Empirical Mode Decomposition Based Deep Learning for Electricity Demand Forecasting

Abstract

Electricity is of great significance for national economic, social, and technological activities, such as material production, healthcare, and education. The nationwide electricity demand has grown rapidly over the past few decades. Therefore, efficient electricity demand estimation and management are required for better strategies planning, energy utilization, waste management, improving revenue, and maintenance of power systems. In this paper, we propose an empirical mode decomposition (EMD)-based deep learning approach which combines the EMD method with the long short-term memory network model to estimate electricity demand for the given season, day, and time interval of a day. For this purpose, the EMD algorithm decomposes a load time series signal into several intrinsic mode functions (IMFs) and residual. Then, a LSTM model is trained separately for each of the extracted IMFs and residual. Finally, the prediction results of all IMFs are combined by summation to determine an aggregated output for electricity demand. To demonstrate the applicability of the proposed approach, it is applied to electricity consumption data of city Chandigarh. Furthermore, the performance of the proposed approach is evaluated by comparing the prediction results with recurrent neural network (RNN), LSTM, and EMD-based RNN (EMD+RNN) models.

Keywords

• Deep learning
• electricity demand prediction
• empirical mode decomposition
• energy analytic
• long short term memory network