Day-Ahead Wind Power Forecasting Based on Wind Load Data Using Hybrid Optimization Algorithm

Accurate wind power forecasting is essential to reduce the negative impact of on operation grid and cost system. Day-ahead plays an important role in day-ahead electricity spot trading market. However, instability series makes forecast difficult. To improve accuracy, a hybrid optimization algorithm established this study, which combines variational mode decomposition (VMD), maximum relevance & minimum redundancy (mRMR), long short-term memory neural network (LSTM), firefly (FA) together. Firstly, original historical sequence decomposed into several characteristic model functions with VMD. Then, mRMR applied obtain best feature set by analyzing correlation between each component. Finally, FA used optimize various parameters LSTM. Adding results all sub-sequences acquires result. It turns out that proposed superior other six comparison algorithms. At same time, additional case provided further verify adaptability stability model.