Short Term Wind Power Prediction Based On Modified EEMD And ELM

Fossil energy is the most important and mature energy in the world.However,fossil energy is limited by its inherent properties,limited storage capacity,and bad impact on the global environment,so it does not have long-term development benefits.Nowadays,the society has a huge demand for energy,so it is necessary to develop and utilize clean and renewable energy.Wind energy is a kind of clean and renewable energy.Wind energy as a natural energy,the strength of the wind is random and unstable,the efficiency of wind power generation is directly affected by the size of the wind,so the accurate prediction of wind power can better ensure the safety of wind power grid,and can provide a more practical and reliable basis for power dispatching scheme,convenient to take measures in advance,for the energy utilization rate of the whole new energy and wind power generation The production and efficiency of electricity are of great significance.How to establish a more effective algorithm model to predict wind power more accurately is the research core and innovation focus of this paper.For short-term wind power prediction,this paper mainly does the following research:(1)This paper describes the development significance of wind power in China,expounds the development status of wind power in China in recent years,summarizes the research focus of wind power at home and abroad in recent years,summarizes the common research methods of wind power prediction,and focuses on the related prediction modeling methods based on wind power time series.(2)According to the characteristics that the original data of wind power is time series data,the original power data is transformed into a time series signal.In view of the non-stationary and non-linear characteristics of the time series signal,a modified Ensemble Empirical Mode Decomposition method is used to decompose the signal in advance,and then the prediction is carried out by the Extreme Learning Machine algorithm optimized by Genetic Algorithm.A new combined prediction model MEEMD-GAELM is proposed.(3)The new combination forecasting method MEEMD-GAELM is tested and compared with the traditional method to verify the effectiveness of the combination forecasting method.(4)The Kernel Extreme Learning Machine(KELM)is introduced as the prediction algorithm to further improve the accuracy.Whale Optimization Algorithm is introduced to optimize KELM to further enhance the prediction performance of KELM.A new short-term wind power combination forecasting algorithm MEEMD-WOA-KELM is proposed.The superiority of this method is proved by experiment.(5)Considering the limitation of Whale Optimization Algorithm in global searching,a new optimization algorithm MWCWOA based on WOA is proposed.Adaptive weight factor and Cauchy mutation operator are introduced to optimize the local search and global search ability of WOA.Compared with Particle Swarm Optimization(PSO),Gray Wolf Optimization(GWO),traditional WOA and an improved algorithm WCWOA,the performance of MWCWOA is proved to be superior.(6)In order to further improve the prediction accuracy,a new combination prediction model MEEMD-MWCWOA-KELM is proposed by using the good optimization performance of the proposed MWCWOA algorithm to optimize KELM and combining with the MEEMD.The experimental results show that the prediction model has higher prediction accuracy.