Short-term wind power prediction based on CNN-LSTM-DIFF Transformer

To tackle the environmental and energy crises from fossil fuel consumption, the power industry is rapidly developing renewable energy, with wind power growing quickly. However, wind turbine output is random and uncertain due to various factors, challenging power system stability. Accurate wind power prediction, especially short-term forecasting, is crucial for grid stability, generation-consumption balance, and cost reduction. This paper proposes a short-term wind power prediction model based on CNN-LSTM-Diff Transformer. The Pearson correlation coefficient (PCC) analyzes wind power and influencing factors to determine input variables. The model extracts complex relationships in input data and captures long-term dependencies in time series. Experimental results show the proposed method outperforms conventional models in prediction accuracy.