I-Corps: Translation potential of an Artificial Intelligence-driven electricity market optimization platform

Award ID: 2628338

This I-Corps project is based on the development of a mathematical framework designed to optimize revenue and manage risk for utility scale energy storage systems participating in wholesale electricity markets. This technology addresses inefficiencies in energy market operations. The current electricity market faces three major challenges: 1) uncertainty in energy generation and tax credits, which drives many bids to zero or even negative prices in wholesale markets; 2) inefficient operation of most energy storage systems in response to price volatility; and 3) extreme price swings, with Day-Ahead prices often diverging significantly from Real-Time prices. These factors reduce the profitability of energy storage. The approach provides an optimized market guidance plan that may help storage operators navigate price volatility, select profitable bidding strategies, and understand risk exposure across multiple market products. This I-Corps project utilizes experiential learning coupled with a first-hand investigation of the industry ecosystem to assess advances in an integrated mathematical optimization framework that may be incorporated into real world decision workflows for energy storage operators. Using advanced machine learning, the platform provides precise forecasts and enhanced optimization for wholesale electricity market prices, plant generation, and electricity demand, enabling more profitable energy storage scheduling, physical trading, and virtual trading. The platform also offers an all-in-one service that combines unified databases, forecasting, optimization, and market intelligence tools that analyze grid data to recommend hedging strategies for trade risk management, all at nearly one-tenth the cost of competitors. The database structure encompasses standardized weather data (including temperature forecasting, wind speed forecasting, wind direction forecasting, sky cover forecasting, and humidity forecasting), electricity grid data (such as grid-level forecasting), and fuel price data. The forecasting service utilizes AI-driven models to generate highly accurate predictions for power plant production and electricity market prices. The optimization feature generates optimal energy storage schedules based on the predictions from the large machine learning models. In addition, a market intelligence tool scans grid-wide data to produce suggested hedge pairs and ratios for electricity trade risk management. The technology has the potential to enhance decision-making and profitability for energy stakeholders. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.

NSF Program Director: Ruth Shuman