Rizqi Fitri Naryanto, Hadromi Hadromi, Ari Dwi Nur Indriawan Musyono, Rizki Setiadi, Wahyu Caesarendra
Diesel engines are extensively used across various sectors—including transportation, petrochemicals, power generation, military, and heavy machinery—with particularly widespread application in the automotive industry. An internal combustion engine burns a mixture of fuel and air to efficiently generate mechanical energy. Engine performance, measured by power output, torque, and fuel efficiently enhances with improved combustion efficiency. To predict diesel engine performance, this study employs a neural network model. The objective is to analyze engine behavior across various fuel blends and identify the most accurate machine learning-based prediction model. The best performance of the biodiesel fuel blends with diesel is in MIX-5, which is 69.8 kW in 4900 rpm. Based on testing on experimental data, the best neural network topology is obtained with three hidden layers. In this neural network topology, training is carried out on the engine performance and the regression value is 0,98416. Copyright © 2025 Reports in Mechanical Engineering. All rights reserved.
Mechanical Engineering Department, Faculty of Engineering, Universitas Negeri Semarang, Indonesia; Automotive Engineering Education, Faculty of Engineering, Universitas Negeri Semarang Indonesia, Indonesia; Department of Mechanical and Mechatronics Engineering, Curtin University, Malaysia
