Research And Application Of A Hybrid Model Based On Multi-objective Whale Optimization Algorithm And Elman Neural Network For Short-term Wind Speed Forecasting

Wind energy,as one of the most important,promising,clean and renewable energy resources,which has been becoming the fastest growing renewable energy resource for electricity generation and is receiving increasing attention all over the world.Recently,in order to alleviate the energy crisis,improve the environment,and achieve economic and human sustainable development,countries all over the world have vigorously developed wind energy.However,in practical operation,the inherent volatility and intermittence of wind energy usually increase the difficulty of wind energy network,and then lead to unstable output power of wind farms,have a great impact on the output power quality and greatly increase the risk of grid safety and stability.Therefore,wind speed forecasting plays an important role in the work of wind power,the accurate and reliable forecasting results not only help the dispatcher to grasp the power change of the wind farm in advance,set up the scheduling operation plan in time,improve the energy conversion efficiency,reduce the risk,and increase the power generation etc.,but also conducive to stable operation of wind power and effective consumption,give a timely warning the safe and stable operation of the grid,and eventually avoid wind power fluctuations caused by random power loss or even power grid collapse.In recent years,domestic and foreign scholars have carried out a large number of wind speed prediction related research,wind speed forecast level has been greatly improved to some extent.The single prediction model is simple and easy to implement,but its prediction accuracy is often low,it is difficult to meet the needs of wind farm power generation.In contrast,the hybrid forecasting model based on optimization algorithms and data decomposition methods can greatly improve the wind speed prediction performance.However,existing hybrid models using single-objective optimization algorithms,only focus on the improvement of the prediction accuracy,which always ignore the improvement of the stability of the forecasting results.Few of them involve the multi-objective optimization algorithm.As a result,these models usually tend to reduce the stability of the forecasting results,bring a huge challenge to the security and stability of wind power generation and wind power grid.Under these backgrounds,this paper proposes a novel hybrid model based on a newly proposed called the MOWOA and ENN,which mainly includes four modules:a data preprocessing module,optimization module,forecasting module,and devaluation module.The main contents can be listed as follows:First,before implementing the wind speed prediction,the CEEMD decomposition method in the data preprocessing is applied to decompose the original wind speed,eliminate redundant noise and extract the primary characteristics.Then,in the optimization module,we propose a new multi-objective optimization algorithm called MOWOA,which is used to optimize the weights and thresholds of ENN neural network.Finally,in the forecasting module we propose a novel hybrid forecasting model i.e.CEEMD-MOWOA-ENN.In order to evaluate the effectiveness and generalization ability of the proposed model,the evaluation module includes hypothesis testing,eight evaluation criteria,four experiments,and ten different wind speed datasets(10-min and 30-min)is introduced perform comprehensive evaluation on this system.The experimental results show that compared with the other fifth forecasting models(i.e.,single prediction models such as WNN,BPNN,ENN,ARIMA,Persistence model,LSSVM and hybrid models:WOA-ENN,EEMD-ENN,CEEMD-ENN,CEEMD-WOA-ERNN,SSA-CS-ENN,VMD-BBO-BPNN,CEEMD-MOALO-ENN,CEEMD-MODA-ENN,CEEMD-MOWOA-WNN),The proposed prediction model can not only possess strong generalization ability,but also greatly improve the wind speed prediction performance with better and more stable forecasting results.The proposed hybrid model in this paper can not only effectively reduce the prediction error,improve the accuracy of wind speed prediction,enhance the stability of prediction results,but also benefit the improvement of wind power generation,and the safety management of wind power grid.The main innovations of this paper can be summarized as follows:First,a new algorithm called the Multi-Objective Whale Optimization Algorithm(MOWOA)is successfully developed in this paper,which provides a novel viable option for solving multi-objective optimization problems.And the experimental results demonstrate that MO WO A outperforms two recently devolved algorithms(MOALO and MOD A)in terms of optimize performance.Then,the CEEMD decomposition method is adopted and applied to decompose,de-noises and reconstruct the wind speed data,which not only effectively excavates the inherent information characteristics of the wind speed sequence,but also is beneficial to the improvement prediction accuracy of wind speed.Third,this paper breaks down the boundaries of the previous single-objective optimization forecasting model.For the first time,the multi-objective optimization algorithm is applied to the predict wind speed time series,which can effectively overcome the shortcomings of the single-objective optimization algorithm,improve prediction accuracy and forecasting results and finally provides a new way of thinking in the field of wind speed prediction.Finally,the results show that the hybrid forecasting model proposed in this paper is superior to the other fifth forecasting models used for comparison.