Optimum Design of Battery-Assisted Photo-Voltaic Energy System for a Commercial Application

This paper focuses on the optimum design of a battery-assisted photo-voltaic (PV) system by using real world data from commercial users. Specifically, we identify the size of PV panels, the capacity of battery energy storage system (BESS), and the optimum scheduling of BESS charging/discharging, such that the long-term average cost, including both energy cost and system cost, can be minimized. The optimum designs are performed by considering the overall cost of the PV system, which usually accounts for a big amount of initial investment, and the aging effects of batteries and solar panels. Practical considerations, such as calendar aging and cycling aging of batteries, inflation of utility cost, and interest rates for investments required for system construction, are considered in the study. The problem is formulated as a mixed integer non-linear programming (MINLP) problem over a time horizon on the order of years to capture the aging effects, whereas almost all existing works on PV system designs consider much shorter time horizons on the order of days or months. The MINLP is transformed into mixed integer linear programming (MILP) and solved by the branch-and-bound (B&B) algorithm. Applying the newly developed algorithms on real-world data from a commercial user in San Francisco reveals that the system achieves the break-even point at the 62nd month and a 38.5% reduction in utility bills.