Numerous studies on wind power forecasting show that random errors found in the\nprediction results cause uncertainty in wind power prediction and cannot be solved effectively using\nconventional point prediction methods. In contrast, interval prediction is gaining increasing attention\nas an effective approach as it can describe the uncertainty of wind power. A wind power interval\nforecasting approach is proposed in this article. First, the original wind power series is decomposed\ninto a series of subseries using variational mode decomposition (VMD); second, the prediction model\nis established through kernel extreme learning machine (KELM). Three indices are taken into account\nin a novel objective function, and the improved artificial bee colony algorithm (IABC) is used to\nsearch for the best wind power intervals. Finally, when compared with other competitive methods,\nthe simulation results show that the proposed approach has much better performance
Loading....