Did solar and wind wilt in the heat?

Critics of renewables are citing last week's heatwave to suggest solar and wind are unreliable and their growth risks power black-outs. Yet they actually proved to be remarkably predictable - key to ensuring reliability.

Renewable energy critics have been keen to use events over last week’s heatwave in Victoria and South Australia to imply that the Renewable Energy Target will place power supplies at risk.

The Australian Financial Review had the headline emblazoned over its weekend edition front page, "Solar and wind can’t stand the heat". The Australian ran a story headlined, "Wind fails test as demand soars", pointing out wind’s low output around midday on Wednesday. It observed that this event “served to highlight a longstanding argument about the impact of intermittent renewable energy on the stability and reliability of electricity networks”.

And Senator John Madigan put out a press release on Friday stating:

“John Madigan has slammed the government’s renewable energy target of ’20 per cent renewable by 2020’ … Every year it gets hot and every year we have both residential and commercial power outages. Increasing our renewable energy target means we are relying more on energy sources that are unreliable.”

Solar and wind power certainly aren’t like a fossil fuel plant whose fuel supply you can readily control (although you can control fuel supply for hydro and bioenergy, which are also supported by the RET). So is their growth about to put us at risk of blackouts?  

The first thing to note is that the RET doesn’t prevent fossil fuel power stations from being built nor mandate their closure – although a number of these plants will, no doubt, run less frequently than they have in the past. For the most part, we have way more fossil fuel capacity than we need, and we can still use fossil fuels in circumstances that, say, the wind is low or the sun isn’t shining.

The second thing one needs to realise is that ensuring power supply reliability is not unlike most other things in life – to cope with challenging events it helps if you can see them coming in advance and prepare accordingly.

To illustrate, consider two alternative situations in which you might find yourself unemployed:

1)     One day at work, after many years of loyal service your boss tells you you’re fired with no redundancy payout. You are taken completely by surprise.  You’ve just used up all your annual leave on a recent holiday and had just bought a new house, mortgaging yourself to the eyeballs... OR

2)     You’ve come to the end of a six-month temporary employment contract. You knew in advance when the contract would expire, so have been looking around for another job and have two offers on the table. In addition, you’ve saved up a bit of money and have no big debts.

Out of the two situations which would you prefer?

There’s no point pretending that solar PV and wind power are some perfect source of energy for meeting peak power needs, just like a six-month temporary employment contract is not ideal. But their behaviour is reasonably predictable and so we can plan around their shortfalls. It is kind of obvious that the solar PV’s output is linked to the how sunny it is. By happy coincidence this happens to correlate to an extent with high peaks in electricity demand over summer. In addition, it is also true that wind speeds vary over time and unfortunately wind speed bears little relationship to electricity demand.

The Australian Energy Market Operator, who is responsible for keeping the lights on, happens to know both these facts and takes them into account in planning how much power supply is needed from other sources to cope with events like last week.

How did solar perform relative to AEMO’s expectations?

In 2012 AEMO released a study they undertook into solar PV which concluded:

In the mainland regions, summer maximum demand typically occurs in the late afternoon, when rooftop PV generation is declining from its midday peak and is operating at an estimated 28 per cent-38 per cent of capacity.

If you look at the solar output charts below from the Australian Photovoltaic Institute you’ll notice that solar PV performed pretty much as AEMO would have expected. For a large proportion of the time when metered demand (demand for electricity from the wholesale market) combined across Victoria and SA exceeded very high levels of 12,500MW, the average proportion of solar capacity producing power across the two states combined was close to the 28-38 per cent range. Tuesday was the only day when the PV output fell short of AEMO’s range around the time of the absolute peak in metered demand, but on the other days it was comfortably higher. What might be surprising to some is that it wasn’t uncommon for solar to be producing greater than a third of installed capacity as late as 5pm (Victorian time). 

Monday (metered demand remained well below 12,500MW throughout day)

Graph for Did solar and wind wilt in the heat?

Tuesday (metered demand above 12,500MW from 3-7pm and peaked around 6pm)

Graph for Did solar and wind wilt in the heat?

Wednesday (metered demand above 12,500MW from 1-7pm and peaked around 4pm)

Graph for Did solar and wind wilt in the heat?

Thursday (metered demand above 12,500MW from 1pm-8pm and peaked around 5.30pm)

Graph for Did solar and wind wilt in the heat?

Friday (metered demand above 12,500MW from 1pm-6pm and peaked around 4.30pm)

Graph for Did solar and wind wilt in the heat?

How did wind perform relative to AEMO’s expectations?

In terms of wind, SA and Victorian farms’ output was highly variable over the heatwave week and there were two occasions around the middle of the day when output was a low percentage of installed capacity. Based on historical data AEMO expects that during the top 10 per cent of summer demand periods, South Australian wind generation contributes at least 8.6 per cent of its installed capacity for 85 per cent of the time and for Victoria its 6.6 per cent.

Now last week during the periods that combined demand across Victoria and SA were very high (above 12,500MW) wind farms in Victoria and SA produced at an average of 22 per cent of their installed capacity. Wednesday was the outlier with very low output of 7.4 per cent, Tuesday it was 20.4 per cent, Thursday 24 per cent and Friday 36 per cent. Across the entire period of demand being above 12,500MW, 92 per cent of the time wind was producing more than 8.6 per cent of its installed capacity. So again, AEMO wasn’t likely to have been caught by surprise.

Furthermore, while wind output tends to less aligned with demand than solar, AEMO, and the market as a whole, have another ace up its sleeve to avoid being taken by surprise – they can forecast the amount of wind power likely to occur some time in advance. 

AEMO’s forecasting system has proved to be remarkably accurate. The chart below illustrates that it’s degree of error in forecasting wind power output 24 hours ahead (light green line) is less than 6%. This gives both AEMO and generators in the market quite a degree of advance notice as to when wind generation is likely to be low such that they can be prepared to fill the gap.

Graph for Did solar and wind wilt in the heat?

Source: Australian Energy Market Operator

Also given that almost all fossil and hydro generators across Victoria and SA have ramp-up rates of 200MW or more within an hour,  the forecasting system’s very fine accuracy within an hour is worth dwelling on.

That chart below shows the system’s forecast for SA in the blue  line and what actually transpired during the heatwave in the red line. As you can clearly see the forecasts were very accurate, although tending to underestimate actual wind power output.  

Graph for Did solar and wind wilt in the heat?

Source: AEMO data, analysis by Infigen Energy

Now it helps to put this all in perspective by also providing a chart of the output of the Loy Yang A generating unit that went missing during the heatwave. Unit 3 went from producing close to 450 megawatts to producing nothing within the space of minutes right as demand was soaring. This outage was certainly not forecast in advance. But just like the variation in output you expect from wind and solar AEMO is prepared because while they can't precisely predict a power station breakdown, they know it's a reasonable possibility and so they plan around it. When determining how much supply is required to meet peak demand they incorporate a contingency or reserve to cope for an event where the biggest generating unit is lost, which is what unfortunately happened.

Power output from LoyYang A’s unit 3 on 14 January

Graph for Did solar and wind wilt in the heat?

Source: Ketan Joshi - www.etwasluft.blogspot.com.au/ 

CORRECTION: The initial version of this article incorrectly stated that the chart showing wind power output forecasts versus actuals illustrated the forecast for 24 hours ahead when it was in fact the forecast for an hour ahead.