Compression ignition engine performance modelling using artificial neural network and hybrid multi criteria decision making techniques for the selection of fish oil biodiesel blend

Abstract

The ever increasing demand and depletion of fossil fuels along newlinewith environmental concern has prompted search for alternate fuels. One such newlinepotential substitute to fossil fuels is biodiesel that ensures sustainable energy newlinesource. Biodiesel is poised to make important contributions to world energy newlinesince it is renewable, bio degradable and non-toxic in nature. Various oils newlinehave been used in biodiesel production owing to their availability among newlinewhich fish oil is a significant one. In the present work, experimental investigations were carried out newlineon a single cylinder four stroke, air cooled, constant speed, direct injection newlinediesel engine with a rated output of 4.4 kW at 1500 rpm at different loads and newlineat different injection timings, 21o, 24o and 27obTDC for studying the newlineperformance, emission and combustion characteristics of diesel engine fuelled newlinewith Ethyl Ester of Fish Oil (EEFO) and its blends. newlineOxides of Nitrogen (NOx), Unburnt Hydrocarbon (UBHC) and newlineCarbon Monoxide (CO) emissions in biodiesel blends were lower than diesel, newlinewhereas smoke was found to be higher. The brake thermal efficiency for B20 newlinewas higher compared to diesel in the entire load spectra. The ignition delay and combustion duration were shorter for biodiesel blends than diesel which newlineresults in lower heat release rate, peak pressure and rate of pressure rise. newlineRetardation of injection timing caused decrease in emission and combustion newlineparameters like NOx, HC, CO, peak pressure, ignition delay, combustion newlineduration and heat release rate which increased with advancement in injection newlinetiming. However smoke and brake thermal efficiency exhibited an opposite newlinetrend with variation in injection timings. Artificial Neural Network (ANN) technique was developed to newlinepredict the engine performance through the limited experimental data.