The Future of AI in Power Grids is HereBy | February 28, 2020
Recently I spoke with Wall Street Journal reporter Asa Fitch for his article “The Key to Keeping the Lights On: Artificial Intelligence.” From the top he noted that Florida Power & Light (FPL) technology investments are paying off especially with the effects of Hurricane Wilma (2005) and of Hurricane Irma (2017). Both category five hurricanes brought havoc and destruction, knocking out power to millions of Floridians. While FPL took 18 days in 2005 to restore power, 12 years later FPL cut that time to restore power by nearly 50%, with residents suffering through an outage for only 10 days.
Asa is exactly right that technology investments are also paying off in our increasingly “complex electricity grid” world. Our own research from 400+ commercial and industrial customers who have installed battery storage with Athena™ AI, Stem’s artificial intelligence (AI)-powered analytics platform, has found that our customers’ investment in energy storage is also paying off. Athena is at the heart of helping forward thinking communities, commercial enterprises, solar project developers and utility providers “keep the lights on” by intelligently managing energy storage solutions for daily and emergency back-up use, while simultaneously enabling organizations to participate in markets for a stable grid while providing capacity to utilities. Stem is highly differentiated: Athena’s AI algorithms learn everything about the customer’s power consumption and grid conditions, then determines when to use battery power in order to maximize the benefits to all parties.
When paired with distributed solar projects, Stem’s energy storage is increasing the financial returns for renewable developers like Syncharpa and Kearsarge. Athena forecasts solar production and wholesale market conditions, optimizing the energy storage system to operate in a manner that maximizes revenue for the developer, while also benefiting the utility via grid services. These grid services, such as capacity, frequency response, and operating reserves, help avoid rolling blackouts during times of high demand or grid disruption.
For utilities like Southern California Edison (SCE), Stem dispatches a fleet of distributed energy storage systems to operate like a virtual power plant. Athena manages each battery to ensure benefits are realized for the hundreds of host customers while at the same time operating sites to respond in aggregate when called upon by SCE. Dispatching hundreds of batteries in aggregate has the same effect as firing up a ‘traditional’ generator like a gas peaker plant but with fewer greenhouse gas emissions.
Another customer, chemical manufacturer Solvay, located in Ontario, Canada, is an example of how commercial and industrial enterprises are benefiting from Stem’s AI-powered energy storage. Stem’s platform predicts what will be the five highest peak hours of the year in Ontario, which contributes to 50-70% of a commercial and industrial (C&I) customer’s annual utility bill. During these forecasted high peak hours, Stem dispatches all our systems in Ontario, thereby reducing Solvay’s customers’ load and ensuring cost avoidance during the highest energy peak points. The grid operator also benefits from this load reduction, because the operator avoids the need for expensive, fossil fuel-powered peaker plants.
Today you hear constant conversation about a decentralized, decarbonized and digitized grid. In 2010, most electricity generation came from gas, coal or nuclear power. In 2020, power flows one way from the utility to the customer but over the next few years, power will flow both directions. Today there are a number of new services such as energy management and efficiencies, back-up power and market participation. By 2030, even more renewables will be online, wind, solar and storage will be ubiquitous and much less expensive than today. Energy storage will orchestrate these technologies for customers that will include C&I, residental, utility and grid operators.
The clean, resilient grid of the future will be one where electricity consumers and producers of all sizes use intelligent software to dynamically interact with the utility. It will be less top-down, centrally managed and operated as interlocking distributed energy markets. AI-enabled energy storage will be critical to this future because it gives everyone high latency, real-time control over how they consume, produce or transmit electricity. AI is needed to optimize the value of an energy storage project or networked fleet of projects in a virtual power plant, as these systems play in more rapid and complex energy markets.
The future of power is bright indeed and AI and energy storage will be at the heart of it. Stay tuned.