Advancing Industrial Energy Efficiency with Data-Driven Methods: A Systematic Review and Roadmap for Research and Practice

Abstract

This review is a critical evaluation of how data-driven solutions revolutionize industrial energy efficiency. Based on our detailed and critical analysis of 162 peer-reviewed studies, empirical methods dominate energy-related research, with machine learning and optimization models being significant in terms of operational efficiency, energy conservation, and predictive maintenance. A closer examination of evaluation practices also shows a persistent gap between technical accuracy and real-world outcomes. The results of our study indicate that the field is advancing rapidly, albeit in a non-linear manner. We also provide a roadmap that identifies hybrid modeling, interoperable frameworks, and cross-sector transferability as essential for achieving sustainable and scalable progress. We thus outline how current barriers can be converted into opportunities for genuine transformation in industrial energy efficiency. Ultimately, the conclusion that emerges is that without rethinking validation and linking it directly to sustainability, AI’s contribution to energy efficiency risks becoming innovation without impact.