Short Term Forecasting of Global Solar Irradiance by k-Nearest Neighbor Multilayer Backpropagation Learning Neural Network Algorithm

The paper proposes based on meteorology data, especially for optimizing the operation of power generating electricity from photovoltaic energy. This paper proposes a novel methodology for very short term forecasting of hourly global solar irradiance (GSI). This methodology is a combination of k-nearest neighbor algorithm (k-NN) modelling and multilayer backpropagation learning neural network (BPLNN) model. The k-NN-multilayer BPLNN model is designed to forecast GSI for 1 hours ahead based on meteorology data for the target PV station which position is surrounded by eight other adjacent PV stations. The forecasting for global solar irradiance using k-NN-multilayer BPLNN modelling is a very powerful technique to determine the behaviour of time series data. The first method implements k-NN as a preprocessing technique prior to backpropagation learning method. The error statistical indicators of k-NN- multilayer BPLNN models used momentum (mc) = 0.8 the root-mean-square error (RMSE) is 176.5 W/m2. The models forecasts are then compared to measured data and validation results indicate that the k-NN-BPLNN based method presented in this study can estimate hourly GSI with satisfactory accuracy.