Photovoltaic parameter estimation using heuristic optimization

Abstract

Photovoltaic (PV) panel accurate modeling is essential for operational studies and generation estimation under actual conditions and power electronic control device designs in grid connected or isolated power systems. Inaccuracy in modeling parameters such as photon's current, diode's saturation current, series and parallel resistances and diodes ideality coefficients causes errors in modeling and simulating photovoltaic panel's operation. In this paper, photovoltaic panel's parameters are estimated according to characteristic curves obtained from experiments. Heuristic optimization methods are utilized to calculate parameters for both singlediode and double-diode models. Results indicate that harmony search, binary particle swarm, artificial bee colony, cuckoo search, firefly and bat algorithms yield better estimation, compared to other algorithms. Model validation is done by Newton algorithm which has more accuracy in double-diode model.