Homepage Books Releases Docs Discourse

$DEACTIVATED_SUBDOMAIN for 1D and 2D elements

Hi All,
To use $DEACTIVATED_SUBDOMAIN, I used two different approaches. But, having some issues. I am using OGS5.

First case: *pcs File:
For this case, I considered THM benchmark as a reference (https://github.com/ufz/ogs5-benchmarks/tree/master/THM), where directly $DEACTIVATED_SUBDOMAIN is used
#PROCESS
$PCS_TYPE
RICHARDS_FLOW
$NEGLECT_H_INI_EFFECT
2
$DEACTIVATED_SUBDOMAIN
1
2
$ELEMENT_MATRIX_OUTPUT
0

However, in my case, if I don’t use list of elements after SUBDOMIAN number, I got the error message as follows
“SURFACE not found or input stream error in GeoObject”
For convenience, I have attached a ppt slide (geometry is not in scale). I need to deactivate the shaded area. In this regard, I also attached *.geo file. I am wondering what the best way is to get rid of this error message without using list of elements? In addition, I was reading GeoSys/RockFlow user manual and I found following lines
$DEACTIVATED_SUBDOMAIN // Select the elements of a subdomain that not required by this process
1 // Number of sub-domain
… // List of the indices of the selected subdomain. The indices are given in element data of mesh file.
In above lines what does the indices means? Is it element list?

Second Case: *.msh file
I found in the same manual that “Each process can be provided with different mesh by means of the $PCS_TYPE subkey word”. Considering the comment, for each process I did not change the $NODES number, but I used relevant $ELEMENTS of the subdomain. Unfortunately, it is also not working

CO2 leakage.pptx (41.8 KB)

#FEM_MSH
$PCS_TYPE
MULTI_PHASE_FLOW
$NODES
4665

$ELEMENTS
9120

#FEM_MSH
$PCS_TYPE
DEFORMATION
$NODES
4665

$ELEMENTS
6770

#STOP
I am also having similar problem for 1D element. Any co-operation is highly appreciable

Thanks in advance.

Regards,
Mohammad
DEACTIVATE SUBDOMAIN.zip (139.1 KB)

Could you please attach the whole project. Otherwise it’s difficult to reproduce your problem.

Anyway, tt seems that you are using a SURFACE as a boundary condition (BC file) or source term (ST file) but your GLI file does not include a surface definition.

That could be it
Thomas

Dear Thomas,
I apologies for late reply. Due to sickness, I took two days off.

I have attached the whole project file. I will also check the SURFACE.

Regards,
Mohammad

DEACTIVATE SUBDOMAIN.zip (144.6 KB)

First of all, the message “SURFACE not found or input stream error in GeoObject” seems to be a bug and should have no influence on youz result. I included to your gli just a meaningless surface and the message disappeared.

#POLYLINE
  $NAME
  Domain_Boundary
  $EPSILON
  0.001
 $POINTS
 0
 1
 2
 3
 0
#SURFACE
 $NAME
  Domain_Boundary
 $POLYLINES
  Domain_Boundary

. . . we may should consider a bug report on Github. However, you don’t have to worry about that part.

Concerning the deactivation of subdomains, the problem seems to be something else.

Thank you again.

Yes, if you add SURFACE in *.gli, the message disappeared. On the other hand, from *.pcs file if we delete element lists, the simulation is not progressing. This situation is more complicated if we need to deactivate a domain which has 1d elements. Because, in this case, after subdomain number (as for example 3), if I need to put element lists (as for example 0,1,2,3…), unfortunately it is deactivating all sub domains. Simulations are progressing, but there are no results.

I discussed with Norihiro Watanabe personally about discrete fracture. For the time being, I have ignored the fracture propagation (However, I am working on it too). As, “discrete fractures are not supported in OGS mechanics simulations”, in fracture case for H2M condition, I also need to "disable fracture elements in mechanics by using $DEACTIVATED_SUBDOMAIN ". If I don’t disable 1d line elements, displacement vectors and stress fields are very high. It is not possible manage them using $NEGLECT_H_INI_EFFECT.

Therefore, it will be appreciable, if you check the bug for both 1d and 2d elements.

Thanks in advance.

Regards,

Mohammad