Artificial Intelligence-Based Ultra-Short-Term Power Load Forecasting

Abstract

In the context of significant renewable energy integration, power load forecasting is viewed as an essential task in energy management and power system operation and scheduling. In an effort to enhance the accuracy and precision of power load prediction, a predictive technique based on Long Short-Term Memory (LSTM) networks enhanced by the quantum-behaved particle swarm optimization (QPSO) is applied to ultra-short-term power load prediction in this paper. Initially, normalization is used to preprocess power load data before it is divided into training and testing datasets. Subsequently, global optimization of the LSTMs essential hyperparameters and network architecture is conducted via QPSO, resulting in the development of a QPSO-LSTM forecasting model. Subsequently, the forecasting model is evaluated by employing the coefficient of determination (R), mean absolute error (MAE), mean absolute percentage error (MAPE), and root mean square error (RMSE) as performance metrics. Finally, comparative experiments are conducted between the proposed model and traditional neural network models. The findings demonstrate that the QPSO-LSTM model offers enhanced forecasting precision and optimal fitting performance.