Artificial intelligence-based forecasting of dual-fuel mode CI engine behaviors powered with the hydrogen-diesel blends

Abstract

The vehicle sector has seen an upsurge in energy demand due to population expansion. Due to the escalating costs and exhaustion of fossil fuels, researchers are now dedicating more of their endeavors to exploring other options. The objective of this investigation, which is being conducted on a single-cylinder DI diesel engine propelled by hydrogen and diesel, is to optimize engine load. We evaluated the investigation on hydrogen/diesel composites against unadulterated diesel (D100). Different variations of fuels with varying percentages of hydrogen supplementation, viz., DH5 (diesel and 5% hydrogen), DH10 (diesel and 10% hydrogen), and DH15 (diesel and 15% hydrogen), were analyzed. In the experimental investigation, we used an injection pressure of more than 220 bar at 18:1 compression ratio for load optimization, and we used a single fuel injector with a diameter of 0.25 mm, The results showd improved brake thermal efficiency of 1.8% for DH5, 5.2% for DH10, and 17.6% for DH15, along with a drop in fuel consumption of 1.7% for DH5, 14.5% for DH10, and 31.6% for DH15. In addition, the addition of hydrogen to diesel demonstrates a promising reduction in smoke emissions of 10.5%, carbon monoxide (CO), and hydrocarbon (HC) emissions by DH15 under full load conditions. The results were subjected to regression analysis using an artificial intelligence network (ANN), to enhance the performance and reduce emissions of fuel mixtures. The ANN proves to be a very good method for the regression of data and prediction. The ANN provides an excellent fit for all the parameters, with a fitting value of more than 99%.