Artificial intelligent-based analysis of VCR engine with biodiesel blends and modelling using uncertainty techniques

Abstract

The increase in population is also a factor that increases the vehicle strength. The biofuel derived from vegetation was found to be suitable and that can be used as biodiesel after chemical conversion. It can be utilized in an existing diesel engine without much modification that can reduce the usage of diesel. In this research, rice bran oil is used as biodiesel since it is available in plenty in south India. The main aim of this work is to create an uncertainty model and to optimize the parameter which gives improved performance by using Taguchi technique. Three factors, three level performance matrix, were considered in order to carry out the experimental investigation through uncertainty. “Design Expert 12.0” software was used for carrying out the uncertainty and graphical analysis of the data collected. The optimum values were obtained for the selected variables through analyzing the response surface contour plots and by solving the regression equation. The validity of the model was checked by analysis of variance (ANOVA) and for finding the significant parameters. Using such a model, the suitable blend which gives maximum performance was identified. Moreover, machine learning models were deployed to predict the brake specific fuel consumption (BSFC) and volumetric efficiency (VOL.E) of the engine tested based on the input features compression ratio (CR), blend (BLEND) and load (LOAD). Gradient boost repressor (GBR) has been found to be the superior model in predicting the multi-output parameters (BSFC and VOL) that decides the engine performance, with R2 of 0.987.