Research On Short-Term Prediction Method Of Photovoltaic Output Power

With the growing contradiction between energy demand and environmental demand,the development and utilization of renewable energy has become the common choice of countries around the world.Among a variety of renewable energy power generation methods,photovoltaic power generation has attracted more and more attention from countries due to its rich reserves,safety and reliability,and green pollution-free characteristics.However,photovoltaic power generation is greatly affected by external factors,and its output power is characterized by randomness and indirectness.Direct connection of photovoltaic power generation to the power grid will have many influences on the power grid dispatching and operation.Therefore,it is of great practical significance to accurately predict the photovoltaic output power.Based on the analysis of the factors affecting photovoltaic output power,this study proposes a photovoltaic output power prediction model based on support vector machines,and conducts in-depth research on parameter optimization problems of support vector machines.The photovoltaic power generation technology is introduced,the basic principles of photovoltaic power generation,the framework composition and the grid-connected manner of photovoltaic power generation systems,and the characteristics of photovoltaic power output power are discussed.The influence factors of photovoltaic output power are analyzed.The correlation degree between various meteorological factors and photovoltaic output power is calculated quantitatively by using the grey correlation degree method,and the factors with higher correlation degree are selected as the input of the prediction model.The principle of support vector machine model is studied,the characteristics of the support vector machine model are analyzed,and a photovoltaic output power prediction model based on support vector machine is established.In order to further improve the accuracy of photovoltaic power generation data,the wavelet transform is used to reduce the noise of various meteorological data and the photovoltaic output power prediction experiment is carried out.The results show that the wavelet denoising can effectively reduce the prediction error of the model.The principle of multiverse optimization algorithm(MVO)is studied,and three improvement strategies are used to improve it.A hybrid improved multiverse optimization algorithm(HIMVO)is proposed.Five typical test functions are used to test the HIMVO algorithm,and the test results are compared with other algorithms.The results show that the optimization accuracy and stability of HIMVO algorithm are better than other algorithms.The HIMVO algorithm is used to optimize the parameters of SVM model,and a prediction model of photovoltaic output power based on HIMVO-SVM is established.The HIMVO-SVM model is used to predict the photovoltaic output power under different weather and different months,and the prediction results are compared with other models.The experimental results show that the average relative error of the HIMVO-SVM model in several sets of prediction experiments is less than 10%.Besides,compared with several other model models,the HIMVO-SVM has higher prediction accuracy and prediction stability,which validates the effectiveness of the proposed model in predicting photovoltaic output power.