| With the rapid growth of the global economy,human society’s demand for energy is increasing.However,the consumption of large amounts of fossil fuels has had a tremendous impact on global environment and climate.Wind energy is widely distributed,clean and pollution-free,and has great development space,so it has attracted widespread attention from all countries.However,wind energy has the characteristics of strong randomness and volatility,which will bring great challenges to the stable operation of the power grid.Therefore,it is necessary to develop a method that can accurately predict wind power,reduce the negative impact of wind power grid integration,and improve the utilization rate of wind power in the power system while ensuring the safe and stable operation of the power system.According to the current research status of short-term wind power forecasting,this paper uses data decomposition technology combined with intelligent optimization algorithm to analyze and research the short-term wind power forecasting method.The main research work is as follows.The influencing factors of wind power are analyzed,and the data of Elia wind farm in Belgium and a wind farm in Turkey are selected as the original data sets for experimental research.Aiming at the problem of some abnormal data in the data set of a wind farm in Turkey,this paper uses the quartile method to remove the abnormal data from the perspective of power and wind speed,and then uses the linear interpolation method to fill in the removed data.In addition,in order to avoid errors due to the different dimensions of each feature in the wind power data,each feature of the wind power data is normalized.In order to alleviate the influence of randomness and volatility of wind power sequence on prediction performance,a short-term wind power forecasting method based on the complete ensemble empirical mode decomposition adaptive noise(CEEMDAN)and least squares support vector machine(LSSVM)is proposed.Firstly,CEEMDAN is used to decompose the original wind power series into several stationary components.then,the LSSVM prediction model is established for each component.In order to improve the prediction performance of the LSSVM model,the grey wolf optimizer(GWO)is used to optimize the regularization parameters and kernel parameters of LSSVM.Finally,the prediction results of all components are superimposed to get the final prediction results.The experimental results show that this model has a better prediction effect than the traditional prediction model.Aiming at the problem that some sub-components still have large fluctuations in the components decomposed by the CEEMDAN method,a short-term wind power combination forecasting model based on the secondary decomposition technique(SDT)and grey wolf optimizer is proposed.First of all,wavelet transform(WT)is used to decompose the sub-components with large volatility after decomposition by CEEMDAN method,which reduces the volatility of wind power sequence.In addition,in view of the limitations of a single prediction model,it is difficult to ensure that all prediction scenarios can be predicted accurately,LSSVM model,ELM model and BPNN model are used to predict the preprocessed wind power respectively.In order to improve the prediction performance,the GWO algorithm is used to optimize the parameters of LSSVM model,ELM model and BPNN model models,and the SDTGWO-LSSVM,SDT-GWO-ELM and SDT-GWO-BPNN are three individual forecasting models.Finally,these three individual forecasting models are weighted and combined to construct a combined forecasting model,and a combined forecasting model weight optimization method based on the GWO algorithm is proposed,which is used to find the optimal weight corresponding to each single predictive model.Experimental results show that the combined forecasting model effectively utilizes the advantages of each individual forecasting model and improves the forecasting accuracy of short-term wind power. |
