Abstract: Deployment of short-duration battery energy storage systems (BESS) has been expanding exponentially and BESS are essential to support the growth of renewable generation from solar and wind farms. BESS are competitive participants in ISO-directed markets and can be operated independently or coupled to a renewable asset. I will discuss daily optimization of BESS dispatch through the lens of stochastic control. Among dispatch objectives that will be addressed are firming of a renewable generator relative to a given dispatch profile; energy arbitrage; peak shaving; and managing battery aging. To efficiently find the dynamic feedback control map I will present a machine-learning algorithm based on Gaussian Process regression and Monte Carlo sampling. Results are illustrated via realistic stochastic models for renewable generation output. 

Bio: Mike Ludkovski is a Professor of Statistics and Applied Probability at University of California Santa Barbara where he co-directs the Center for Financial Mathematics and Actuarial Research. Among his research interests are renewable energy markets, Gaussian process models,  stochastic control, computational finance and mortality analysis. His research has been supported by NSF, DOE, ARPA- E among others and includes over 70 peer-reviewed publications. He has worked in the area of stochastic models for energy finance for 15+ years, including co-editing a Springer volume on “Commodities, Energy and Environmental Finance” and teaching multiple summer mini-courses on this topic. He holds a Ph.D. in Operations Research and Financial Engineering from Princeton University and has held visiting positions at London School of Economics and Paris Dauphine University.