Supervised learning technique to develop predictive regression model and performance evaluation of dual bio-diesel blends

The experimental investigation was carried out to know the influence of dual bio-diesel blends on a performance of diesel engine and its emission characteristics. Jatropha (J), Mahua (M) and Diesel (D) were taken into account to prepare dual bio-diesel blends and tested on single cylinder four-stroke variable compression ratio (VCR) diesel engine test rig. Experimental test was done at 50% load with constant speed of 1500 rpm by varying compression ratio from 15:1 to 17:1, brake power as 1, 2 and 3 KW with different dual bio-diesel blends. The maximum BSFC is observed to be 0.62Kg/KW-hrs. For 10% bio-diesel blend and maximum EGT is found to be 184 °C for 30% bio-diesel blend. At constant compression ratio, as percentage of dual bio-diesel increase, the exhaust gas temperature (EGT) increased by 11.51%. However, BSFC seen to be decrease by 54.83%. The novelty of this paper is to determine the optimal levels for BSFC and EGT and to develop linear regression model by using machine learning technique to validate the experimental results.