On the modeling of convective heat transfer coefficient of hydrogen fueled diesel engine as affected by combustion parameters using a coupled numerical-artificial neural network approach

• Prediction of convective heat transfer coefficient of hydrogen fueled diesel engine.
• Application of a coupled CFD-ANN approach for modeling convective HTC.
• Direct Injection Compression Ignition Hydrogen Fueled Engine was assessed.
• Effect of Temperature, Equivalence ratio and Liquid Mass Evaporated on convective HTC was studied.

It has long been recognized that injector and combustion parameters are vital to the performance of hydrogen fueled diesel engine as well as thermal properties. However, until today, it has not been possible to assess the convective heat transfer coefficient of hydrogen fueled diesel engine for head, liner and piston walls as affected by equivalence ratio, liquid mass evaporated and temperature. This study has made a significant step in advancing the field through modeling the phenomena using the computational fluid dynamics code coupled with the predicting ability of artificial neural network approach. The results indicated that the heat transfer coefficient values of the walls are tangibly greater at 3500 rpm than those of 2500 rpm. The impact of the aforementioned parameters on heat transfer coefficient at diversified ranges was covered. The result of different modeling implementations using various training algorithms at diversified neurons revealed that a multilayer perceptron neural network with back propagation learning algorithm using 3-17-3 structure denotes the best model with root mean square error equal to 9.13. Coefficient of determination (R2) for the three parts of liner, piston and head were obtained as 0.9870, 0.9975, and 0.9942, respectively in the training step.