High Pressure Direct-Injection (HPDI) is a technology option for engines used in mobile equipment applications where use of LNG as a fuel is desired. Using the combination of a diesel pilot injection and direct gas injection, HPDI has the potential to deliver low emissions, excellent transient performance, high efficiency, and high gas substitution. When the HPDI program was initially undertaken, in order to aid in initial hardware design, 3-dimensional computational fluid dynamic modeling was conducted to understand the mixing and reaction processes in the combustion chamber of an HPDI engine. Gaining insight into qualitative trends of operation parameters and hardware configurations was a first critical step toward delivering a hardware set to demonstrate HPDI natural gas combustion system capabilities.
To model the combustion of multi-component fuel at arbitrary constituent ratios, a combustion model based on a detailed chemical kinetics approach was employed. Several published mechanisms and combinations of established mechanisms were tested by comparing results with existing fumigated dual fuel engine results. The result shows that some of combined mechanisms for n-heptane combustion and methane combustion are capable of adequately predicting combustion behavior in diesel-natural gas dual fuel combustion systems. One of the reduced n-heptane mechanisms (by Patel et al.) also matched dual fuel combustion results reasonably well.
This preliminary simulation study was conducted with typical trapped air conditions and fuel quantities matching the energy delivery for a 100 % load condition in existing DI diesel engines. A full 360-degree mesh at intake valve closing was constructed and a detailed geometry of the gas injector nozzle and sac area was modeled in locally refined grids using a Caterpillar proprietary CFD code that accepts industry standard mechanisms. The diesel pilot injection followed by gas injection and resulting combustion inside an HPDI engine was simulated from IVC through the compression and combustion strokes. The operating parameters — such as diesel pilot injection timing, pilot injection amount, and start of gas injection — were varied, and the effect on IMEP, NOx, CO and cylinder pressure were investigated. It was shown that the start of gas injection is the strongest parameter for control of combustion. Subsequent to the work discussed in this paper, the hardware configuration established as optimal during the modeling work was carried forward to the physical engine testing and was successful in delivering the performance and emissions goals without modification, demonstrating the accuracy and value of modern combustion modeling.