Short-Term Load Forecasting Based on a Hybrid Deep Learning Model

Short term load prediction plays a critical role in the planning and operations of electric power systems especially in the modern days with high emphasis on integration of renewable energy resources. In this work, a hybrid deep learning model for short term load forecasting (STLF) is presented. The proposed method first decomposes the time series data into several intrinsic mode functions (IMF) using Empirical Mode Decomposition (EMD) and a reconstruction of the original series is obtained by suppressing the irrelevant IMFs. Detrended fluctuation analysis (DFA) is applied to each IMF to determine their scaling exponents for robust denoising performance. The denoised data is then used as input to the Deep Belief Network (DBN) model for modeling and prediction. Real data which represents hourly load consumption from the Electric Reliability Council of Texas (ERCOT) was used to evaluate the efficacy of the proposed method.