Dual fuel diesel-methane low temperature combustion (LTC) has been investigated by various research groups, showing high potential for emissions reduction (especially oxides of nitrogen (NOx) and particulate matter (PM)) without adversely affecting fuel conversion efficiency in comparison with conventional diesel combustion.
However, when operated at low load conditions, dual fuel LTC typically exhibit poor combustion efficiencies. This behavior is mainly due to low bulk gas temperatures under lean conditions, resulting in unacceptably high carbon monoxide (CO) and unburned hydrocarbon (UHC) emissions. A feasible and rather innovative solution may be to split the pilot injection of liquid fuel into two injection pulses, with the second pilot injection supporting the methane combustion once the process is initiated by the first one.
In this work, diesel-methane dual fuel LTC is investigated numerically in a single-cylinder heavy-duty engine operating at 5 bar brake mean effective pressure (BMEP) at 85% and 75% percentage of energy substitution (PES) by methane (taken as a natural gas surrogate). A multidimensional model is first validated in comparison with experimental data obtained on the same single-cylinder engine for early single pilot diesel injection at 310 CAD and 500 bar rail pressure. With the single pilot injection case as baseline, the effects of multiple pilot injections and different rail pressures on combustion emissions are investigated, again showing good agreement with experimental data.
Apparent heat release rate and cylinder pressure histories as well as combustion efficiency trends are correctly captured by the numerical model. Results prove that higher rail pressures yield reductions of HC and CO by 90% and 75%, respectively, at the expense of NOx emissions, which increase by ∼30% from baseline. Furthermore, it is shown that post-injection during the expansion stroke does not support the stable development of the combustion front as the combustion process is confined close to the diesel spray core.