Phase II of the project has focused on improving the initial analysis performed in the first phase by enhancing the various aspects of predictive combustion for lean burn spark ignition natural gas engines under variable composition fueling. These enhancements have incorporated validation data from a Cooper-Bessemer GMVH-10C3 engine located in New Jersey, which improves upon the lack of field data to bound the scope of composition variation. In simulation related endeavors, effort was made to improve the fundamental combustion physics related parameters, namely laminar flame speed, by developing...
Phase II of the project has focused on improving the initial analysis performed in the first phase by enhancing the various aspects of predictive combustion for lean burn spark ignition natural gas engines under variable composition fueling. These enhancements have incorporated validation data from a Cooper-Bessemer GMVH-10C3 engine located in New Jersey, which improves upon the lack of field data to bound the scope of composition variation. In simulation related endeavors, effort was made to improve the fundamental combustion physics related parameters, namely laminar flame speed, by developing a code base for distributed computing of the chemical kinetics solver, Cantera, and was key to improving upon the chemistry modeling used in the previous phase. Methods to improve numerical convergence were employed to reduce the time to solve large mechanisms, such as the Saudi Aramco Mechanism (v1.3).
Modeling of pre-chamber combustion from first principles, common input experimental heat release analysis and simulated heat release generation were additional components of improving model matching with pre-chambered engines. In its current state, manual optimization is required to tune the heat release curves based on guesses about the initial charge mass state, scavenging efficiency, fuel delivery and thereby the trapped equivalence ratio.