Application of Machine Learning Approach in Internal Combustion Engine: A Comprehensive Review

Abstract

Alternative fuel combustion in internal combustion engines is a multidisciplinary topic of study combining chemical kinetics and nonlinear physical processes on length and time scales, including intricate molecular reactions. Alternative fuels such as biodiesel are becoming a popular source of energy for moving automobiles, heating houses, generating electricity, cooking meals, and so on. The two primary goals of alternative fuel research are to improve combustion efficiency while minimizing hazardous emissions. Machine learning enables data analysis-based methodologies for analyzing enormous volumes of combustion information collected from simulation or experiments at several spectral-temporal scales, thus exposing highly complex patterns and furthering combustion research. This article presents a comprehensive review of current research on the employment of machine learning in the disciplines of combustion phenomena. The study is an endeavor to review machine learning concepts and applications in chemical processes, modeling of combustion, combustion indices measurement, prediction of engine efficiency, and parametric optimization. Furthermore, the benefits and drawbacks of applying machine learning to engine emission performance are discussed. The present study aims to present a picture of the present position of machine learning application in combustion research, as well as to motivate researchers to continue their work. Machine learning methods are quickly developing in this age of big data, and there is tremendous promise for combining machine learning with combustion research to achieve some remarkable discoveries in the near future.