Load Forecasting based photovoltaic power using New Particle Swarm Neural Networks Model

摘要

The load forecasting is required in power system management and ensures electricity providing for customers. Photovoltaic power forecasting aims to reduce the fuel consumption and play important role in the supervisory control for a hybrid energy system. This paper presents the application of new model using neural networks (NN) and Particle Swarm Optimization (PSO) to determine the net load forecasting. In this study, instead of the method of back-propagation of the gradient, the optimization technique by swarms of particles is well tested for training neural network that optimizes the forecast error. Simulations were run and the results are discussed showing that New Hybrid Technique (PSO-NN) is capable to decrease the forecasting error.