A monster project that relies on Australian company RayGen’s thermal storage technology has attracted serious backing.

The way Will Mosley sees it, there is an excess of unreliable generation in the energy market (including ageing coal assets) and a substantial deficit of reliable generation. “On the hottest day of the year when prices are spiking, can you rely on the ageing existing coal fleet to be there?” says the head of business development at Australian solar thermal storage company RayGen.

The solution, of course, is to save some of that unreliable energy – solar and wind, not coal – and use it later.

To that end, RayGen is working with Amsterdam-headquartered Photon Energy to develop a 300MW/3.6GWh storage project with 150MW of grid-connection capacity on 1,200 hectares in South Australia – what it says is bigger than the biggest energy storage project of any type worldwide (not counting pumped hydro).

The numbers take some explaining: the 300MW is what can be collected by RayGen’s 40%-efficient multi-junction PV receivers and the 3.6GWh of electricity will be produced by turbines turned by ammonia heated by the thermal differential between large, covered reservoirs.

“If you need more storage you simply expand the size of the hole and add more water,” Mosley says.

Similar but much larger pits are used in the European district energy system, Mosley says, especially in Denmark, where pits are heated during the summer by large solar arrays to be used in towns during the winter.

Hot and cold

Reservoirs for hot and cold water have been dug at the 50MWh Carwarp trial project.

The inverted pyramid design allows heat energy to be stratified, he says, so that the density of water separates the charge and discharge temperatures. The water in a RayGen reservoir is hotter – about 90°C – near the top, where the “charge” comes in. Down the bottom, the “discharge” temperature is about 20°C lower. The level between the charge and discharge layers is the “thermocline”.

To charge, the operators draw water from the base of the pit, heat it using concentrated solar energy and add it to the top of the pit. This causes the thermocline to descend.

To discharge, they draw water from the top, use it to heat ammonia to create steam to drive an organic Rankine cycle engine, and then send that water back into the base of the pit. This causes the thermocline to rise.

“You can store [the heated water] for six months and lose 5% of the energy, as demonstrated in Denmark,” Mosley says. “Then, when you need the power, you take the water heated to 90°C, use it to boil ammonia to spin a turbine to generate electricity, and then condense the ammonia with the cold storage [from a reservoir kept close to freezing].”

Once the hot water has transferred the energy needed to boil the ammonia, it returns to the base of the pit at 70°C. The pit is insulated using a rubber membrane made in Australia by Fabtech.

To trial this solution for the first time, RayGen is building a 4MW/50MWh solar energy-plus-storage plant in Carwarp, Victoria, due for completion in mid-2022.

Towers of power

Three of RayGen’s PV Ultra receivers have proven themselves at the Newbridge site.

At the 1,200-hectare Photon project, Mosley expects there to be six cold pits and six hot pits, with some probably being utilised more than others. Each pair of pits will be matched to a turbine. The 1,200 hectares reserved for the project is “almost entirely for solar,” he says. “The storage component is very small, typically less than 5%.” Each megawatt of solar takes up about 2.5 to 3 hectares of land, roughly the same as a solar plant with single-axis tracking.

The core of the technology is RayGen’s receiver, where more than 400 PV Ultra modules, each with 100 triple-junction cells (which are about 40% efficient), are arranged within about four square metres. Each receiver can generate about 1MW of electricity and 2MW of heat. About a third of the sunlight that hits the receiver is converted to electricity. Behind the solar cells a heat sink absorbs the energy sent from a field of heliostats below. “It’s a beam strong enough to melt steel,” Mosley says. “We need to extract heat away from those cells [using a water-based coolant].”

Deep thinker

Multi-junction cells harvest a wider section of the spectrum than regular PV wafers by stacking layers that capture “more colours of the rainbow”. RayGen’s tech is triple-junction (for now), meaning it has three layers.

The heliostats – parabolic mirrors – are kept focused on the receiver throughout the day using software developed in-house. “There is a lot of intellectual property in how we’ve managed to make the mirrors point at the receiver,” Mosley says, where optical imaging allows the heliostats to understand how they each need to move so that the sun’s rays are sent to the right receiver. “We always know exactly where the beam is going.”

A commercial demonstration project has been running in Newbridge, Victoria, since 2015, with two more towers added in 2018 that have been performing at 98% availability. “This a really robust technology; we understand the technology really well now.”

These sophisticated autonomous devices, which communicate in wireless unison, must withstand whatever the atmosphere wants to throw at them, however. One windstorm that tore through the Newbridge site was so severe it pulled the roofs off neighbouring properties, and yet only one of more than 200 heliostats was shifted in its foundations. Although that unit now leans on an edge, Mosley says, the algorithm corrected itself the following day so that it still sends reflected sunlight to the receiver. “That was a demonstration of the ability of our algorithm to compensate for changes.”

Tame the sun

RayGen’s multi-junction PV module is about 40% efficient.

Solar energy is surging into the NEM, from almost nothing 10 years ago to around 11% off all supply this year, rooftop and utility-scale combined. In states where solar dominates, however, this means it is a price-taker. When prices fall to near-nothing or veer into negative territory around midday, owners of solar plants in states such as South Australia feel the pain.

Storage deployed at enormous scale will be needed to store that solar for use in the evening, benefiting the entire grid. But Mosley says owners of storage assets can benefit two ways. “When pricing is negative we become a customer to the grid; we get paid to charge,” he says. “We absorb energy from the market but we also absorb energy from local renewable plants that are constrained. Typically you have two sources of value: the market generally, and the [local] node … that is saturated. There is a lot of zero-value power – and we can take advantage of both of those things.”

In RayGen’s case, electricity from the grid or from its PV receivers is used to chill the cold pits close to freezing. “You need to match the heat you are generating with the cold, which is a sink for the ammonia turbine.” If there is solar in even greater abundance, it can be purchased to run a heat pump to charge the hot pits.

In need of balance

Negative and low prices are clear signals that the market does not need more unreliable generation. Wind and solar developers should include storage in their plans, he says, and it’s safe to assume some coal assets will retire ahead of schedule. “That’s what negative pricing suggests.”

Retailer AGL has signed on as off-taker to the Carwarp project and is financing $5 million in construction costs. It is also looking into the feasibility of using RayGen’s technology for storage at the Liddell coal power station in NSW, which is due to retire its first unit in April 2022. “We are working with our partners to understand the best model for uptake,” Mosley says. Photon Energy, an investor in RayGen, is relying on the merchant market for sales from its Leeton and Fiveborough solar plants in NSW.

A recent $55 million equity raise included Photon, AGL Energy, Schlumberger New Energy, Chevron Technology Ventures and Equinor Ventures.

The long and short of it

With not enough reliable generation in the NEM, the market for “a hedging product” will increase, Mosley tells EcoGeneration. Take a look at pumped-hydro projects, he says. They are not being valued on their day-to-day power supply but for their potential to supply for hours at a time when the price is spiking. Batteries, on the other hand, are responding to shallow ancillary services markets in short, sharp discharges.

“Unless you can do a long duration of storage, you can’t offer capacity,” Mosley says. “The price is going to spike at 4pm on a hot day in summer and it’s not going to finish at 6pm, it’s going to keep spiking through the evening while all the air-conditioners are on.”

The NEM requires much, much more wind and solar to get clean energy making up anywhere near 80% of supply. At the same time, the demand for electricity will increase as industry and transport are electrified. If storage is built close to demand, that would only mean wasteful overspending on transmission linking solar and wind resources with our cities, many hundreds of kilometres away.

“A 30-40% capacity factor on a transmission line from a solar renewable energy zone to a city is not sufficient to justify that new build,” he says.

Sure, storage at the customer end makes sense but enormous amounts of storage at points of supply will smooth generation and make sure transmission lines are fully utilised. “We see a huge demand for long-duration storage that bulk shifts renewable energy from when it isn’t needed to when it is,” he says. “Realistically you can’t do that with pumped hydro because it isn’t near the solar or wind.”