, on the first day of Scott Morrison’s prime ministership, lightning strikes took out major transmission links between NSW and Queensland, prompting a series of events that caused two states to be “islanded”, tens of thousands of people to lose power, and a realisation that all is not well with the electricity grid.

That incident in August last year became the catalyst of a major rethink about the way frequency control and ancillary services – an incredibly important response to system faults that can more or less decide whether the lights will go out or stay on – is managed in the country’s main grid.

While South Australia sailed through without issues and was the only grid to emerge unscathed in that August event – thanks in part to the quick and versatile response of its big batteries – the Queensland grid was shaken by the failure of many big coal and gas plants to do what had been expected of them, and respond to the frequency changes.

See also: On first day as PM, Morrison learns difference between Big Battery and Big Banana

The Australian Energy Market Operator – and many energy experts – say it highlights the need for changes to rules to try and arrest the decline in frequency performance in the NEM, sheeted home to the relaxing, or in some cases, the absence of governor controls in big coal and gas generators.

They pointed to growing concerns that big coal and gas generators had been relaxing their governor controls, and altering their “dead-band” settings to the point where they were not responding to major frequency variations until after they had moved beyond safe operating levels. This was confirmed by the August event in Queensland.

“The lack of PFR (primary frequency response) from some generating systems contributed to significant technical challenges in arresting and controlling power system frequency, particularly in the earlier stages of the event,” AEMO noted of the Queensland event.

Some experts, such as Advisian’s Bruce Miller, describe the lack of frequency control on the main grid as like driving a car with the steering wheel getting looser and looser.

He and others blame the relaxation of settings to the way the market is designed. They say market signals – from 45 different FCAS markets – are taking precedence over the technical requirements of the grid.

“The FCAS markets are not providing the necessary signals to keep the system frequency under control,” he said in a recent presentation at a seminar called to discuss the issue. “The entire system is at risk.”

Miller worked with power engineering expert Peter Sokolowski – and a group of others including Kate Summers, Jack Bryant and Lasantha Meegahapola – to put forward a change in the rules, largely in line with that wished by AEMO, and it appears they have been largely successful.

The major rule-maker, the Australian Energy Market Commission, has relaxed is dogmatic adherence to market signals and accepted an interim proposal to make some sort of frequency response compulsory on all generators.

That idea is not without controversy. The Clean Energy Council, and many renewable energy developers, suggest requiring the entire generation fleet to meet the PFR requirements may be inefficient and a case of “overkill”, when a more targeted method such as direct contracting may be more cost effective  interim measure to restoring frequency control.

Battery storage providers, with an eye on the growing FCAS revenue market, and knowing that they are probably best placed to provide the range of PFR – thanks to their speed and versatility, agreed.

In the end, AEMO will be allowed some discretion, and individual generators can negotiate with the market operator. Miller suggests wind and solar farms are going to have to get used to providing PFR (it may require a small dialling down of their output to give them manouvrabililty – if the system is to move to 100 per cent renewables.

Another energy expert, Kate Summers, says that in order to ensure that the power system can withstand events in the future, new renewable plant should in the future provide a dynamic active power response that is sensitive to frequency.

But she notes that some older type 1 and 2 wind turbines may not be able to deliver this service, and should not be required to because their influence on the power system is small. “They must not be unduly penalised for this as it was not a requirement at the time they connected and their influence on the system is small.”

Miller notes that frequency control is effectively free because it is a control function that all generators should be capable of. On the other hand the provision of “reserve” i.e. the ability to respond, is not free because it affects the dispatch of generation and accordingly attracts an operating cost. Battery (or other types of ) storage has a role to play in reducing provision of reserve by generators on the system.

AEMO chief Audrey Zibelman welcomed the changes.

“The changing nature of the power system under-pinning the NEM has resulted in a lack of effective frequency control, which poses risks to system security,” she said in a statement.

“This rule change calls upon capable generation plant, both existing and new, to automatically respond to frequency changes and ensure that power system frequency is adequately controlled. It is a critical step towards improving system resilience and predictability.”

(The AEMC draft decision on FCAS was one of 16 major decisions released by the rule maker on Thursday, along with another dozen or so reports from AEMO and the energy regulator. The normal pre-Xmas rush).

 

 

Comments

3 responses to “Major blackouts prompt rethink about frequency control on the grid”

  1. Ian Avatar
    Ian

    Traditionally, when AC was produced by large spinning generators, the frequency was a function of the rotatory movement of these and the load applied to them. Ie the more load applied, the more these things would slow down. This was compensated for, by cranking up the turbine driving the generator either by giving it more steam or more high pressure water etc. and the reverse when the load was light.

    Electronic inverters attached to windmills and solar panels don’t naturally have this simple relationship between frequency and load. These are designed to produce any frequency you like.

    Someone might correct me, but inverters could export electricity at a frequency of 51hz or 49Hz which would make for some interesting harmonics on the grid line. They could even export large quantities of electricity at exactly 50hz regardless of the load, presumably then the line voltage should rise.

    As more electronically generated AC enters the grid, so frequency as a signal for rotating generators to increase or decrease output will be less effective. Presumably phase shift will also become more of an issue.

    Here’s a question: Given that inverters could be designed to export any frequency and can electronically adjust any phase shift could such an electronic devise be installed on the grid circuit to import say electricity from a struggling grid at a slightly low voltage and low frequency , modify this and export it at the correct voltage and frequency without any batteries etc

  2. Seriously...? Avatar
    Seriously…?

    Maybe they should talk to some audio engineers. Any piece of basic recording software has to control frequency and sync to other devices at rates up to 128kHz. Now, sure, that’s not putting much electrical load on. But adjusting to frequency and voltage at rates way over 50Hz should be child’s play, as long as there’s fast-acting spare capacity (batteries) in place. Now, managing phase over a network of thousands of kilometres might be more complex–maybe even a fraction as complex as syncing hundreds of tracks and multiple plug-ins and external devices etc. at 128kHz (and perfectly in phase at every point). I mean, we’re talking consumer electronics here. Actually, most people playing group games over the internet are probably creating more complex syncing problems than the electrical network faces.

    But I guess the reality is that the electricity network is still built around 1950s or even 40s or 30s technology. Time to make the giant leap to the 21st century, and digital control. The place to start would be SA, move to digital control and syncing there as an exercise.

  3. Ray Miller Avatar
    Ray Miller

    One has to seriously question the technical and managerial capabilities of Australian engineering with such a fundamental flaw built into the NEM. Maybe it goes back to a combination of Uni training, lack of critical thinking skills and not standing up to the bean counters, but the deficit is laid bare for all to see. Now you add the sudden revelation by AEMO that we have an unstable grid connection to a number of commissioned Solar Farms who went through the current approvals process only to find out at the operation stage that again our power engineering people failed to undertake the system modelling they should have done years ago. The Solar business should be able to claim damages against AEMO. Then we have the “gold plating” of the transmission network due to a number of reasons both technical and planning with the consequences still playing out in the high price of energy for years to come. The consumers should be able to bring a class action against the transmission companies for their engineering and management failure.

    Any way you view these major failures everyone needs to acknowledge some fundamental truths about Australia’s engineering capabilities, they are clearly sub par and something needs to be done urgently.

Get up to 3 quotes from pre-vetted solar (and battery) installers.