Adaptive Zoning: HRR for mapping

egidio.cassone · September 4, 2019, 02:39

Hello, I was wondering on what is the effect of this option in the mapping strategy (it is not actually described in the 2.4.26 manual neither in the 3.0 one).
We are analyzing a Turbulent Jet ignited Methane fueled engine and having trouble with the pressure prediction. HRR is "more relaxed" (as the pressure trace) in respect of the experimental data, that's why we think that the chemical solver may need tuning.
We are using SAGE with the GriMech 3.0 reaction mechanism.
We have set a permanent 1mm grid size in the cylinder and in the pre-chamber with additional AMR with a maximum grid size of 0.125mm in the combustion phase.

Thankyou for any reply

MFGT · September 4, 2019, 02:58

Hi,

have you run multiple cycles?
Especially for pre chamber engines I think you need a first cycle that washes out your initial conditions. Furthermore, I would reduce the permanent grid size in the pre chamber to 0.5mm.

Blanco · September 4, 2019, 06:27

Hi,

considering the typical size of pre-chamber's holes I would suggest to put a fixed embedding surrounding holes and set it active for the whole simulation. Local fixed embedding level has to be decided considering holes diameter but I suspect at least 0.25 mm are needed. I agree w/ MFGT that multiple cycles are probably required, it depends on how you initialize the simulation.

Have you tested also other mechanisms in addition to GriMech?

egidio.cassone · September 10, 2019, 07:14

Hello and thank you for your answers.
However, the mesh is already refined as you sudgested (it is 0.5mm in the PC, my mistake in the first post) and around the PC holse I've created a spherical embedding with a 0.5mm grid. it is set as permanent.

I was trying to compare the laminar flamespeed obtained with 3 different mechanism and with the standard linear correlation.
at 10 bar and 300,350 and 400K the grimech flame is the fastest between the mechanism included in Converge and a skeletal mechanism for Methane (derived from grimech

) combustion while the laminar flame correlation describes "well" the behavior in an equivalence ratio between 0.7 and 1 while (extremely) over estimates the speed at higher E.R.
the reason for this comparison is that I've tried to run the PC computation with the G-eq approach (with sage inside and on flame) and results are quite better : we have now an over extimation of the pressure trace.

the next step is the sandia mechanism even if it seems too heavy.

i'll keep you up to date

September 4, 2019, 02:39	Adaptive Zoning: HRR for mapping	#1
egidio.cassone New Member Egidio Cassone Join Date: Sep 2019 Posts: 10 Rep Power: 6	Hello, I was wondering on what is the effect of this option in the mapping strategy (it is not actually described in the 2.4.26 manual neither in the 3.0 one). We are analyzing a Turbulent Jet ignited Methane fueled engine and having trouble with the pressure prediction. HRR is "more relaxed" (as the pressure trace) in respect of the experimental data, that's why we think that the chemical solver may need tuning. We are using SAGE with the GriMech 3.0 reaction mechanism. We have set a permanent 1mm grid size in the cylinder and in the pre-chamber with additional AMR with a maximum grid size of 0.125mm in the combustion phase. Thankyou for any reply

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
How to activate adaptive zoning strategy (multi-zone)	amin_u50	CONVERGE	2	November 3, 2017 14:05
Scalar Transport with Adaptive Meshing	emrahh	OpenFOAM Programming & Development	0	October 4, 2017 16:18
DES (DES,DDES,IDDES) turbulence model: Fixed or Adaptive Time step?	rohani	FLUENT	0	January 31, 2017 14:38
Dynamic Adaptive Mesh Coarsening	Will Anderson	FLUENT	0	October 26, 2010 17:26
Adaptive Timestepping: Possible Glitch?	Josh	CFX	9	August 25, 2010 21:52

September 4, 2019, 02:58		#2
MFGT Senior Member Tobias Join Date: May 2016 Location: Germany Posts: 265 Rep Power: 11	Hi, have you run multiple cycles? Especially for pre chamber engines I think you need a first cycle that washes out your initial conditions. Furthermore, I would reduce the permanent grid size in the pre chamber to 0.5mm.

September 4, 2019, 06:27		#3
Blanco Senior Member Blanco Join Date: Mar 2009 Location: Torino, Italy Posts: 193 Rep Power: 17	Hi, considering the typical size of pre-chamber's holes I would suggest to put a fixed embedding surrounding holes and set it active for the whole simulation. Local fixed embedding level has to be decided considering holes diameter but I suspect at least 0.25 mm are needed. I agree w/ MFGT that multiple cycles are probably required, it depends on how you initialize the simulation. Have you tested also other mechanisms in addition to GriMech?

September 10, 2019, 07:14		#4
egidio.cassone New Member Egidio Cassone Join Date: Sep 2019 Posts: 10 Rep Power: 6	Hello and thank you for your answers. However, the mesh is already refined as you sudgested (it is 0.5mm in the PC, my mistake in the first post) and around the PC holse I've created a spherical embedding with a 0.5mm grid. it is set as permanent. I was trying to compare the laminar flamespeed obtained with 3 different mechanism and with the standard linear correlation. at 10 bar and 300,350 and 400K the grimech flame is the fastest between the mechanism included in Converge and a skeletal mechanism for Methane (derived from grimech ) combustion while the laminar flame correlation describes "well" the behavior in an equivalence ratio between 0.7 and 1 while (extremely) over estimates the speed at higher E.R. the reason for this comparison is that I've tried to run the PC computation with the G-eq approach (with sage inside and on flame) and results are quite better : we have now an over extimation of the pressure trace. the next step is the sandia mechanism even if it seems too heavy. i'll keep you up to date