Electrical Load Forecasting Using LSTM, GRU, and RNN Algorithms

Forecasting the electrical load is essential in power system design and growth. It critical from both a technical financial standpoint as it improves performance, reliability, safety, stability well lowers operating costs. The main aim of this paper to make forecasting models accurately estimate based on measurements current loads electricity company. importance having predicting future loads, which will lead reducing costs resources, better electric distribution for companies. In paper, deep learning algorithms are used forecast loads; namely: (1) Long Short-Term Memory (LSTM), (2) Gated Recurrent Units (GRU), (3) Neural Networks (RNN). were tested, GRU model achieved best performance terms accuracy lowest error. Results show that an R-squared 90.228%, Mean Square Error (MSE) 0.00215, Absolute (MAE) 0.03266.