ELM-QR-Based Nonparametric Probabilistic Prediction Method for Wind Power

Wind power has significant randomness. Probabilistic prediction of wind is necessary to solve the problem safe and stable grid dispatching with integration large-scale power. Therefore, this paper proposes a novel nonparametric probabilistic model for based on extreme learning machine-quantile regression (ELM-QR). Firstly, ELM-QR models multiple quantiles are established, then new comprehensive index (NCI) optimized by particle swarm optimization (PSO) obtain weighting coefficients corresponding lower upper bounds intervals. The final interval obtained integrating outputs coefficients. Finally, case studies carried out real farm operation data, simulation results show that proposed algorithm can narrower intervals while ensuring high reliability. Through sensitivity analysis comparison other algorithms, effectiveness further verified.