PERFORMANCE EVALUATION AND CLASSIFICATION OF SOLAR CELLS EFFICIENCY ACCORDING TO AREA USING NEURAL NETWORKS رسالة ماجستير

Abstract

The prediction of the output power of PV power system in a given place has always been an important factor in planning the installation of PV power system, and guiding electrical companies to control, manage and distribute the energy to their electricity networks properly. The production of the electricity sector in Palestine using PV power system is a promising sector; the aim of this research is to propose two models that are used to predict the future output power values of solar cells, which provides individuals and companies with predicted future information, and thus they can organize their activities in the electricity sector. Firstly; we aim creating a model capable of connecting time, place, and the PV power system output relations between randomly distributed PV power systems. The proposed model analyzes collected data from units through solar cells distributed in different places in Palestine. Multi-layer Feed-Forward with Back propagation Neural Networks (MFFNNBP) is used to predict the power output of the solar cells in different places in Palestine. The model depends on predicting the future production of the power output of PV power system depending on the real power output of the previous values. The data used in this thesis depends on data collection of one day, month, and year. Secondly; we propose an Enhanced Radial Basis Function Neural Networks (RBFNN) model that depends on the standard RBF existing in Matlab (newrb). This enhancement on newrb depends on the use of intelligent algorithms like K-means Clustering, K-Nearest Neighbor (K-NN,) and Singular Value Decomposition (SVD), to optimize the centers c, radii r, and weights w of the RBFNN, which replace the mathematical calculation used to find these parameters in newrb. This enhanced model is applied to predict the solar cells energy production in Palestine using already installed PV power system in Jericho. Solar irradiance and daily temperature used as an input training data set for the proposed model, v with the real output power of (2015) as the training supervisor. The model is applied to predict the output power within one month and one year. Finally, a traditional power output equation was optimized to calculate the solar energy depending on the daily irradiance and temperature with an acceptable accuracy. The experimental results show that the enhanced model performs more precisely in prediction than the Multilayer Perceptron Neural Network (MLPNN) algorithm, with low Mean Square Error (MSE) of relatively few neurons on the hidden layer (RBF). The proposed models conduct a systematic process, which predicts the power output of PV power system in selected area. The enhanced RBF model is more precise that MLP and traditional RBF models with low RMSE with relatively few neurons in the hidden layer, and so we can determine the suitable place for an installation of solar cell panel in different areas of Palestine