Residential Electricity Demand Forecasting Employing a Highly Accurate BiLSTM Intelligent Model

Abstract

The proper operation of the electricity generation and distribution sections of microgrids requires the accurate prediction of electricity consumption. Furthermore, forecasting electric demand plays a significant role in the management and development of microgrids that can reduce power and economic losses. A novel model based on bidirectional Long Short-Term Memory (BLSTM) in pursuit of precise demand prediction is introduced in this paper. The proposed model performs impressively and achieves substantial improvements in key performance indicators (KPIs), shown by comparing its performance to five other AI-based models. To assess the effectiveness of our proposed method, a dataset of electricity consumption in Toronto, Canada from 2017 to 2021 is utilized. The outcomes of the simulations demonstrate the high accuracy of the proposed approach. © 2024 Elsevier B.V., All rights reserved.