How to set rigid displacement of blocks in coupling analysis of tensor mechanics and heat conduction?

[lucuo]

I am now using heat conduction and tensor displacement, and there is thermal expansion of the material. As shown, I set [InclineNodisplacementBC] boundary conditions on the hypotenuse of the green block. If I want to set the rigid body displacement in the direction of the red arrow throughout the green block, how to set it.
Are there any mature code in MOOSE or which source files can be modified?

[GiudGiud]

@recuero @jiangwen84
I think we want a normal displacement BC here

[lucuo]

You mean DirichletBC? However, the DirichletBC displacement specified at the boundary will affect the original thermal expansion strain of the block.
Or are you using other code? Can you give me a specific case? Thanks a lot

[aikubo]

Just learning MOOSE so I’ll just try to guess.

You could try creating a pinned node in the middle of edge of the domain like this

But in the boundary conditions add value = non zero number for the disp variable in the desired axis. You could also use a function as needed with the Function systems and function BCs.

[yfun0815]

I'm also curious about this question.
I'm wondering is there any function in MOOSE which can fix a variable such as displacement or pore pressure to a constant value during nonlinear iterations. In most FEM program, this function is applicable, but I didn't find any actions in MOOSE.

[GiudGiud]

hello

during nonlinear iterations

a separate value for each nonlinear iterations?
You can make a DirichletBC depend on the current iteration number. You could code it or you could use parsed expressions involving this postprocessor
https://mooseframework.inl.gov/source/postprocessors/NumNonlinearIterations.html

[yfun0815]

No, I just wanna set a fixed value all over the calculation. Do you mean setting a value to a boundary using DirichletBC? This is only acting on boundary, but I want to act on block.

[GiudGiud]

so you are not solving any equation on that block?

[yfun0815]

Yes, I just want to solve displacement but fix pore pressure to 0 on a block.

[GiudGiud]

How does MOOSE treat the mutual point between mesh with different pore pressure?

MOOSE solves the nonlinear problem as posed. You have a nodal variable with a degree of freedom at every node, and you integrate an equation over quadrature points over each domain.

I expect the dashed line should be with 0 pore pressure, acting as a drained boundary, by which water can flow from soil to rock.

Would that really be the result of a single line of nodes with 0 pore pressure? Shouldnt you need elements to make a channel in which water can flow? I don't know.

[yfun0815]

Did you flip step 1 and step 2 here? It seems step 2 has the 0 pore pressure in the rock blocks

No, I didn't. Rock has a super low pore pressure but not 0 in step 2. Both step 1 and step 2 are from restarting.

Would that really be the result of a single line of nodes with 0 pore pressure? Shouldnt you need elements to make a channel in which water can flow? I don't know.

I see rock has a very large permeability, and linking to air, so the pore pressure should be nearly 0.

Because it is a continuous finite element family, the pore pressure value has to transition sharply over that node.

I think I did set a 0 pore pressure to rock. However, it is overwritten by a non-zero pore pressure by soil at this point.

[GiudGiud]

I think I did set a 0 pore pressure to rock. However, it is overwritten by a non-zero pore pressure by soil at this point.

Try flipping the ICs in the input? We should be able to get one to execute after the other.

[yfun0815]

Try flipping the ICs in the input? We should be able to get one to execute after the other.

No difference. I don't think there is difference in executing order.
Now my method is to apply a 0 pore pressure BC to this dashed line. Just like helping the kernel to apply 0 pore pressure. It looks like fitting my expectation, but it's a little bit silly by doing so.

[GiudGiud]

No difference. I don't think there is difference in executing order.

you are right. The IC projection algorithm will actually use the values from the two ICs and blend them. To get continuity to work as expected.
We could adapt it notably by leveraging the boundary parameter