Contributed by Matt Wytock, Founder & CEO, Gridmatic
Over the past decade, the U.S. electrical grid has experienced significant transformations, with renewable energy growth and a focus on grid resiliency to mitigate the impact of weather volatility. However, traditional management approaches and outdated systems struggle to handle the complexities, hindering grid optimization and supply-demand balance.
AI is now crucial for addressing these challenges. Relying on traditional methods increases risks, including disruptions, economic inefficiencies, and slower renewable energy adoption.
Storage optimization
Storage has long been lauded for its potential to help balance the grid and smooth the adoption of intermittent clean energy. In ERCOT, the rapid growth and adoption of battery storage is evident in its complete takeover of the grid-stabilizing Regulation Down ancillary market, crowding out gas and coal in just three years.
The graph suggests that storage is approaching saturation for revenue generated from Regulation Down services. The opportunity cost for storage to provide Regulation Down is usually lower than that of a gas or coal power plant, so as storage becomes the marginal resource, Regulation Down prices decrease. This trend extends beyond a single ancillary service, each of which is a fraction of the market size of the overall electricity market.
GO DEEPER: Check out the Factor This! energy storage podcast playlist, including episodes on battery storage, long-duration energy storage, gravity storage, and more. Subscribe wherever you get your podcasts.
To sustain returns, storage operators must focus on alternative sources of monetization, such as Real-Time Energy. However, capturing these opportunities is complicated due to the unpredictability of real-time price spikes and the need to balance battery commitments made during participation in the day-ahead market. Outdated storage management approaches are inadequate, whereas advanced forecasting and AI-based battery optimization, including accurate day-ahead forecasts, have proven effective in capitalizing on revenue opportunities.
To illustrate further, the graphs below depict:
- The Actual revenue of a representative system in ERCOT (operated, as most systems are today, without an AI optimization scheduling it)
- The hypothetical Perfect revenue (the maximum revenue achievable with perfect market knowledge),
- The real-world potential revenue achievable with an AI optimizer. The stacked colors demonstrate the AI optimizer’s capability to capture diverse revenue streams more effectively.
Another challenge for manual storage optimization is the recent accelerated pace of market changes, exemplified by ERCOT’s introduction of the ERCOT Contingency Reserve Service (ECRS), the first new daily procured Ancillary Services to be introduced in over two decades. The implementation details were revealed on May 5, 2023, with a rollout scheduled for June 9, 2023.
Renewable Energy Procurement
With increasing demands from investors and consumers for companies to prioritize ESG goals, the focus on decarbonization is intensifying. Google, Microsoft, and many other corporations are embracing the goal of achieving 24/7 Carbon-free Energy (CFE) by sourcing carbon-free energy for every kW-hour they consume, throughout each hour of the day.
However, the traditional methods of sourcing renewables make 24/7 CFE extremely challenging. To address renewables’ intermittency, companies often need to procure excess wind and solar capacity to ensure hourly matching. Moreover, the availability of real-time data is limited, even with two-thirds of industrial meters being smart meters, resulting in delays of at least one day, and up to 30 days, in accessing consumption information. Similarly, on the generation side, there is often a lag in production data due to the absence of real-time information sharing by market operators, further complicating the achievement of precise load matching.
AI has the ability to utilize advanced forecasting techniques to provide accurate predictions of renewable energy sources like wind and solar, even in the absence of real-time consumption and production data. By leveraging these predictions, AI can effectively assist in assembling a carbon-free energy supply portfolio. Additionally, AI can aid in optimizing the matching of intermittent carbon-free and dispatchable resources on an hourly basis, while also leveraging load flexibility through curtailment or onsite generation.
Furthermore, AI plays a crucial role in optimizing the operation of onsite energy storage systems, ensuring the efficient utilization of stored energy during periods of limited renewable energy generation, thus enabling a continuous supply of carbon-free energy when needed.
Leveraging flexible loads
AI’s forecasting capabilities empower energy consumers to maximize benefits from various peak demand reduction programs, leading to energy savings. Flexible loads can also capitalize on real-time market opportunities, for example, the ability to bid into ERCOT’s Day Ahead Ancillary Services Market. This not only opens up new revenue streams but also supports grid reliability. By optimizing load in strategic ways, participants can significantly cut down on energy-related charges and contribute to sustainability goals.
However, the key challenge lies in accurately predicting the timing of these peak events and striking a balance between load curtailment and business operational needs. With advanced knowledge of the expected timing of peak events and energy prices, end-use customers gain the power to strategically curtail their operations.
By harnessing the potential of AI algorithms and advanced forecasting techniques, energy consumers can optimize their engagement in demand reduction programs and the ancillary services market. AI’s predictive capabilities enable the identification of when and where peak demand is likely to occur, empowering businesses and individuals to effectively manage their energy usage and minimize consumption during these critical periods.
Through the intelligent application of AI, energy consumers can maximize the benefits of demand reduction programs, leading to cost savings, reduced environmental impact, and enhanced operational efficiency.
Relying on traditional approaches hinders grid optimization, increases disruptions, limits revenue generation, and slows down the transition to renewable energy. Grid optimization, renewables procurement, and flexible loads are still just a fraction of the ways in which AI can assist stakeholders in adapting to a rapidly changing landscape. By embracing AI, the energy ecosystem can navigate the challenges, unlock new opportunities, and accelerate the transition towards a sustainable and resilient future.
