Ultra-Short-Term Wind Speed Forecast Based On The LSTM Optimized By The Sparrow Search Algorithm

With the increase of energy demand and the emphasis on ecological environment safety,wind energy as one of the renewable energy sources has attracted global attention.With the rapid development of wind energy,wind energy has the potential to support sustainable economic development and protect environmental safety.However,the randomness and nonstationarity of wind have adverse effects on high precision wind power generation.Reliable and accurate wind speed prediction is the basis of effective utilization of wind energy.Accurate wind speed prediction can effectively reduce the risk of wind uncertainty to the power system,which plays an important role in the stable operation of wind farms and power grids.At present,hybrid model is still the focus of wind speed prediction research,but how to choose single prediction model and optimization algorithm still need to be further studied.Based on the above problems,the thesis has carried out the research of ultra-short-term wind speed prediction.In view of real-time mutation and dependence of wind speed data,long-term memory network(LSTM)is selected as the basic prediction model for ultra-short-term wind speed prediction because it takes into account the dependence relationship between continuous events,has a unique memory structure and long-term memory ability,and is very suitable for dealing with problems highly related to time series.Since the prediction result will be affected by the initial value of the model network,the optimization algorithm can be used to optimize the parameters and generate the optimal parameters adaptively to improve the prediction performance of the model.Sparrow search algorithm(SSA)has the characteristics of strong stability,strong robustness,fast convergence speed and so on.It has certain local development ability and good global search ability,and has certain help to search the optimal parameters.In order to improve the accuracy of wind speed prediction,an ultra-short-term wind speed prediction model based on sparrow search algorithm and long short-term memory neural network is established.The sparrow search algorithm is used to optimize the weights and thresholds in the LSTM model,so that the model can set parameters adaptively and predict the wind speed data under the condition of the optimal parameters.In order to evaluate the prediction performance of LSTM model after sparrow search algorithm optimization,particle swarm optimization(PSO)was selected to optimize the LSTM model and compare with it.Finally,in order to further verify the prediction effect of SSA-LSTM prediction model,the SSA-LSTM model is compared with other single prediction models.For the above experiments,the experimental results are as follows.Firstly,through the simulation experiments of the SSA-LSTM model and the LSTM model in different wind speed data sets,it is found that compared with the LSTM model,the root mean square error of the SSA-LSTM model in the four data sets is reduced by at most 25%,and the average absolute error by up to 30%,the mean absolute percent error was reduced by up to 26%.It shows that the sparrow search algorithm has a significant effect on optimizing the parameters of the LSTM model.Secondly,the comparison between the SSA-LSTM model and the PSOLSTM model in different data sets shows that the values of each evaluation index of PSOLSTM model is higher than that of SSA-LSTM model,indicating the parameter optimization effect of the particle swarm optimization algorithm on the LSTM model.It is not as good as the sparrow search algorithm,which proves the optimal performance of the sparrow search algorithm.Finally,the SSA-LSTM model is compared with models such as support vector machine,BP neural network,and bidirectional long short-term memory neural network,and it is found that the indicators of the SSA-LSTM model are the lowest among different models.To sum up,through the comparison of different experiments,the results show that the wind speed prediction model based on the sparrow search algorithm optimized by LSTM proposed in this thesis reduces the prediction error,obtains a higher prediction accuracy than other models,and is better than other comparative models.It is suitable for ultra-short-term wind speed prediction and has high effectiveness and practical value.