Need Help with ElasticSolve problem

[alok]

Hey Everybody,
I am a complete noob when it comes to solid mechanics so I was hoping to get some help with a problem. I am a paleontologist working using finite element methods to remove deformation from fossils that have been squished over time. To do this I want to stretch a 3d mesh of the model along a given axis. So far I have generated a 3D mesh of a box with the shape of the fossil in the mesh with in the box. I have two boundaries declared (on two opposing sides of the box). I want to force the box to stretch to 1.5x it original length and then note the position of certain nodes inside the box. I want to set the problem so one of the box boundaries is fixed and the other is pulled away from it. Here come the dumb questions:

1.Is it correct to assume if what I want to detect are large deformations I should use non linear analysis?
2.Can I use a transitive solver with ElasticSolve so I can look at intermediate stretches (ie below the full 1.5x stretch)?
2.Will someone help me with the .sif file?

Right now the mesh is composed of 1st order tetrahedral elements. Thanks in advance for any and all help!
Cheers,
Alok

[mzenker]

Hi,

stupid question: If all you want to do is to stretch your model along a given axis, why do you want to use FEM? Purely geometric transformations might do what you want. Naively, I would try to use geometrical transformation tools of my favorite mesher or a CAD program or, if that fails, try to process the mesh file with Matlab or Scilab.

HTH,

Matthias

[alok]

Hey Matthias,

Not a stupid question at all, that is in fact what I have been doing for years and it generally works quite well! We have just run into fossils that have very odd deformations and a pure geometric stretch doesn't seem to quite fit the bill (skulls in particular because the teeth add rigidity to the jaw). I know with the conditions I have started with for the FEA trials would not make much difference but we just want to lay the building blocks for more complicated trials (ie: shear stresses instead of a simple stretch, modeling the box as sediment, differential material properties for parts of the fossil).

Cheers,
Alok

[mzenker]

I see.
As I am not a mechanics guy I hope someone else with more knowledge will answer...

Matthias

[Takala]

Hey Everybody,
1.Is it correct to assume if what I want to detect are large deformations I should use non linear analysis?
2.Can I use a transitive solver with ElasticSolve so I can look at intermediate stretches (ie below the full 1.5x stretch)?
2.Will someone help me with the .sif file?

Hi Alok,

1. You should check your material properties if your strains are off the elastic region (in other words, you have yielding). If yes, then you cannot use the linear elasticity. I think that is your case.
2. Do you mean transient? Yes you could do that. Then you just have to define time dependent displacements in order to get the intermediate states. Also you should use high enough timesteps to kill all the vibrations that may occur in transient simulation.
3. I think if you show some tenacity (in trying to create the model yourself) people tend to "throw you a bone" ( a streched one )

Welcome to the forum and Good luck!

Cheers,

Eelis

[mzenker]

Hi Alok,

you can use ElmerGUI to make your first sif file. Then you can decide if you continue working with ElmerGUI, or if you quit ElmerGUI and use manual sif file editing and the command line. It is mostly a question of personal taste.

HTH,

Matthias

[alok]

Thanks for the replies, both have helped massively! Especially liked the pun :p Combined with finally getting my automated meshing script to work this has been a great day!!!!

Cheers,
Alok

[alok]

Hey gain,
I have generated a basic sif file that is starting to do what I think I want it to do but a few things are not working, here is the sif:

Code: Select all

Header
  Mesh DB "." "mesh"
End

Constants
  Gravity(4) = 0 0 -1 9.81
  Stefan Boltzmann = 5.67e-08
  Permittivity of Vacuum = 8.8542e-12
  Boltzmann Constant = 1.3807e-23
  Unit Charge = 1.602e-19
End

Simulation
! to increase verbosity increase the value
  Max Output Level = 30
  Coordinate System = Cartesian 3D 
  Simulation Type = scanning

  Timestep Intervals = 5
  Timestep Sizes = 1
  Steady State Max Iterations = 1
  Steady State Min Iterations = 1
  Output Intervals = 1
  Output File = "stretch.result"
  Post File = "stretch.ep"
End

Body 1
  Name = "box"
  Equation = 1	
  Material = 1
  !Body Force = 1
End


Material 1
  Name = "matrix"
  Density = 1850
  Youngs Modulus = 1e8
  Poisson Ratio = 0.3
End


Solver 1
  Equation = Elasticity Solver
  Variable = Displacement
  Variable DOFs = 3
  Procedure = "ElasticSolve" "ElasticSolver"

  Calculate Loads = True
  !Calculate Stresses = True
  !Calculate Strains = True
  Calculate Boundary Weights = True 
  Stabilize = True
  !Bubbles = True
  Optimize Bandwidth = True

  !Neo-Hookean Material = Logical True
  
  Linear System Solver = Iterative
  Linear System Iterative Method = BiCGStabl
  BiCgStabl Polynomial Degree = 5  
  Linear System Preconditioning = ILU1
  !Linear System ILUT Tolerance = 1e-3    
  Linear System Max Iterations = 1500
  Linear System Convergence Tolerance = 1.0e-9
  Linear System Residual Output = 20

  Nonlinear System Newton After Tolerance = 1.0e-3
  !Nonlinear System Newton After Iterations = 5
  Nonlinear System Max Iterations = 100
  Nonlinear System Convergence Tolerance = 1.0e-6
  Nonlinear System Relaxation Factor = 0.5
  Steady State Convergence Measure = "solution"
  Steady State Convergence Tolerance = 1.0e-6
  
  End
  
Equation 1
  Active Solvers = 1
End


Boundary Condition 1
  Name = "contact"
  Target Boundaries = 1

  Displacement 3 = 0.0
  Save Scalars = Logical True
End


Boundary Condition 2
  Name = "front"
  Target Boundaries = 2
  Displacement 3 = Variable time
	Real MATC "10"
End

Boundry one is supposed to be fixed in the Y direction and boundary 2 is supposed to be displaced by 10 units for each time step(for 5 timesteps). I am pretty sure I did the mobile boundary wrong as the Elmer post file I generate is refuses to load a color mesh. There will most likely be some inconsistencies as I based it on a sif for a different ElasticSolve trial. Any help fixing this up would be extremely appreciated (Also: For the end result I want to compare the displaced coordinates of certain nodes to determine if the fossil inside the box is more or less symmetrical than when it started, can I do this by referencing the output files?)
Cheers,
Alok

[Takala]

Hi Alok,

try this

Code: Select all

  Displacement 2 = Variable time
   Real MATC "10*tx"

where "tx" is the representative of "Variable time". This is the convention used in sif, don't ask me why it is "tx". If you have several variables, then you will have "tx(0)", "tx(1)", etc...

Moreover, normally Coordinate 3 is the z-coordinate. If your displacement is along y-coordinate, maybe you should fix the coordinate 2?

Cheers,

Eelis