The Study On Optimization Criteriaand Fusion Strategy Of Wind Power Forecasting Models

With the aggravation of environmental pollution and energy crisis,the countries around the world began to look for clean,renewable green energy to replace the traditional non-renewable energy.Wind energy has been the attention of all countries due to the wide distribution,pollution-free and high development potential.The main form of using wind energy is wind power generation.With the rapid development of wind power generation,the proportion of wind power in power grid is becoming bigger and bigger.However,the intermittence and randomness of wind energy will affect the power grid and will restrict the development of wind power.Therefore,accurate forecasting of wind power is very important for the development of wind power and the stable operation of power grid.In order to further improve the forecast accuracy of wind power,this paper aims at the key problems in wind power forecasting modeling,the improvement of power forecasting model is studied.The evaluation and optimization methods of the forecasting model are studied emphatically based on the forecasting models established by the research group,as well as the fusion strategy of the forecasting model.The main contents of this paper are as follows:(1)This paper summarizes the background and research status of wind power forecasting and introduces the classification of wind power forecasting model,the classical single forecasting model and combined forecasting model.It summarizes the problems existing in wind power forecast and the significance and current situation of model evaluation.(2)Taking wavelet neural network as an example,the key problems in wind power forecasting are studied.In order to solve the problem that the parameters and structure of the wavelet neural network are difficult to be determined,the particle swarm optimization algorithm is used to optimize the parameters of the wavelet network.The gray correlation deletion method is used to determine the number of hidden layer nodes in the wavelet network.The influence of wind power forecasting accuracy is analyzed in detail factor and its data characteristics,the input variables of the wavelet network are determined.The simulation results show that the forecasting accuracy of the proposed wavelet neural network based on PSO parameter optimization and gray relational deletion is improved obviously.(3)In this paper,a multi-index wind power forecasting model evaluation method is proposed.This paper establishes the evaluation index system of the forecasting model,uses the entropy method to determine the objective weight of the index,and analyzes the subjective weight of the index by the analytic hierarchy process.Thus the comprehensive weight of the index and the comprehensive evaluation value of the model was obtained,and thecomprehensive evaluation and the sorting of the model forecasting effect are realized.(4)In order to further improve the forecasting accuracy of forecasting model,a model fusion method based on multi-index optimization is proposed.On the basis of the multi-index evaluation of the model,the model is selected and the redundant judgment is made.The single model of fusion is optimized.Five kinds of combination methods are used for fusion modeling.Simulation experiments are carried out by a large amount of data.The results show that the model optimization can help to improve the forecasting accuracy of the fusion model.The entropy method,the Shapley method and the IOWA combination method can effectively improve the forecasting accuracy of the model,and the IOWA fusion model has a better prediction results.