Inferential Statistics and Machine Learning Models for Short-Term Wind Power Forecasting

The inherent randomness, intermittence and volatility of wind power generation compromise the quality system, resulting in uncertainty system's optimal scheduling. As a result, it's critical to improve assure real-time grid scheduling grid-connected farm operation. Inferred statistics are utilized this research infer general features based on selected information, confirming that there differences between two forecasting categories: Forecast Category 1 (0–11 h ahead) 2 (12–23 ahead). In z-tests, null hypothesis provides corresponding quantitative findings. To verify final performance prediction findings, five benchmark methodologies used: Persistence model, LMNN (Multilayer Perceptron with LM learning methods), NARX (Nonlinear autoregressive exogenous neural network model), LMRNN (RNNs training methods) LSTM (Long short-term memory network). Experiments using real dataset show has highest accuracy when compared other approaches including persistence LMNN, network, LMRNN, 23-steps improved by 19.61%.