The increasing dependency on electricity and demand for renewable energy sources means that distributed system operators face new challenges in their grid. Accurate forecasts of electric load can solve these challenges. In recent years deep neural networks have become increasingly popular in research, and researchers have carried out many experiments to create the most accurate deep learning models. Players in the energy sector can exploit the increasing amount of energy-related data collected from smart meters to improve the grid’s operating quality. This review investigates state-of-the-art methodologies relating to energy load forecasting using deep neural networks. A thorough literature search is conducted, which outlines and analyses essential aspects regarding deep learning load forecasts in the energy domain. The literature suggests two main perspectives: demand-side management and grid control on the supply side. Each perspective has multiple applications with its challenges to achieve accurate forecasts; households, buildings, and grids. This paper recommends using a hybrid deep learning multivariate model consisting of a convolutional and recurrent neural network based on the scoping review. The suggested input variables should be historical consumption, weather, and day features. Combining the convolutional and recurrent networks ensures that the model learns as many repeating patterns and features in the data as possible.
Energy systems face challenges due to climate change, distributed energy resources, and political agenda, especially distribution system operators (DSOs) responsible for ensuring grid stability. Accurate predictions of the electricity load can help DSOs better plan and maintain their grids. The study aims to test a systematic data identification and selection process to forecast the electricity load of Danish residential areas. The five-ecosystem CSTEP framework maps relevant independent variables on the cultural, societal, technological, economic, and political dimensions. Based on the literature, a recurrent neural network (RNN), long-short-term memory network (LSTM), gated recurrent unit (GRU), and feed-forward network (FFN) are evaluated and compared. The models are trained and tested using different data inputs and forecasting horizons to assess the impact of the systematic approach and the practical flexibility of the models. The findings show that the models achieve equal performances of around 0.96 adjusted R2 score and 4–5% absolute percentage error for the 1-h predictions. Forecasting 24 h gave an adjusted R2 of around 0.91 and increased the error slightly to 6–7% absolute percentage error. The impact of the systematic identification approach depended on the type of neural network, with the FFN showing the highest increase in error when removing the supporting variables. The GRU and LSTM did not rely on the identified variables, showing minimal changes in performance with or without them. The systematic approach to data identification can help researchers better understand the data inputs and their impact on the target variable. The results indicate that a focus on curating data inputs affects the performance more than choosing a specific type of neural network architecture.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.