Capturing Cyclic Variability in Exhaust Gas Recirculation Dilute Spark-Ignition Combustion Using Multicycle RANS

Dilute combustion is an effective approach to increase the thermal efficiency of spark-ignition (SI) internal combustion engines (ICEs). However, high dilution levels typically result in large cycle-to-cycle variations (CCV) and poor combustion stability, therefore limiting the efficiency improvement. In order to extend the dilution tolerance of SI engines, advanced ignition systems are the subject of extensive research. When simulating the effect of the ignition characteristics on CCV, providing a numerical result matching the measured average in-cylinder pressure trace does not deliver useful information regarding combustion stability. Typically large eddy simulations (LES) are performed to simulate cyclic engine variations, since Reynolds-averaged Navier–Stokes (RANS) modeling is expected to deliver an ensemble-averaged result. In this paper, it is shown that, when using RANS, the cyclic perturbations coming from different initial conditions at each cycle are not damped out even after many simulated cycles. As a result, multicycle RANS results feature cyclic variability. This allows evaluating the effect of advanced ignition sources on combustion stability but requires validation against the entire cycle-resolved experimental dataset. A single-cylinder gasoline direct injection (GDI) research engine is simulated using RANS and the numerical results for 20 consecutive engine cycles are evaluated for several operating conditions, including stoichiometric as well as exhaust gas recirculation (EGR) dilute operation. The effect of the ignition characteristics on CCV is also evaluated. Results show not only that multicycle RANS simulations can capture cyclic variability and deliver similar trends as the experimental data but more importantly that RANS might be an effective, lower-cost alternative to LES for the evaluation of ignition strategies for combustion systems that operate close to the stability limit.