The announcement of a $650 million solar power plant near the South Australian town of Port Augusta brought forth the usual statements that such a plant would produce reliable or dispatchable power.

But anyone with experience of renewable plants would have looked in vain in the announcements for usable information on the plant’s performance, including an estimate of the plant’s capacity factor, or for an explanation of why the South Australian government of Jay Weatherill is getting power apparently so cheaply from an expensive and still largely unproven way to generate electricity.

The capacity factor may be thought of as a generator’s average output. Although this factor is not a true measure of the plant’s reliability it’s a start and, for the record, a similar although smaller plant built by the same company, SolarReserve, in the Nevada desert in the US has been reported as having a capacity factor of 52 per cent (the reported figures vary, and are not properly sourced).

That means its actual average output is slightly more than half its theoretical total output.

But even that figure is an estimate. At about the time of the August announcement of the Port Augusta plant, the Adelaide Advertiser reported that the 110 megawatt Crescent Dunes plant in Nevada stopped operating in October of last year and only resumed generating power in July, due to a leak in the molten salt storage container. As the facility was still ramping up to its full output over 2016, and has yet to reach the 55 per cent capacity figure, the plant does not have much of a production record on which to base an assessment.

In contrast, the two brown coal plants that used to operate at Port Augusta – the Playford B (240 MW) and Northern (520 MW) – would have had nominal capacity factors of more than 80 per cent. Although they may not have operated 80 per cent of the time, they produced electricity on demand and up to full capacity when required, and for as long as necessary.

Like the Crescent Dunes facility, the new Port Augusta generator will use a host of mirrors scattered around the base of a tower to concentrate sunlight at the top of the tower. The resulting intense heat melts salt, and that molten salt is pumped away to boil water. The resulting steam drives the turbines which generates the electricity. The publicity material repeatedly states that the molten salt will hold heat for up to eight hours after going through the tower, and so produce full power at night. That means the project has characteristics of both an energy store (a battery) and a generator, but only for eight hours after the sun sets.

The publicity material produced for the plant states that the plant in Nevada managed to operate for 120 hours (five days) continuously in July, America’s hottest month. That is fine as far as it goes, but was the plant at full capacity during those days and, while we are on the subject, what happened on the other days? Production profiles for four days where the facility managed periods of continuous 100 megawatt production for certain periods were given, but again observers are left wondering what happened on the other days.

It is typical of the media coverage of this and other projects that consumers are told little more than what’s in the media releases plus a few quotes from the managers or politicians about how the project is the start of the green future.

There are plenty of other solar power stations worldwide but these mostly use water instead of salt and do not store energy. Apart from the Crescent Dunes plant in Nevada, one of the few comparable plants is the Solana generating station in Arizona which uses parabolic troughs rather than a tower and has molten salt equivalent to six hours of energy storage.

Completed in 2013, the plant finished its second full year of production in 2016, but US Energy Information Administration figures indicate that output was 25 per cent below projections.

One often-mentioned plant is the Ivanpah Solar Electric Generating System in the Mojave Desert, which is also infamous for the number of birds killed from flying through the concentrated sunlight. This has a capacity of 392 megawatts, with a planned 27.4 per cent capacity factor, but that facility uses water, rather than molten salt, so there is no effective energy storage and it has to be kick-started in the morning with gas power. The plant uses so much natural gas that, as the Orange County Register reported in 2015, the facility should have been registered as a greenhouse gas emitter under state law.

One feature shared by these plants is that they are expensive to run, but the electricity from the Port Augusta plant will be strangely cheap. Media reports put the wholesale sale price of power from the Nevada installation at $US135 ($A170) a megawatt hour, after a United States Federal government loan guarantee and 30 per cent of the project costs being reimbursed through tax concessions.

The Port Augusta plant gets a $110 million ‘concessional equity loan’ (what this means is not explained) from the federal government, but the South Australian government says it will only pay $A78 for power from the Port Augusta plant, which is a little less than the ruling wholesale price for electricity but well under half the price of power from the similar US plant.

Although the state government is contracting directly to buy the electricity, the power still has to go into a grid. The grid operators then face all the problems and costs of keeping the grid’s voltage and frequency stable 24 hours a day, despite the addition of yet another, substantial, renewable source. This increases costs for everybody.

The Port Augusta project looks good in press releases but many, unexplained aspects of it puzzle the observer, along with its still largely unproven technology. Alas, the project appears to have all the makings of yet another renewable white elephant.