Neural Network Participation to Enhance Hydrogen-Biofuel CI Engine Performance and Combat Emissions

Abstract

To efficiently run, combat emissions, and predictive maintenance of the compression ignition engine (CIE), artificial intelligence participation plays a vital function. In this study, a one-cylinder CIE was outfitted with blended fuel (hydrogen, biofuel, and water) to combat emissions and enhance engine running efficiency. Simulink was used to collect data, which was then preprocessed and analyzed to predict emission characteristics of CIEs using deep reinforcement learning (DRL) and artificial neural networks (ANN). This paper presents a hybrid model. In this study, mean square error (MSE), R2, and accuracy are evaluated to show how well the prediction model worked to improve CIEs' running characteristics and fuel types. The proposed method is promising to enhance engine performance and combat emissions. Additionally, the model was found to have a low MSE, indicating that it can make accurate predictions for engine running characteristics and fuel types.