‘Reliable’ renewables: What cost battery storage and structural inflation?

Renewables power is intermittent. Intermittency is variable. Variation is uncertain. Intermittency is also seasonal. Reliable electricity, 24/7 all year, means renewables' capacity must allow for these realities. There are many possible outcomes each year. Consider one solar illustration.

Start from my Georgia on my mind opinion piece (OLO, 22/05/24). Now suppose solar power is 15% full-on one day, but all-off the next due to heavy cloud. Averaged over the two days, solar power is generated 7.5% of the time. For reliability, generation capacity must allow 48 hours of power to be produced in 7.5% of the time. That's 13.3 times the capacity of always-on base-load power. Battery storage must be 12.8 times the generation capacity of base-load power.

Assume this one day on/one day off cycle applies on average over solar panel lives. Over the 80-year life cycle for conventional nuclear power plants, with 30 years life for solar panels, and 10-20 years for batteries, installed solar generation capacity must be at least 40 times base-load capacity, and, for batteries, about 51-102 times base-load, both measured over 80 years.

Intermittency is more variable and uncertain, day to day, than this example. For longer sunless periods, renewables generation and battery storage capacity must be larger, compared with always-on base-load power, to deliver the same reliability, especially measured over 80 years.

Seasonality is somewhat more predictable each year. But it has larger effects on solar capacity, especially battery storage, required for reliability, compared with base-load power.

Consider this conservative example for solar power. Assume 'full-on' power averages 16% per day in the warmer half of the year, and 14% per day in the cooler half. Of the 16% in the warmer half year, 1 percentage point must be stored in batteries (assuming a 100% renewables-only plus 100% batteries scenario) for discharge during the cooler half of the year. The year average 'full on' solar power is still the evidence-based 15%.

For solar generation, harvesting the 'summer surplus' requires 6.7% extra generation capacity per day compared with baseload power. All of that extra must be stored in batteries during the warmer half year, to be fully discharged in the cooler half.

This extra daily generation production must be stored, on average, for 182.5 days. This daily 'summer surplus' is an energy flow. On average it must accumulate a dischargeable energy stock in batteries for 182.5 days. That stock peaks at (0.067 x 182.5 = 12.2) times the average daily 'summer surplus' extra generation. It is then fully discharged over the cooler half year.

Over 80 years (the life of a nuclear base-load power plant), the average daily extra 'summer surplus' solar generation must be 20% more than otherwise, assuming a 30-year PV life and reinvesting three times over 80 years. The extra 'summer surplus' battery capacity must be 48.8 – 97.6 times the 'summer surplus' daily solar generation, assuming batteries last for 5 – 10 years and reinvesting in new ones 4 – 8 times over 80 years.

What would that mean for the National Electricity Market (the NEM) if reliability is to be maintained year-round, all seasons? Assume a fossil fuel-free NEM using all-solar generation, and all-battery seasonal storage.

Last financial year, the average daily power produced by the NEM was over 520,000mwh. Suppose 'summer surplus' generation is added. That's 6.7% more. That is an extra 34,667mwh per day. That extra, on average, must be stored every day for 182.5 days. That's a storage total of 6,326,660mwh. If batteries last 20 years, over 80 years, battery capacity investment must deliver 25,306,661mwh. If batteries last ten years, storage capacity doubles to 50,613,283mwh.

The Tesla Hornsdale 'big battery' in South Australia is claimed to be able to discharge 194mwh from fully charged to zero. For NEM storage, capacity of 130,447 – 260,893 Hornsdales would be needed just for seasonal storage and discharge (ignoring efficiency losses).

That's expensive. If, as claimed, a Hornsdale 'big battery' costs $90 million, and batteries last 20 years, the cost is (Australian dollars) $11,740,194 million (A$11.7 trillion). If the batteries only last 10 years, the cost is A$23,480,389 million (A$23.5 trillion).

Technology improvements and scale operations might well cut unit energy storage costs a lot. Suppose such costs average just 10% of the cost for the Hornsdale SA 'big battery'. At between A$1.2 trillion and A$2.3 trillion, seasonal battery storage is still very costly.

Before allowing for the extra costs of needed solar generation capacity, and new transmission capacity everywhere, such battery costs for the NEM under a 100% solar plus 100% batteries regime are ruinous. Either seasonal reliability is sacrificed, or customer costs soar even more.

Or both.

I've not allowed for other incipient electricity demands, such as those driven by AI, cryptocurrencies, electric vehicles, 'green' production of various 'green' metals, electrolysed hydrogen, and the like. These would magnify renewables power supply problems.

I'm sceptical about 'distributed' renewables generation and transmission, and so-called 'demand response' (really a euphemism for power rationing). Who pays for all that, anyway?

Switching to a 'reliable' renewables-only policy for the NEM is a major structural inflation driver with decades still to run. Australian living standards, inevitably, will fall more.

The economic realities of intermittent renewables, plus the growing demand for maintaining reliable electric power, will force consideration of alternatives.

Note that, even if unit costs of storing energy in batteries fall to 10% of the Hornsdale 'big battery' cost, the total cost of investing in the required number of Hornsdale-equivalent energy storage units won't happen.

Why? The cost averages about the same as the current dollar value of Australian GDP.