Australia is transitioning from fossil fuels to renewables. Are the latter reliable? What generation and storage capacity, at what cost, is needed for renewables providing all power? We seem to be heading there.
Politicians duck these questions as hard as they push renewables. They hide behind using remaining fossil-fuelled power, 'demand response' (rationing), other selective outages, and crossed fingers, as back-up.
As renewables rise due to ambitious state renewable energy targets (RETs), reliability imperatives will force these crucial questions to the fore. What happens as we get to 100% renewables?
Consider two solar renewables costs not shared with fossil-fuels, one of which is avoidable – at a high price. The maximum generation capacity (call this TC) of solar panels (PVs) is the sum of:
- Power generated and used ('used power' or UP), plus
- Power generated but not used ('unused power' or UUP – sometimes called 'curtailed supply'), plus
- Power not generated at all ('no power', or NP).
In Australia today, UP, UUP and NP all have positive values adding up to TC.
On very sunny days, solar panels generate much power, but often it's not used to power appliances, charge batteries (if they're fully charged) or fed into the grid. Grid feed-in is stopped to protect it (eg, because voltage from excess supply increases above required levels). That power is 'unused', or 'wasted'.
So PVs generate power we use, plus some power we don't – or can't. Used and unused power are a job lot produced by the same PVs. We have to cop the latter to get the former.
Very windy and sunny SA is a good example. Last financial year, AEMO reported used solar power was about 15% of total PV capacity in SA. The remaining 85%, was 'unused power' plus 'no power'. AEMO records used power (UP) and total capacity (TC). 'Unused power' and 'no power' aren't shown separately.
Suppose we want solar to deliver the same power as the fossil fuel plant it's replacing. Call this number 100. What solar generation capacity is needed, with a 15% 'efficiency rating', to deliver power of 100? We need about 6.7 times the generation capacity of the fossil-fuel plant. That's 100 divided by 15. With 15% 'efficiency', solar capacity must be sufficiently larger to generate enough power to match the continuous power generation of fossil fuels (in this case a value of 100). We need 6.7 times more generation grunt to do the same power supply job. This is solar intermittency in action.
There's more: 15%-of-capacity power generation, 24/7, can't happen. We'll see 100% of capacity for 3.6 hours (15%) of each day, on average, and zero for the other 20.4 hours. The additional generation capacity required is the same either way. The 3.6 hours scenario also needs batteries to cover the other 20.4 hours.
Of this solar power set at 100, 15% can be used as generated. The other 85% must be stored for use when the sun doesn't shine. We need up to another 5.7 times the generation capacity equivalent of the fossil-fuel plant being replaced to store the extra 85% of solar power. That's 85 divided by 15. Intermittency again.
So, based on 2016-17 AEMO results for SA, replacing 'on tap' fossil-fuel power with PVs delivering the same power requires up to 12.3 times the generation-equivalent capacity in solar panels and batteries.
There's more: 'unused power' (UUP).
The 85% residual from the 2016-17 AEMO SA data includes 'unused power' (ie, 'wasted' power) plus 'no power'. Does it matter if we don't know how much of each was in the 85% result? In one sense, no. AEMO provides an overall 'used power' rating (15%) for total solar capacity. Only used power counts – for users.
Knowing the value of 'unused power' does matter for power costs. It is a measure of the extent to which we need to 'over-invest' in renewables capacity to deliver a given value for UP plus UUP. If we don't 'over-invest', we won't have enough capacity to deliver a target value for used power. We must generate both.
'Unused power' also measures how much additional storage we need to convert it into used power. We could reduce or eliminate unused power if we had more batteries in which to store it. If we could store presently unused power, the measured 'efficiency' of solar panels would rise. If, say, unused power was 10% of used power, and we could store all of it for later use, then AEMO's 2016-17 SA 'efficiency rating' for solar would increase from 15% to 16.5%.
We'd need fewer solar panels to deliver the same used power (in this case 100). But here's the rub. We'd also need even more batteries, not fewer, to store hitherto unused power. Required PV generation capacity falls to under 6.1 times the fossil-fuel plant it's replacing. The generation-equivalent capacity of needed batteries increases to over 6.3 times the fossil-fuel plant it's replacing. The all-up required generation capacity equivalent rises to about 12.4 times the fossil-fuel plant it's replacing.
We don't know the % of unused power to used power. We can do 'what if' scenarios assuming different proportions of unused power to used power, using the 15% AEMO 2016-17 SA estimate for the latter.
If we want (rising) renewables capacity to deliver the same power as fossil-fuels, while eliminating waste of renewables 'unused power' (by storing it), required capacity and costs rise. A lot. See chart below.
To replace fossil-fuel generators while delivering the same power requires solar generation plus battery storage capacity over 12 times larger if we don't capture unused power. If we do capture unused power at, say, 70% of used power, it's over 14 times. These are big numbers, combining intermittency, capturing unused power, and required battery storage capacity. Unused power will rise as renewables' power share rises. More renewables power will be wasted without proportionally more storage. We'll pay, used or not.
AEMO's 2016-17 efficiency estimate for wind in SA is about 29%. The required wind capacity multiplier to replace the same fossil-fuel power is nearly 6 times (ignoring 'unused power'). This is a big number too.
There's still more. Additional costs using renewables do not end with intermittency and unused power. The intermittency of renewables is uncertain. We don't know for certain when the sun will shine, the wind will blow, or water availability will be drought-affected.
The SA intermittency effects measured by the AEMO (eg, for solar, 15% 'efficiency' in 2016-17) are an 'on average' measure for the financial year as a whole. Through the year, there will be substantial variations around that average (season- and weather-related). Reliability requires sufficient generation and storage capacity to deal with these variations around the average as well as year-average intermittency.
Uncertainty is not about averages. It's about the spread around them. Pluses and minuses around the average don't cancel out. They increase total required renewables capacity for reliability.
We know there will be long periods (eg, during tropical storms, cyclones, etc) when the sun will not shine for days or even weeks, as well as heatwave conditions, again for days or more. Prolonged periods of little or no wind (even too much) are common. Droughts are a cyclical Australian feature.
Long periods of no renewables supply, or too much, threaten power reliability. As RETs head towards 100%, using renewables as back-up requires (i) even more generation capacity (to generate more power in 'good' times), and (ii) even more storage capacity (to hoard more power for 'bad' times). Both add yet more to renewables' costs relative to fossil-fuel generators delivering the same reliable power.
What's the cost of reliability-equivalent renewables versus fossil-fuel plant? We're not told. For now, politicians rely on fossil-fuel, rationing, and hope, as back-up. Their 'power' concerns are more personal.
I'm no renewables fan. I'd like to see all support – Federal and State – abolished ASAP. They're ineffective, inefficient, costly, unreliable, and unfair to the poor (see my OLO Power Failure piece of 30 January 2018).
It's incumbent upon die-hard fans of up to 100% renewables to respond publicly to the multiplied generation and storage capacity arithmetic outlined above. Are these multiples right or wrong? If wrong, why? If right, what will they cost? If they continue to ignore these questions, why? More dissembling and dishonesty?
Most politicians, incumbent governments and other, give lip-service to power affordability and reliability, and to renewables. OK. Tell us what happens in the power end-game where we have 100% renewables. Tell us what 100% reliance on renewables requires in multiplied generation and storage capacity.
And tell us the bill we'll cop for that.
Wither (the spelling is deliberate) affordable, reliable, power?