Based On Improved BP Neural Network Model Generating Power Predicting For PV System

In this work the artificial neural networks (ANN) and the optimization algorithm of nonlinear damping least squares (Levenberg-Marquardt) were applied to estimate the generating power of photovoltaic system in China. And the MATLAB was applied to establish prediction model. Finally, the training samples were measured data of 30 days, 90 days and 180 days. Under the three samples, it researched generating output power forecasting of photovoltaic system. The predicted results show LMBP overcomes the shortcomings of BP neural network; it has better convergence and accuracy. After compare with all predicting data, predicting results of 30 days data is the most accurate among three training samples.