SOLAR CELL MODELLING USING RADIAL BASIS FUNCTIONS NEURAL NETWORK

摘要

This paper presents solar cell modelling using single diode equivalent circuit as base line to train radial basis functions neural network (RBFNN). The simulation was run using a random set of data (solar radiation and ambient temperature) for the purpose to study the characteristics of solar cells at diffident situations. Afterwards, RBFNN has been used to simulate the whole characteristics of solar cell in general and specifically at maximum power points (MPP). Finally the RBFNN has been used to simulate a real data collected from Ramah site (East of Riyadh, Saudi Arabia). Two sets of data have been prepared to train the RBFNN. It has been shown that RBFNN can predict the MPP with minimal least mean square error (LMSE).