Performance and emission prediction using ANN (artificial neural network) on H2-assisted Garcinia gummi-gutta biofuel doped with nano additives

Abstract

The current work focuses on utilization of ANN (artificial neural network) for the prediction of performance and tailpipe emissions of Garcinia gummigutta methyl ester (GGME) enriched with H2 and TiO2 nano additives. For experimentation, H2 gas was introduced to the mixes containing TiO2 nanoparticles. Diesel, B10 blend (10% GGME biofuel + 90% Diesel), B20 (20% GGME biofuel + 80% Diesel), Diesel-TiO2 (Mineral Diesel with 100 ppm TiO2 nano additives), B10-H2-TiO2 (B10 blend with 100 ppm nano additives + 5 L/min of H2) and B20-H2-TiO2 (B20 blend with 100 ppm nanoparticles + 5 L/min of H2) were considered for experimentation. A constant mass flow rate of 10 L/min was used for the hydrogen flow throughout the test procedures. Test results were carefully analyzed to determine the performance and emission measures. Different speeds between 1800 and 2800 rpm were used for each test. When combined with pure Diesel and mixtures of biodiesel, these nanoparticles and hydrogen enhanced the performance data. For instance, the brake-specific fuel consumption was reduced but the power, torque, and thermal efficiency were increased. Although there was a modest rise in NO emissions, the primary goal of lowering CO, CO2, and other UHC emissions was met. The ANN models confirm and agreed the Diesel engine experimental work possesses minimal root mean square error (RMSE) and correlation coefficient values were estimated. This ideal model predicts and optimizes the engine output at a higher accuracy level, which gives better results compared with other empirical and theoretical models.