Ultra-short-term Wind Speed Forecasting Based on a Hybrid FEEMD-ICS-LSSVM Method

摘要

Because of its adverse impact on the environment, it is increasingly important to reduce our dependence on fossil fuel resources. Therefore, as a renewable energy source without pollutant emission, wind energy has attracted a growing number of attention. However, due to the randomness of wind speed, the establishment of a high-precision wind speed prediction model is of great significance for the use of wind energy. So, this article projects a hybrid model. First, the original time series is decomposed into some intrinsic mode functions (IMFs) by using the fast ensemble empirical mode decomposition algorithm (FEEMD). Then, using the tent mapping and simplex algorithm, the parameters of the LSSVM are searched by the improved cuckoo search (ICS) algorithm. At last, use the optimized LSSVM model to predict the IMFs. The final result can be obtained by superposing the prediction results of each IMF. Ultra-short-term wind speed data from a Spanish wind farm in April 2019 are used to verify the model. The simulation results show the effectiveness and practicability of the prediction model.