Short-term wind power forecasting model based on improved Tasmanian Devil algorithm

In order to solve the problem of insufficient convergence speed and convergence accuracy of the prediction model caused by the variability and volatility of wind speed, a combined prediction model of SSA-VMD-ITDO-LSTM based on variational mode decomposition, improved Tasmanian devil optimization algorithm and long short-term memory network was proposed. Firstly, the SSA-VMD method is used to decompose the original wind speed series to obtain relatively stable wind speed sub-series, and the LSTM model is used to obtain the time correlation of the input data series. Then, the Brownian random walk strategy is used to update the population position and the stochastic center of gravity differential evolution strategy are used to optimize the global and improve the robustness to find the optimal solution. The parameters of the LSTM model are optimized with the improved Tasmanian devil optimization algorithm, and finally the final predicted value of wind power is obtained by superimposing the predicted value of each subsequences. The model is applied to the actual wind power prediction, and compared with other eight models, the experimental results show that the three average errors of the stroke power value of the model are the smallest, and the RMSE, MAE and MAPE values are 0.83367, 0.67062 and 1.5439%, respectively.