SA black-outs because energy market operator ignored BoM forecasts
The Australian Energy Market Operator has been relying on private weather forecasters who have been predicting milder conditions than the Bureau of Meteorology, and likely led to the AEMO being under-prepared for the recent weather-related blackouts in South Australia.
It is now clear – from its latest report into the rolling blackout, or load shedding, in South Australia on February 8 – that AEMO does not use BoM forecasters, but instead relies on predictions from private operators Telvent and Weatherzone.
The problem for the market operator on February 8 was that it was relying on forecasts that were 3-4°C lower than what the BoM was forecasting.
Which may explain these words of exasperation by Greens Senator Sarah Hanson-Young in the recent Senate committee hearings into the resilience of Australia’s electricity grid.
“My 10-year-old daughter knew all week it was going to be stinking hot in Adelaide. All the kids at school were talking about how it was going to be stinking hot and they would go swimming after school. I cannot see how the energy operator did not foresee that demand would be high.”
It turns out that it was because they were not looking at BoM forecasts at all.
“For South Australia demand forecasting, AEMO uses an equal weighted average of hourly weather forecasts provided by WeatherZone and Telvent5 based on measurements taken at Bureau of Meteorology (BoM) weather stations at Adelaide Airport and Adelaide.
This was crucial for those blackouts. As AEMO has revealed, it completely botched its forecasts for that afternoon, progressively expecting lower temperatures and more wind, when the opposite happened.
By the time it realised its mistake, it was too late to call on the country’s most efficient gas-fired generator, Pelican Point, to fire up its idle second unit.
The failure and absence of another 300MW of gas-fired generation meant it had not enough capacity to meet demand, forcing it into “load shedding” or cutting power from consumers.
Andrew Stock, a former senior energy industry executive and member of the Climate Council, says he doesn’t understand why AEMO wouldn’t be using the BoM forecasts as well.
“Surely if you have a circumstance like what happened on February 8, and BoM said it will be 41°C two days ahead of time, and 40°C three days ahead of time, why wouldn’t you take that into consideration.”
The difference in forecasts equates to roughly 500MW of demand – that is how much extra air conditioning is switched on to deal with such high temperatures.
With the benefit of advance notice, AEMO not only could have got Pelican Point ready, it could have – like it did in NSW – given consumers sufficient warning and joined with state ministers in asking them to moderate their electricity use.
“I believe the SA load shedding could have been completely avoided if AEMO had been using BoM forecasts,” Stock said.
That advanced warning and pleas to moderate use had a dramatic effect in NSW, cutting around 200MW of demand.
That was crucial for NSW because it now appears that the interconnectors to Victoria were overloaded, 1GW of coal generation in NSW was out of action, nearly 1GW of gas generators failed just at the time of peak demand, and rolling blackouts were only narrowly averted thanks to a decision to cut 300MW of capacity to the Tomago aluminium smelter.
The non-use of BoM data and predictions has stunned many in the industry, and led for calls for AEMO to change its operating practices, which have been criticised intensely in the Senate hearing.
Dr Don Russell, the former Keating adviser who now serves as head of the Premier’s department in South Australia, lamented why AEMO had allowed the inter-connector from Victoria to run at full pelt and called for no back-up as the devastating storms built up in South Australia, eventually knocking down three transmission lines.
“A lot of people have commented that that was quite a risky strategy in an environment where a lot of people knew that an extreme weather event was approaching,” Russell told another sitting of the inquiry in Adelaide this week.
“You might think that looking out the window and listening to the radio might have led them to take a more cautious view, but that would have been straying outside their procedures,” Russell said.
That may be unfair, considering that while AEMO has an office in Adelaide, its main operations rooms are elsewhere.
But the fact that AEMO has been taking weather forecasts from private operators and not the government forecaster has deepened the frustration, particularly in the renewable energy industry, which has copped the blame from conservative politicians, some in business, and many in mainstream media.
Update:
Weatherzone managing director Charles Solomon sent in this statement after the story:
“The comparison offered by your publication is materially incorrect. The statement incorrectly compares a composite hourly temperature forecast with a single daily maximum temperature. The AEMO does not rely solely on single daily maximum temperature forecasts, but instead utilises a composition of forecasts from various sources at its discretion.”
This is what AEMO said in its report on the outages:
Energy use in summer is highly dependent on ambient temperatures, with increased energy use for cooling increasing actual and forecast demand.
For South Australia demand forecasting, AEMO uses an equal weighted average of hourly weather forecasts provided by WeatherZone and Telvent5 based on measurements taken at Bureau of Meteorology (BoM) weather stations at Adelaide Airport and Adelaide. AEMO’s demand forecasting models use a 50/50 combination of forecast and actual temperatures from the same weather stations. Temperatures quoted in this report are calculated as averages of both weather station and weather information provider.
These forecasts are updated hourly and were used to forecast the operational demand for Wednesday 8 February. Figure 7 shows the evolution of the temperature forecast during the day. Errors in the temperature forecasts led to errors in the demand forecast.
Original article posted on Renew Economy, 24th of February 2017 (Link)
