Exploring the performance of biodiesel-hydrogen blends with diverse nanoparticles in diesel engine: A hybrid machine learning K-means clustering approach with weighted performance metrics

Abstract

Biodiesel, an eco-friendly fuel with lower greenhouse gas emissions, is a crucial alternative to traditional fossil fuels. Adding hydrogen and nanoparticles to biodiesel enhances diesel engine performance, reduces emissions, and improves fuel efficiency through better combustion and fuel atomization. In this study, various nanoparticles were experimented, and their performance, emission, and acoustic outcomes were systematically evaluated, resulting in a comprehensive ranking and clustering. The k-means clustering-Entropy-TOPSIS method is applied to weigh performance outcomes, rank the nanoparticles, and create clusters among them among best and worst, considering multiple criteria's in this study. This study further divides diverse nanoparticles into clusters, employing a hybrid machine learning k-means clustering method. The CO parameter carries the highest weight (57%), followed by UBHC (22%), owing to their significant variation induced by the addition of nanoparticles, surpassing the variability observed in other parameters. Manganese oxide nanoparticles exhibited superior performance across various critical parameters among all nanoparticles with a minimum centroidal distance of 25.97. Moreover, enrichment with hydrogen gas, combined with nanoparticles, significantly enhanced diesel engine performance, manifesting in an approximate 8 % increase in Brake Thermal Efficiency (BTE) and 23 % reduction in CO emissions. Identifying the optimal nanoparticles is crucial for enhancing biodiesel properties, and the paramount importance of hydrogen mixing lies in its potential to significantly improve combustion efficiency, emissions, and overall diesel engine performance.