Resarch On Short-Term Wind Power Prediction Based On Improved Chicken Swarm Optimization Algorithm

With the development of society,the energy demand is increasing,the depletion of fossil energy is inevitable,and the problem of environmental degradation is becoming increasingly prominent.There is an urgent need to vigorously extract low-carbon,renewable energy to solve the energy problem.Wind energy,as a low-carbon renewable energy source,has received widespread attention and development in recent years.However,the large-scale integration of wind power will have a greater impact on the smooth and safe operation of the power grid,so it is necessary to predict wind power.Short-term wind power prediction can also provide a basis for grid scheduling and control,effectively reduce the impact of wind power on the grid,and increase the security,stability,and reliability of the grid.Therefore,research on short-term wind power prediction is necessary.Firstly,this paper summarizes the relationship between wind speed and wind power prediction,and the different classification methods of wind power prediction.Then the first mock exam data of a wind farm are analyzed and processed,and three common single wind power algorithms,including wavelet neural network,Elman neural network and support vector machine(SVM),are used to conduct short-term wind power forecasting model and simulation.The advantages and disadvantages of different algorithms are compared and the prediction results are analyzed.The experimental results show that the prediction effect of WNN network model in three single models is effective.In order to further improve the effect of wind power prediction,then use the chicken swarm optimization(CSO)and ICSO algorithm to optimize the WNN network,establish the CSO-WNN and ICSO-WNN power prediction models,and carry out simulation analysis.Through comparing the prediction results and errors,it is proved that CSO has better global convergence and calculation robustness,and can find out faster The advantages of the global optimal solution effectively improve the prediction accuracy,and the predictionaccuracy of ICSO-WNN combined prediction model is higher than that of CSO-WNN model.Finally,a combined prediction method based on deep learning is proposed.This method is based on a deep belief network,which optimizes the deep belief network through an improved chicken flock algorithm.After the parameters are reversely fine-tuned through the Kalman network,a combined prediction model based on ICSO-DBN-Kalman is established.Analysis shows that the combined forecasting model can improve wind power forecasting accuracy and has certain applicability.