Short-Term Electricity-Load Forecasting by Deep Learning: A Comprehensive Survey

TL;DR

This survey defines STELF as predicting future load $\,\hat{y}_{t+h}$ from historical data and external variables $X_t$, formalized as $\hat{y}_{t+h} = f(y_t, y_{t-1}, \ldots, y_{t-n+1}, X_t)$. It comprehensively maps the end-to-end STELF process to eight research questions, covering datasets, preprocessing, feature extraction, modeling, training optimization, evaluation, and future directions. The review highlights a rapid growth of DL-based STELF literature (2014–2023), a shift toward public datasets and sophisticated architectures (including RNNs, CNNs, Transformers, and graph-based models), and a broad range of evaluation metrics for deterministic and probabilistic forecasts. It also identifies key challenges—standardization, generalization, interpretability, and real-time applicability—and points to future trends in graph neural networks, image-inspired techniques, and potential roles for large-language models in time-series forecasting. The work provides a structured, actionable roadmap for researchers and practitioners aiming to deploy robust, scalable DL-based STELF in modern power systems.

Abstract

Short-Term Electricity-Load Forecasting (STELF) refers to the prediction of the immediate demand (in the next few hours to several days) for the power system. Various external factors, such as weather changes and the emergence of new electricity consumption scenarios, can impact electricity demand, causing load data to fluctuate and become non-linear, which increases the complexity and difficulty of STELF. In the past decade, deep learning has been applied to STELF, modeling and predicting electricity demand with high accuracy, and contributing significantly to the development of STELF. This paper provides a comprehensive survey on deep-learning-based STELF over the past ten years. It examines the entire forecasting process, including data pre-processing, feature extraction, deep-learning modeling and optimization, and results evaluation. This paper also identifies some research challenges and potential research directions to be further investigated in future work.

Figures (3)

• Figure 1: The structure of this survey paper.
• Figure 2: STELF papers published between January 1, 2014, and December 31, 2023.
• Figure 3: Venue distribution of surveyed papers.