SHORT-TERM FORECASTING OF ELECTRICAL LOAD DEMAND IN HANOI BASED ON EXTREME LEARNING MACHINE MODEL

Abstract

Accurate forecasting of the electrical load is a critical element for grid operators to make well-informed decisions concerning electricity generation, transmission, and distribution. In this study, an Extreme Learning Machine (ELM) model was proposed and compared with four other machine learning models including Artificial Neural Networks (ANN), Convolutional Neural Networks (CNN), Long Short-Term Memory (LSTM) and Gated Recurrent Unit (GRU). The dataset utilized for evaluating machine learning models were procured from the statistical analysis of the electrical load in the city of Hanoi, Vietnam. Prior to its utilization, the dataset underwent preprocessing procedures involving the removal of outliers and handling of missing values, thereby enhancing the computational efficiency of the models. According to the study results, the proposed model has superior performance when compared with the other four models, achieving the lowest error value. These outcomes substantiate the efficacy of the model, making it a good option for short-term load forecasting.