Improving the predictive capability of empirical heat transfer correlations for hydrogen internal combustion engines

Hydrogen internal combustion is widely considered a viable technology to achieve near-zero tailpipe CO2 and NOx emissions for difficult-to-electrify applications due to the maturity of ICE technology and production facilities. One-dimensional/zero-dimensional (0D) modeling is a valuable tool for engine development due to its relatively low computational requirements, but hydrogen combustion models still require further development. A large factor is gas-to-wall heat transfer, which is higher for hydrogen combustion due to higher flame temperatures and shorter quenching distance. For accurate prediction of in-cylinder temperatures, and therefore combustion rates and knock propensity, a well calibrated heat transfer model is essential. This paper evaluates existing heat transfer models against previously published experimental cylinder pressure and heat flux data from a Cooperative Fuel Research (CFR) engine with hydrogen Port Fuel Injection (PFI). A new heat transfer correlation is developed, utilizing a new fluid properties correlation to better represent the change in viscosity and conductivity with changing hydrogen concentration. Recent developments in 0D turbulence models improve the characteristic velocity calculation, which is augmented with a combustion term. This model is tested against a second dataset from the CFR engine with lambda from 1.0 to 4.0 and compression ratios of 9–13, showing improved performance versus previously published models. Whilst the new model provides more consistent results during combustion for variations in lambda and compression ratio, it requires improvement in its prediction of heat loss during expansion, and further validation at higher engine speeds and different engine configurations.