The Short-term Load Forecasting Of Electric Power System Based On A New Algorithm

Based on thorough research to the characteristic of the short-term load forecasting of electric power system and the existing forecasting methods,a short-term load forecasting model was proposed, which adopted a new algorithm designed and developed in this paper. Firstly,the SGOA manifesting a thought of the adequate fusion was designed. Then a new Feed Forward Neural Network based on the Early-restart Changed-structure Algorithm was proposed after dynamic optimizing to the GaussNewton Feed Forward Neural Network. When determineing the number of layer of network and the neuron of each layer,the influence of factors such as climate, temperature were considered and carried on fuzzy processing before taking them as the inputs of the network. Finally the SGOA algorithm is used to optimize and train the neural network,then is applied in the short-term load forecasting of electric power system.The results indicated that the improved neural network proposed in this paper sufficiently displayed the superiority of the SGOA and the Early-restart Changed-structure Feed Forward Neural Network in processing non-linear problem,and effectively enhanced the training efficiency of and forecast precision of neural network.