A California start-up backed by former California governor and actor Arnold Schwarzenegger has launched what it says is the first strategic bidding software backed by artificial intelligence in the world that can help wind and solar farms dodge negative prices and also deliver a significant lift in revenue for battery storage.
The firm, AMS, says it has some 2,000MW of wind and solar farms in Australia, and one soon-tobe-commissioned battery storage unit under contract, and has already been demonstrating its abilities at the 100MW Snowtown 1 wind farm in South Australia.
On July 10, for instance, there were more than 20 negative pricing events – mostly caused by a network constraint that stopped excess renewables being exported to Victoria. The AI system developed by AMS helped it dodge those events – when it might normally have been exposed to significant payments – and maximise revenue.
CEO and founder Susan Kennedy says the growing share of wind and solar, the introduction of battery storage, and the appearance of so-called solar “duck curves” and volatile trading means that the excel spread sheets that once served the coal and gas industry can no longer keep up.
Kennedy, a former chief of staff to Schwarzenegger, whose pursuit of climate goals and renewable energy resulted in some 16GW of new wind and solar capacity in that state, and help set off a renewables revolution in other countries, says Australia has now taken a leadership position in the world.
“What is happening here will be replicated around the world …. and you need to cost-effectively manage the integration of renewables,” Kennedy tells RenewEconomy in an interview from the company’s base in California.
“The requires batteries and other storage, but batteries cannot be optimised with Excel spreadsheets. Whether it is solar, wind, pumped hydro, batteries or gas, all these assets have to be transacted in markets that are very very fast and very volatile.”
Kennedy says the AMS software can increase revenue from a wind or solar farm by up to 10 per cent. Staggeringly, Kennedy says it will lift the revenue from a battery three-to-four fold. That’s because a human trader simply can’t keep up with the thousands of possibilities that might appear in any five minute trading period.
That will be of interest to the many developers and operators looking to add batteries into the grid. The Tesla big battery, for instance, has already impressed with the speed, accuracy and flexibility of its operations, and it has already been joined by three other big batteries into the grid, with another four under construction or being commissioned, and a huge number of projects in the pipeline.
Being able to dial down output instead of paying someone to take the output will save wind and solar farms potentially millions of dollars. Some solar farms, such as Tailem Bend, have already had to switch off during negative pricing periods, but those decisions do not have the flexibility of the trading exampled in the graph above.
Deion Campbell, the CEO of Tilt Renewables which owns the 100MW Snowtown 1 wind farm, says he was impressed with the results right from the outset as the technology “self-executed” a “precision trading” strategy through the wave of volatility and negative prices in South Australia on that July day.
“AI software gives us an edge to manage risk and improve margins in energy markets that are becoming more challenging with high penetrations of renewable generation,” Campbell says. “Being an early adopter of innovation like this demonstrates Tilt Renewables’ leading approach to renewable asset management and we look forward to applying this technology to other assets shortly.”
Tilt, as RenewEconomy reported earlier this month, is looking at making a final decision on its proposed battery that will accompany then Snowtown 1 wind farm, although this is not the battery unit that has already been contracted to AMS.
Schwarzenegger, who starred as the robot in the Terminator series, said in a statement: “This time, artificial intelligence is being used to save the planet. If we’re going to address climate change, we need to build renewable energy on a massive scale. Electric grids around the world can’t handle renewable energy on a massive scale without AI.”
Kennedy (pictured above Schwarzenegger) with founded AMS (then Advanced Microgrid Systems) in 2014, and the company has been mostly known for operating “hybrid-electric buildings” that delivered more than 2,000 MWh of flexible capacity for one of California’s largest utilities as part of a 70MW/360MWh “virtual power plant”. Its early backers also included Macquarie Capital.
In 2018, AMS expanded to Australia where high penetrations of solar and wind have caused big swings in wholesale energy prices, and where the world’s biggest lithium-ion battery – the 100MW/129MWh Tesla big battery at Hornsdale in South Australia – was installed.
Kennedy liken the shift from spreadsheets to AI software to that of the move from analogue to digital.
“Every region of the globe is experiencing the challenges of integrating solar, wind, electric vehicles and distributed energy into electric grids designed for coal, oil and gas,” Kennedy says.
“As renewable penetration grows, market volatility will increase. Assets need to be smart, fast and able to accurately predict the forces impacting the electric grid ….batteries cannot be optimised with excel spreadsheets.”