Biodiesel derived from non-edible feedstocks has attracted significant interest because of the increasing need for cleaner alternative fuels. Marotti oil biodiesel is a novel renewable fuel source that paves the way for utilizing more non-arable lands for cultivation and has reduced emission characteristics compared to other non-edible fuel resources. In the present work, the performance and emission characteristics of a single-cylinder diesel engine running on Marotti oil biodiesel-diesel blends (10–40%) were experimentally investigated, and predictive models were developed using response surface methodology (RSM) and artificial neural network (ANN) techniques. The engine was operated at a constant speed (1500 rpm) under different load conditions. The experimental results showed that the brake thermal efficiency decreased up to 9.6% when the percentage of Marotti biodiesel was increased, whereas the brake-specific fuel consumption increased by 19.6%, mainly because of both the higher viscosity and lower calorific values of the blends. The exhaust gas temperature also decreases by 7.7% with a higher percentage of biodiesel content. In terms of emissions, significant reductions in carbon monoxide (from 50% to 29%), unburned hydrocarbons (at most 26%), and smoke opacity were detected during neat diesel operation. Nitrogen oxide emissions, on the other hand, rose more moderately at around 16 percent. Using the experimental data, predictive models were developed for the performance and emission parameters using the RSM and ANN methods. The prediction errors based on RSM and ANN models were in the range of 2.9% to 3.51% and 5.2% to 2.12%, respectively. The excellent agreement between the proposed and experimental results indicated that these two models possessed superior predictive ability. The results indicate that RSM and ANN can be used as powerful approaches for engine performance predictions with biodiesel blends of Marotti oil, significantly reducing the amount of experimental work required, and hence, time, cost, and effort.
Lalvani et al. (Fri,) studied this question.