mapFields question
Hi all,
I have problems with mapFields. What I want is a fully developed profile at the inlet of my pipe. So in the first case, I run a piece of pipe with cyclic b.c. whose outlet ("PER_PIPE_OUTLET") lies exactly at the inlet ("PIPE_INLET") of my second case. I created all meshes with ICEM. Now, the first case runs, everything is fine. I created a mapFieldsDict in the system directory of my second case that says Code:
/*--------------------------------*- C++ -*----------------------------------*\ http://www.cfd-online.com/Forums/ope...tml#post264825 i run Code:
mapFields -sourceTime 20630 ../../pipe/k_omega_sst_lowRe/ and get the error: Code:
/*---------------------------------------------------------------------------*\ Code:
patchMap ( PIPE_INLET PER_PIPE_OUTLET); Anyway, mapFields runs, but ends with like thousands of errors of Code:
--> FOAM Warning : |
1 Attachment(s)
Ok, finally I could get it running, it was some mistake in the /0 folder of the target case. I don't get the final "FOAM FATAL ERROR" any longer, but the warnings remain:
Code:
--> FOAM Warning : Attachment 33487 Does anyone know, what all these "--->FOAM Warning" with "Invalid normal for source face..." means? |
1 Attachment(s)
|
Ok, to anyone who get's problems with mapFields: Check whether you installed openFoam 2.3.x. It has a new implementation of mapFields with tons of bugs. I installed 2.2.2 just for mapping. Now it works.
|
Quote:
I ues ICEM to generate the mesh, and I got the same error with yours. Then I tried with 22x, error again. Im sure this is not about the field, its about the mesh. Then I tried neglect "-consistent", it works. I dont know why.~ Best, |
Hello Rodrigez. I try the same thing as you: first I run a simulation with cyclic boundary condiitons and than I want to map the solution to a new simultion. I got the error:
--> FOAM FATAL ERROR: Attempt to cast type patch to type lduInterface What did you do to solve your problem? |
Michael, you probably forgot to change the cyclic face in the boundary file of from wall to patch after converting the mesh to openFoam...
|
Hello Phillpp
ok. now I got it almost running if I change the boundary conditions in the cyclic case to e.g. zero gradient. but in the mapped case the zeroGradient bounardy conditions still rimains. Instead I want a noniniform List. Have you a hint to achieve this? |
Quote:
|
Actually, I have a 2D channel where I impose a constant pressure gradient as source term. In order to get the homogeous solution, I have cyclic boundary conditions in streamwise direction.
After i get a converged solution for this problem, I have a second channel with the same inlet mesh as my periodic problem. For the second channel I want to apply a zero Gradient boundary condition at the outlet. Now I try the map somehow the solution of the periodic problem to the inlet of my channel where I want to apply the zeroGradient condition at the outlet. I think it is quite the same problem as you had. |
Ok, now I got it.
Did you do what I wrote in the first posts? What goes wrong? |
I don't know exactly what is going wrong.
I had to change the cyclic condition in the blockMeshdict to type patch and also the boundary conditions in the fields (U, p, T etc. ) of my periodic simulation from cyclic to somethion to for example to zeroGradient. Then the mapField application worked. But unfortunately the application did not write a nonuniform list in my inflow condition. Now I solved it by using the results of the sample application as inflow condition |
If mappedField run successfully, it overwrites the files in the \0 folder. If you want, you can post the terminal output of your commands and we can try to find the error.
|
Create databases as time
Case : ../kepsilon_periodic_q10_map nProcs : 1 Source time: 200000 Target time: 200000 Create meshes Source mesh size: 120 Target mesh size: 1200 Creating and mapping fields for time 200000 Creating mesh-to-mesh addressing for region0 and region0 regions using cellVolumeWeight Overlap volume: 400000 Creating AMI between source patch IN and target patch IN using faceAreaWeightAMI AMI: Creating addressing and weights between 40 source faces and 40 target faces AMI: Patch source sum(weights) min/max/average = 1, 1, 1 AMI: Patch target sum(weights) min/max/average = 1, 1, 1 interpolating alphat interpolating nut interpolating Pk interpolating p_rgh interpolating p interpolating T interpolating k interpolating epsilon interpolating B interpolating U interpolating GradT End |
So this did non change the boundary values in your 200000 folder? How did you check that?
What kind of boundary is this inlet? Interpolating both pressure and velocity makes mostly no sense... you need to delete all files from the source folder except the ones you want to interpolate... |
I had similar problems.
using mapFields of 2.2.2 instead of 2.3.x works wonderfull!!! |
Hi Phillip,
I'm facing a similar problem with trying to get the outlet of the first part as my inlet for the second part. The mapFields utility works. The problem is that the pressure field remains 0 because the bc for pressure is uniform zero at the outlet of the first part. This creates a problem while running the second part as I don't get a smooth transition of flow from the first to the second part. What Bcs for u and p did you use at the outlet of the first part? Thanks! Best, Scram |
All times are GMT -4. The time now is 07:58. |