Electricity and Heat Load Forecasting for Integrated Energy System Based on Variational Modal Decomposition and Deep Neural Networks

Load forecasting is crucial for the optimal scheduling of integrated energy systems(IES). With the increasing complexity of system load due to demand diversification and load coupling, this paper proposes a method for electricity and heat load forecasting based on Variational Mode Decomposition(VMD) and Deep Neural Networks . First, VMD decomposes the loads into components with different central frequencies, providing more stable load eigenmode functions. Then, a hybrid model combining Convolutional Neural Networks(CNN) and Long Short-Term Memory networks(LSTM) is used to extract time-series features, with parameter optimization performed using the Salp Swarm Algorithm(SSA). A case study of a real IES demonstrates the model's superior predictive performance.