Wind Power Prediction Based On Whale Optimization Algorithm And Least Square Support Vector Machine

In recent years,the consumption of traditional energy sources is increasing and the energy crisis is becoming more and more serious.In order to solve the problem of en-ergy shortage,all countries are working on the development of new energy sources.As a green and clean energy source with huge reserves,wind power has received widespread attention and development,and the scale of wind power and the number of grid connec-tion are increasing day by day.However,due to the random intermittent and fluctuating nature of wind power,the large-scale integration of wind power into the power system brings great impact on the system operation.If we make plans in advance to meet the demand for electricity according to the expected power generation,we can not only maximize the utilization of resources,but also improve the safety and reliable operation of the power system.Therefore,the prediction of wind power is particularly important.Research on the prediction accuracy of wind power,the main work is as follows:(1)According to the wind power patterns,Complementary Ensemble Empirical Mode Decomposition(CEEMD)is selected to decompose the wind power data,which effectively eliminates the noise interference problem in Empirical Mode Decomposi-tion(EMD)by adding bi-directional white noise.By adding two-way white noise,it effectively eliminates the noise interference problem in Empirical Mode Decomposi-tion(EMD)and Ensemble Empirical Mode Decomposition(EEMD).Least Squares Support Veotor Maohine(LSSVM)is used to predict each component,and inequality constraints are replaced by equation constraints to reduce the difficulty of nonlinear wind power prediction.Algorithm(WOA)is used to optimize the penalty factor and kernel parameters of LSSVM.A CEEMD-WOA-LSSVM model is developed to de-compose the raw power data into several sub-series components by CEEMD,and the WOA-LSSVM model is used to predict each sub-series component separately,and the prediction results of each sub-series are superimposed to obtain the final prediction re-sults.The final prediction results are obtained by superimposing the prediction results of each sub-sequence.(2)Analyze the shortcomings of WOA in terms of convergence speed and merit-seeking ability.Propose adaptive power.The improved strategy of weight and prob-ability threshold and the introduction of the inverse Cauchy cumulative distribution variance function to improve convergence speed and accuracy,and the improved whale optimization algorithm(IWOA)is tested using five benchmark test functions.(3)To address the problems of CEEMD decomposition,Variational Mode Decom-position(VMD)is introduced to form a new prediction model VMD-IWOA-LSSVM model.The VMD decomposition is used to reduce the prediction difficulty,and the IWOA optimized LSSVM is used to predict each component of the VMD decompo-sition,and then the prediction results are superimposed to obtain the final results to effectively improve the prediction accuracy.