Mortar BC,contact + heat

[FredrikR]

Hi!

Is it possible to use mortar BC contact result in mortar boundary used in statcurrent and heatsolve?

I have set up mortar BC contact between two bodies in a stress solver simulation. In the simulation the contact area is increasing due increased load (imagine a spherical blunt pressing into a flat body). That part of the simulation is working fine. I would like to transfer electric current and heat from the blunt to the flat body via the active region of mortar boundary (the part of the boundary where the two bodies are in contact). When adding StatCurrentSolver or HeatSolver, the electric current or heat flow seems to be transferred through the whole mortar boundary and not by the active region only.

I can see in paraview a field variable "displacement contact active" that shows the contact region. I cant find out how to use that variable in the sif-file to set the active region for the statcurrent and heatsolver.

I hope someone can guide me here.

[raback]

Hi

There is a old test case "ContactFrictionHeating" that uses soft limiters to find the elastic contact and some conditional contacts for heating. Later in the development we would have used the mortar BCs which are optimal in accuracy. However, I'm dubious whether it is possible to combine current mortars for elasticity (elastic contact) and heat equation. This test case uses different machineries for the two. The mortar projector for elasticity extends larger than just the area of contact and the contact conditions pick the active nodes.

There is a keyword (that I found excatly this one use) "Level Projector Condition" that selects the active nodes in the time when the projector is build. If you try

Code: Select all

  Level Projector Condition = Equals "Displacement contact active"

it will fail because this keyword would be needed already when building the original mortar projector for which the active contact is not created. The elasticity also needs a potentially wider area.

There should be a field-specific mask for picking active rows of the mortar projector when building the constrained system in style:

Code: Select all

  Temperature Projector Condition = Equals "Displacement contact active"

Such thing does not unfortunately yet exist.

-Peter

[FredrikR]

Hi!

Thanks' for the clarification. Based on how and when the mortar is created I now understand why active contact didn't work.

Another alternative might be,
Considering the physical situation the conductivity in the interface might depend on the "pressure" between the two bodies. Some literature suggests that the contact resistivity, besides the bulk resistivity, also depends on the contact pressure and the flow stress. I made a very quick test using the contact normalload to control the conductivity of a thin layer on top of the flat surface. The conductivity of the thin layer:

Code: Select all

Electric Conductivity = Variable "displacement contact normalload", "vonmises"
    Real MATC "3*37.73e6*abs(tx(0))/(abs(tx(1)) + 1)"

The constant 3 I just arbitrary. 37.7e6 is the bulk conductivity. I need to go deeper into the physics behind contact conductivity and exactly what field parameters to use to model this properly. However, a quick test looks promising. The voltage is applied at the very top and bottom faces. Of course, the voltage seen at the non-loaded top face of the thin layer is physically faulty but the layer is effectively attenuating the current flow down to the body.

Any suggestions to this alternative would be appreciated.

Thanks to the Elmer team. A great FEM framework.

a.png

(161.38 KiB) Not downloaded yet

b.png

(179.74 KiB) Not downloaded yet

[raback]

Hi

I think you could get even more physical by getting inspiration from here:

Code: Select all

elmeruser@elmeruser-VirtualBox:~/elmerfem/fem/src/tests$ grep 'mortar bc coefficient' -ir
/home/elmeruser/elmerfem/fem/tests/DisContBoundaryMortarJumpB/discont.sif:  Mortar BC Coefficient Temperature = Real 1.0
/home/elmeruser/elmerfem/fem/tests/DisContBoundaryMortarJump/discont.sif:  Mortar BC Coefficient Temperature = Real 1.0

-Peter