Energy Storage State of Health Optimization for Variable Generation Power Systems

The rapid growth of residential renewable energy in the United States has driven consumers to purchase energy storage systems to counter the unpredictable variation of solar power and wind generation. To counter intermittency problems from weather-related events such as cloud cover that occur with residential photovoltaic generation, intelligent power management interfaces have been designed to optimize the performance of the consumer's renewable energy system while minimizing the cost to the consumer. This approach determines a state of charge schedule for the battery based on a dynamic programming algorithm that minimizes consumer cost and incorporates the lifetime of the most expensive system component, the energy storage system. The battery state of health was introduced into the model as an ageing coefficient that forces conservative battery behavior to preserve its lifetime with continued use. This thesis provides an adapted predictive dynamic programming algorithm that optimizes a battery charge/discharge cycle for the model of an existing 24 V single-phase, grid-connected renewable energy system located in Drexel University's Center for Electric Power Engineering (CEPE). Simulations of the system model analyzed four solar irradiance datasets and corresponding load profiles to indicate changes in the percentage of photovoltaic generation used to charge the battery bank over the course of a year. This study details the approach to integrate the adapted predictive dynamic programming algorithm to the physical hardware of the renewable energy system located in the CEPE.M.S., Electrical Engineering -- Drexel University, 201