International Journal of I.C. Engines and Gas Turbines
Optimizing Internal Combustion Engine Performance and Fleet Fuel Economy through Machine Learning and AI using Big Data Analytics
Fuel Cell Product Development, Ashok Leyland Technical Centre, Chennai, Tamilnadu, Pin-600103, India
Abstract
Abstract
This paper presents a software-defined operational intelligence approach for optimizing internal combustion engine performance and fleet fuel economy using Big Data Analytics combined with Machine Learning and Artificial Intelligence. Real-world vehicle data is transformed into engineered features and virtual sensing outputs to estimate fuel consumption, driver behaviour, and component health without additional hardware. The proposed system integrates fuel optimization, predictive maintenance, driver assistance, and sustainability assessment into a unified one-stop platform. Machine learning enables accurate prediction, while an AI decision layer converts these predictions into practical operational actions such as driving guidance and maintenance alerts. The results demonstrate reduced fuel consumption, lower maintenance cost, improved fleet availability, and decreased environmental impact. The proposed ML–AI framework offers a scalable and cost-effective solution with future OEM integration potential, contributing toward reduced total cost of ownership and supporting sustainable mobility.
An AI-driven decision layer converts these predictive insights into practical operational actions such as personalized driving guidance, adaptive fuel-efficiency recommendations, early maintenance alerts, route-level optimization suggestions, and fleet-level performance dashboards. The architecture supports continuous learning, allowing models to improve over time as more operational data becomes available.
Experimental and field-level results demonstrate measurable reductions in fuel consumption, lower unplanned maintenance costs, improved fleet availability, extended component life, and decreased environmental impact through optimized combustion and reduced emissions. The proposed ML–AI framework offers a scalable, hardware-independent, and cost-effective solution with strong future OEM integration potential. By enabling data-driven operational intelligence, the system contributes toward reduced total cost of ownership while supporting regulatory compliance and long-term sustainable mobility objectives.
Keywords
Machine Learning; Artificial Intelligence; Big Data Analytics; Virtual Sensing; Fleet Fuel Economy; Sustainable Mobility.