Research On Short-term Wind Speed Prediction Of Wind Farm Based On Hybrid Optimization Model

In recent years,due to the acceleration of environmental pollution and energy consumption,renewable energy has received more and more attention.Wind energy,as a kind of green energy,can replace the use of fossil fuels,thus reducing greenhouse gas emissions and effectively alleviating energy crisis.Because the wind speed has a strong randomness,the energy produced is extremely unstable.Therefore,accurate wind speed forecast is particularly important.In addition,the accurate prediction of wind speed will also help to reduce the operating cost and rotation reserve of the power system,which is of great significance for energy trading and power system management.In order to improve the accuracy of wind speed prediction,a hybrid prediction model is proposed,which combines the improved adaptive noise set empirical mode decomposition(ICEEMDAN),the optimal variational mode decomposition(OVMD),the improved Harris Hawk optimization algorithm(IHHO),the phase space reconstruction(PSR)and the outlier robust extreme learning machine(ORELM).In this model,the initial wind speed is decomposed into a series of intrinsic mode numbers using ICEEMDAN.In order to further reduce the nonstationary characteristics of the data,based on the VMD,the center frequency method and the minimum residual index were used to achieve the optimal variational mode decomposition(OVMD)for the subsequence with the highest frequency in the above eigenmodes.Then,in order to enhance the prediction ability of the hybrid model,an improved Harris Hawk optimization algorithm(IHHO)was proposed to optimize the parameters affecting the prediction performance in PSR and ORELM.IHHO with discrete integer encoding was used to synchronize optimization of embedded dimension,delay coefficient in PSR and hidden layer neurons in ORELM.The regularization coefficient and weight threshold in ORELM are optimized by using real-coded IHHO to further improve the prediction accuracy.Finally,the results of all the sub-sequences are accumulated to calculate the final forecast results of the original wind speed.In order to evaluate the prediction performance of the hybrid model proposed in this paper,the feasibility and effectiveness of the hybrid model were verified by using the actual data of a wind farm in central China.The results show that the hybrid model has a high prediction accuracy and can meet the actual demand.