Application of machine learning tools for energy efficiency in industry: A review

• Published in: Energy reports Vol. 6; pp. 1181 - 1199
• Main Authors: Narciso, Diogo A. C, Martins, F. G
• Format: Journal Article
• Language: English
• Published: Amsterdam Elsevier 01-11-2020; Elsevier Ltd
• Subjects: Energy efficiency; Industrial data; Machine learning; Manufacturing industry; Process industry; Energy efficiency; Process industry; Industrial data; Machine learning; Manufacturing industry
• ISSN: 2352-4847; 2352-4847
• DOI: 10.1016/j.egyr.2020.04.035

The current industrial context favors the generation of large amounts of data, most of which still seems to remain unexplored by the majority of enterprises. This paper presents a literature review on methodologies reported in the scientific literature exploring the potential value of industrial data via the utilization of Machine Learning tools for energy efficiency related goals. This work identifies and examines in detail the scientific contributions published up to date. A total of 42 published papers are found to present original contributions in this field, and addressing multiple energy efficiency challenges. A descriptive analysis is presented and demonstrates that the number of published works in this field is rapidly growing. The majority of contributions address challenges in petrochemical industries, and namely in ethylene production. There is still a very limited number of published papers addressing the application of Machine Learning tools on energy related objectives in other types of industries. The technical content of all identified papers is thoroughly reviewed and their key features and objectives are highlighted. A number of important themes across the final list of papers emerges, addressing challenges such as energy consumption forecast, energy analysis and energy optimization. A framework identifying the key goals reported on the set of 42 papers and the tools proposed to address them is suggested. This framework provides a summary on existing tools and facilitates the identification of research needs in this field. Additionally, the proposed framework serves as a reference guideline for the manufacturing and process industries on the selection of adequate Machine Learning tools for energy efficiency objectives via the utilization of industrial data.