Contact Mechanics - Friction Test

Numerical methods and mathematical models of Elmer
kevinarden
Posts: 2237
Joined: 25 Jan 2019, 01:28
Antispam: Yes

Re: Contact Mechanics - Friction Test

Post by kevinarden »

Did the same steps, received the same results, normal loads way too high. Built a similar model from scratch and made my own sif, same result, normal loads way too high.
maike151
Posts: 41
Joined: 10 May 2019, 10:00
Antispam: Yes
Location: Cologne, Germany

Re: Contact Mechanics - Friction Test

Post by maike151 »

Thank you so much for your effort.

Unfortunately I cannot find out how and where "NormalLoadVar % Values" gets its values. I also tried to put the Normalload as a Boundary Condition at the bottom of the cube (Force 3 = -9.81 * 7810.0 and Force 3 = -9.81 * 7810.0 * 10.0e-6, because I did not know if volume is considered in a bc) instead of the body force. But the values are stil way too high. I don't know where NormalLoadVar" gets its values from..
maike151
Posts: 41
Joined: 10 May 2019, 10:00
Antispam: Yes
Location: Cologne, Germany

Re: Contact Mechanics - Friction Test

Post by maike151 »

I think this is the first time NormalLoadVar is called in SolverUtils:

NormalLoadVar => VariableGet( Model % Variables,&
TRIM(VarName)//' Contact Normalload')

But "Displacement Contact Normalload" is a variable that is calculated and outputted in vtu or if you want through SaveScalars, but not a variable you could set as a bc. If you try to set it as a bc, you get an error. I cannot find another line where NormalLoadVar gets any values..
kevinarden
Posts: 2237
Joined: 25 Jan 2019, 01:28
Antispam: Yes

Re: Contact Mechanics - Friction Test

Post by kevinarden »

Do loop starting at line 3052 in Subroutine CalculateContactPressure
sums the nodal forces into NodalForce then calculate NormalLoadVector
DO i=1,n
j = NormalLoadVar % Perm( Indexes(i) )

IF( .NOT. NodeDone( Indexes(i) ) ) THEN
NodeDone( Indexes(i) ) = .TRUE.
WeightVar % Values(j) = 0.0_dp
NormalLoadVar % Values(j) = 0.0_dp
SlipLoadVar % Values(j) = 0.0_dp
END IF

k = FieldPerm( Indexes(i) )
IF( k == 0 .OR. j == 0 ) CYCLE
DO l=1,dofs
NodalForce(l) = LoadValues(dofs*(k-1)+l)
END DO

IF( ThisRotatedContact ) THEN
NormalForce = NodalForce(1)
ELSE
NormalForce = SUM( NodalForce * Normal )
END IF
SlipForce = SQRT( SUM( NodalForce**2 ) - NormalForce**2 )

NormalLoadVar % Values(j) = NormalLoadVar % Values(j) - &
s * Basis(i) * NormalForce
SlipLoadVar % Values(j) = SlipLoadVar % Values(j) + &
s * Basis(i) * SlipForce

WeightVar % Values(j) = WeightVar % Values(j) + s * Basis(i)
END DO
maike151
Posts: 41
Joined: 10 May 2019, 10:00
Antispam: Yes
Location: Cologne, Germany

Re: Contact Mechanics - Friction Test

Post by maike151 »

That is interesting. In the comment above it says
! The nodal loads may be given either in cartesian or n-t coordinate system.

and in the comment above the Subroutine CalculateContactLoad() ist also says
! If we have normal-tangential coordinate system then also the load is in
! the same coordinate system.

Maybe one needs to define the load differently, I mean instead of using Force 1,2,3 Force N or T or something like that? Is that even possible?
maike151
Posts: 41
Joined: 10 May 2019, 10:00
Antispam: Yes
Location: Cologne, Germany

Re: Contact Mechanics - Friction Test

Post by maike151 »

Well I tried to put Normal Force as a bc but still wrong results..


So after trying to get this Fortran Code for a second I am even more confused:

So CalculateContactPressure gets its normalloads from LoadValues, LoadValues is

LoadVar => CalculateContactLoad()
LoadValues => LoadVar % Values,

CalculateContactLoad calls CalculateLoads using LoadVar as the Parameter "NodalLoads"

LoadVar => VariableGet( Model % Variables, &
TRIM(VarName) // ' Contact Load',ThisOnly = .TRUE. ).

But you cannot set Displacement Contact Load anywhere, but only get it as an output. I don't quite understand the sense of calculating loads using loads that are to be calculated.

In line 6345 and following in subroutine SetNodalLoads it appears that they are calculating loads based on the body force one has defined. But I don't know when and where they are used, apparently not for the normalload we are outputting.
maike151
Posts: 41
Joined: 10 May 2019, 10:00
Antispam: Yes
Location: Cologne, Germany

Re: Contact Mechanics - Friction Test

Post by maike151 »

or as in line 5521 in ElasticSolve.F90. I wonder if one could just get the NodalLoad in SolverUtils like that too.
kevinarden
Posts: 2237
Joined: 25 Jan 2019, 01:28
Antispam: Yes

Re: Contact Mechanics - Friction Test

Post by kevinarden »

Changed Contact Friction = Logical True to Contact Friction = Logical False and the normal load and slip load are much more reasonable. The value of the coefficient of friction still does not change the results.
mu_3_false.png
(34.57 KiB) Not downloaded yet
kevinarden
Posts: 2237
Joined: 25 Jan 2019, 01:28
Antispam: Yes

Re: Contact Mechanics - Friction Test

Post by kevinarden »

Using Contact Friction = Logical False, turned off the gravity (body force load) applied a Normal Force = Real -10. to the top face of the cube and the contact normal load is exactly 10.
NormalForce.png
(37.11 KiB) Not downloaded yet
maike151
Posts: 41
Joined: 10 May 2019, 10:00
Antispam: Yes
Location: Cologne, Germany

Re: Contact Mechanics - Friction Test

Post by maike151 »

If you turn off friction contact, the friction coefficients do not have any effect, because in the subroutine TangentContactSet the if-condition

IF( FrictionContact .AND. &
ListGetLogical( BC,'Stick Contact Global',Found ) ) THEN

is not met (when there is no friction the tangent contact set is empty).

I tried the same thing but I don't get Displacement Contact Normalload = 10 neither in Paraview nor in my SaveScalars-Output.. can you send me your sif?
Post Reply