Photovoltaic Power Forecasting Using Cascade Forward Neural Network Based On Levenberg-Marquardt Algorithm

Abstract

Renewable energy is useful energy collected from renewable resources, and renewable resources are naturally replenished within human time. Photovoltaic (PV) is a generator that converts solar energy into electrical energy. Depending on the weather, photovoltaic output is an occasional output. Therefore, this study will use the Cascade Forward Neural Network (CFNN) method with the Levenberg-Marquardt algorithm to predict photovoltaic power plants one day in advance. The measure of accuracy error from the simulation result in this study is calculated using Mean Square Error (MSE). From the simulation results, it is obtained that The Cascade Forward Neural Network (CFNN) method with the Levenberg-Marquardt could give the better MSE at the learning rate of 0.1 by mean MSE of 0,308% while the learning rate of 0.05 by mean MSE of 0,326% and learning rate of 0.01 by mean MSE of 0,322%. It is also obtained that Cascade Forward Neural Network (CFNN) method also eligible for solving photovoltaic power forecasting problem due to its accuracy and should be eligible for another renewable energy electricity source power forecasting.