[m102.992]

Hello Folks,
I am simulating a spark ignition IC engine case for combustion of premixed air-methane-mixture. I am using Sage with adaptive zoning. As turbulence model RANS k-eps RNG is used. The general simulation setup is executable without any major issues. Parameters regarding grid sizing, AMR, solver parameters, simulation time parameters, turbulence modelling etc. were set according to the Converge examples.
Unfortunately the heat release seems to be pretty slow and magnitudes lower compared to the data measured at the real engine. While the measured peak pressure at the test bench was about 42 bar, the simulation only reached 20.5 bar in peek, which is only approx. 1.5 bar increase compared to the motored case. Also the heat release graph from simulation looks not realistic. Please see the attached graph.
The load point is given with 3000 rpm at 40% throttle with a premix equivalence ratio of 1. Throtteling is imposed by using a pressure profile from file for the inflow boundary which was measured at the real engine. With that I am getting a good match for the in-cylinder pressure for the motored case (engine powered by the dynamometer).
I tried the SanDiego Mechanism as well as the GRI 3.0 Mechanism, which both should be suitable. But the Problem occurs with both.
I did not change Turbulent Schmidt Number or SAGE reaction multiplier so far. But it seems to me that the mismatch is originated by a more fundamental problem.
By the fact, that heat release starts within a reasonable time after ignition, I assume the problem is not the ignition delay from reaction mechanism. It is more like the simulated flame speed is much slower than in reality.
I have a few ideas left, but don't know how to approach this problem in the best way.

I am looking forward to hear your thoughts and ideas.

Best regards
Christopher