Capturing Pressure Oscillations in Numerical Simulations of Internal Combustion Engines

Abstract

In an earlier publication [1] the authors compared numerical predictions of the mean cylinder pressure of diesel and dual-fuel combustion, to that of measured pressure data from a medium-speed, large-bore engine. In these earlier comparisons, measured data from a flush-mounted in-cylinder pressure transducer showed notable and repeatable pressure oscillations which were not evident in the mean cylinder pressure predictions from CFD. In this paper, the authors present a methodology for predicting and reporting the local cylinder pressure consistent with that of a measurement location. Such predictions for large-bore, medium-speed engine operation demonstrate pressure oscillations in accordance with those measured. The temporal occurrences of notable pressure oscillations were during the start of combustion and around the time of maximum cylinder pressure. With appropriate resolutions in time steps and mesh sizes, the local cell static pressure predicted for the transducer location showed oscillations in both diesel and dual-fuel combustion modes which agreed with those observed in the experimental data. Fast Fourier Transform (FFT) analysis on both experimental and calculated pressure traces revealed that the CFD predictions successfully captured both the amplitude and frequency range of the oscillations. Resolving propagating pressure waves with the smaller time steps and grid sizes necessary to achieve these results required a significant increase in computer resources.

A novel air management system for a large bore two-stroke naturally aspirated gas engine to reduce emissions

“…A direct comparison of the rectangular dotted black line representing the AMS1 resulted from the simulation to the measurement data shows a very close match to the experimental values. Some wiggles presented in the experiments could be predicted by the CFD simulations as shown in Gubba et al 13 However, this requires huge computational resources as described in Gubba et al 13 ; hence, this level of fidelity was not attempted in this work. A point to be noted is that, the stuffing box pressure at EPO must be different while using the different AMS designs.…”