A Novel Short-Term Load Forecasting Method by Combining the Deep Learning With Singular Spectrum Analysis

One of the major issues about operation power systems is prediction load demand. Moreover, forecasting prime concern to system operators. Recently, integration elements, such as renewable energy sources, storages and electricity vehicle, brings more challenges, particularly when there are large fluctuations in cycle. This study concentrates on short-term demand proposes a hybrid method that combines Singular Spectrum Analysis (SSA) with deep-learning Neural Network (NN) techniques. In beginning, SSA technique applied an initial filter remove noises. Next, neural network, including Backpropagation (BPNN) Long Short-Term Memory (LSTM), developed trained. Then, trained network used core algorithm. Each has different forms combine networks. The performance proposed algorithm demonstrated using data recorded Taiwan. Furthermore, this compares results by five models, SSA, SSA-BPNN, ANN, SSA-LSTM, Wavelet (WNN) LSTM. reveal model provides best forecasts.