Computational Modeling and Multi-objective Optimization of Engine Performance with Waste Soya Oil-Based Biodiesel Using Genetic Algorithm and Utility Function

Climate change and other adverse effects of environmental pollution are a very serious challenge to mankind. As part of pollution mitigation measures, exploitation of bioresources for renewable energy and conversion of waste to wealth are greatly encouraged. Consequently, in this study, the engine performance of biodiesel produced from used soya oil was optimized. The multi-objective optimization was carried out using non-dominated sorting genetic algorithm II (NSGA-II) which used an artificial neural network (ANN) as its fitness function and a utility function coupled with ANN and optimized with particle swarm optimization (PSO). The optimized input parameters were blend and engine speed, while brake power (BP), basic specific fuel consumption (BSFC), brake thermal efficiency (BTE), torque, CO emission, NOx emission, HC emission and unit cost were the response variables. The results show that the optimum blend and speed obtained from the utility function were 91.52% and 1719.17 rpm, respectively. The Pareto front solution from the NSGA-II algorithm with relatively high utility values was found to be an excellent guide for engine designers.