Optimized Battery Energy Storage Systems in the ERCOT Electricity Grid

Abstract

This thesis develops a large-scale DC optimal power flow (DC-OPF) framework to evaluate reliability in an ERCOT-like transmission network using the ACTIVSg2000 synthetic grid. Base-case operation as well as transmission and generator contingencies are simulated using time-series load and renewable generation data. Reliability is quantified using non-served energy (NSE), enabling identification of locations where demand cannot be met under stressed conditions. The analysis shows that transmission failures dominate reliability outcomes, producing substantial variation in NSE, while generator outages have minimal impact due to sufficient reservecapacity. Spatial patterns of NSE reveal persistent bottlenecks concentrated in the Dallas–Fort Worth region. Further, correlation analysis demonstrates that many high-impact contingencies represent redundant system states, allowing reduction to a smaller set of distinct failure modes. A two-stage stochastic optimization model is used to determine optimal battery energy storage placement. Results show that storage is highly concentrated at a small number of locations and remains robust across scenario formulations. Optimal placement is driven by structural transmission bottlenecks rather than scenario-specific assumptions, with storage providing the greatest reliability benefit when located at transmission-constrained load centers. These findings suggest that battery energy storage should be viewed as a tool for alleviating transmission constraints rather than solely for managing renewable variability. The methodology developed in this work provides a scalable approach for identifying high-value storage locations and supports more effective planning of reliable, low-carbon power systems.