Modeling of Photovoltaic Cell Based on BP Neural Networks Improved by MEA

Abstract

Photovoltaic cell is the central part of solar photovoltaic power generation system, so how to establish effective and accurate photovoltaic cell model is crucial. Based on the research of PV cell characteristic as well as the mind evolutionary algorithm and neural network, this paper put forward a new kind of BP neural network modeling method based on MEA, and used for photovoltaic battery modeling. In this model, using MEA for neural network parameter optimization to overcome defects of the traditional BP neural network which has slow convergence speed, easy to fall into local optimal solution and other shortcomings, and then improved the modeling accuracy and reliability. The test and simulation results showed that the improved neural network model is high precision, little error, short training time and it is effective.