Analysis on Engine Performance Using Azzola Biodiesel in a DI Diesel Engine Using ANN

In the present investigation, biodiesel prepared from azzola has been used as fuel in 4-stroke water-cooled engine. Biodiesel blended with diesel and ethanol has been used as test fuels. Three fuels diesel, B5D95, B5E5D90 have been used to conduct experiments by varying load as input parameters. Effects of different fuels and varying load on output parameters BP, IP, BTE, ITE, BSFC and ISFC have been studied. Experimental results revealed that blended fuel performed better than conventional diesel. At 4 kg load values obtained were maximum for IP and BP. IP and BP were read to be 2.84 kW and 1.54 kW, respectively, for BED95. ITE and BTE were found to be maximum at 4 kg load with B5E5D90 fuel. ITE was found to be 33%, and BTE was found to be 20.8%. Brake-specific fuel consumption was found to be lowest 0.412 Kg/Kw-hr at 4 kg with B5E5D90 fuel. The paper also investigates on the engine parameters using artificial neural network technique. In this, an ANN model has been created for prediction and optimization of engine parameters. Data obtained from experiments has been used for training the neural network. Error analysis revealed that B5D95 fuel performed better than diesel fuel at 4 kg load. Predicted values of ITE and BTE suggest that B5E5D90 performed better when compared to diesel. Regression coefficient was found to be 0.99893 for the entire neural network. The experimental and ANN studies suggest that both B5D95 and B5E5D90 have been better fuel when compared to diesel.