Optimization of Engine Performance Parameters using PSO and GA for a Diesel Engine Operated with Karanja Biodiesel

Abstract

This study investigates the enhancement of the braking thermal efficiency (BThEff) in a diesel engine running on Karanja biodiesel by optimizing four key engine performance parameters: engine speed, load, fuel blend ratio, and compression ratio. To approximate the link between input parameters and BThEff based on experimental data, a Random Forest (RF) model was used as a stand-in. Particle Swarm Optimization (PSO) and Genetic Algorithm (GA), two metaheuristic optimization methods, were used and contrasted throughout several separate experiments. The best parameter combinations were found via PSO and GA, both of which produced maximum BThEff values above 30%. GA converged more quickly (133.68 seconds) and had a marginally higher average BThEff (30.0665%) than PSO (30.0476%, 210.67 seconds). There was no discernible difference between the two approaches, according to statistical analysis employing t-tests and bootstrap validation (p >0.05). The results show that both algorithms work well for increasing engine efficiency with biodiesel blends, however GA shows a little advantage in terms of consistency and computational economy. By optimizing engine tune, our effort enables the use of Karanja biodiesel as a sustainable fuel substitute.