A statistical prediction of thermal efficacy improvement and exhausts reduc-tion of CRDI engine energized with ternary blends using linear regression machine learning model

The intensifying environmental and energy-related issues have led to a significant surge in the need for discovering alternative fuel sources. Additionally, the labor and time involved in identifying the appropriate fuel for the engine being examined is considerable. In order to mitigate this workload, the utilization of statistical analysis might be employed in the pursuit of identifying suitable fuel blends. The present work utilizes a machine learning algorithm (MLA) to predict the emission and performance characteristics of a common rail direct injection engine fueled by waste pedicel biodiesel and Butanol 20 dual fuel. The research focuses on examining ignition, efficiency, and exhaust traits of various fuel blends with varying Injection timing (21º, 19º & 17º bTDC), with Exhaust Gas Recirculation of 10% and 20% also employed during experimentation. After performing a series of investigations PBD20Bu is used to draw maximum performance with minimized emissions with other test samples. This study employed the linear regression model. As a result, the R2 values of 0.98, 0.977, 0.975, 0.98, and 0.947, respectively, the findings suggest that the total accuracy of the predictions exceeds 94%. Shows that the linear regression model created exhibits a high level of accuracy in predicting the levels of NOx, CO, HC, smoke, and BTE.