Investigation of the sources of combustion instability in low-temperature combustion engines using response surface models

In the search for more efficient and cleaner engines, low-temperature combustion engines have emerged with the potential to simultaneously increase engine efficiency and reduce nitrogen oxides and soot emissions. Although promising, low-temperature combustion engines have not been widely implemented due to difficulties controlling combustion phasing, combustion duration, and cycle-to-cycle variation. This article develops a methodology using detailed computational fluid dynamics simulations to investigate cycle-to-cycle instability of homogeneous charge compression ignition and reactivity-controlled compression ignition engines. Using computational fluid dynamics modeling, a large design of experiment is performed with small perturbations to the intake and fueling conditions. A response surface model is then fit to the design of experiment results to predict the combustion characteristics and exercised to determine the main sources of cycle-to-cycle variation. As expected, the results show that reactivity-controlled compression ignition and homogeneous charge compression ignition have significantly more variation than conventional diesel combustion. Reactivity-controlled compression ignition combustion phasing (CA50-crank angle when 50% of fuel is burned) is most sensitive to variations in diesel fuel mass, level of exhaust gas recirculation, and charge gas temperature. The peak pressure rise rate of reactivity-controlled compression ignition combustion is most sensitive to variations in gasoline fuel mass. The sources of variation for homogeneous charge compression ignition are similar to those of reactivity-controlled compression ignition combustion; however, trapped gas pressure and cylinder liner temperature become significant factors. Because of the late combustion phasing required for homogeneous charge compression ignition to maintain acceptable pressure rise rate, its cycle-to-cycle variation was found to be higher than that of reactivity-controlled compression ignition combustion for the same input variations.