Predicting Renewable Energy Based On Optimized Neural Network

With the energy crisis and environmental pollution problem become more and more serious.Renewable energy becomes the best alternative due to its cleanness and high efficiency,and the study of renewable energy at home and abroad goes more and more in-depth.Among them,the technique of photovoltaic(PV)power generation and wind power generation are the most mature.However,the output of photovoltaic power and wind power are influenced by many environmental factors,it has the intrinsic intermittence and fluctuation,the integration into the grid will lead to the power fluctuations.Therefore,increasing the forecast accuracy benefit the grid schedule,reduce the bad effects on power grid,achieve the full utilization of renewable sources and avoid the waste of energy.On the basis of evolutionary algorithms and artificial neural network,the paper analyzes the influence of the improved neural network on renewable energy,we further study the prediction ways and gain some achievements.Firstly,the prediction of wind speed always contains many inputs,the dimension of variables is large,the paper proposes to predict renewable energy using a cooperative-coevolution(CC)genetic algorithm based on back propagation(GABP)neural network.The way proposed in this paper use each neuron in hidden layer as a reference.Experiments found that CC-GABP model improves the accuracy of the prediction of wind speed.Then,through the analysis of photovoltaic power prediction,a simple yet effective method for PV power prediction is proposed.This method improves the relevancy of train samples and prediction samples.And the simulation results of PV power generation show that the proposed method achieves higher prediction precision.Last but not the least,some PV power prediction also involve a large number of input variables,this type of PV power is predicted by CC-GABP.The contrast experiments show that CC-GABP has good fitting precision and predicted precision,further prove the potential of CC-GABP.