Learning Forum

[vvaibhav1002]

Hello all.

I need to simulate stresses in a rectangular glass slab during cooling from a high temperature. As the conductivity of glass slab is low, stresses will be generated due to non linear thermal expansion within the glass slab even though it is free to expand from everywhere. The problem is that I am unable to apply boundry condition in which I can specify that the slab is free to expand from everywhere. I am trying to do it, but it is giving me pivotal error. Please let me know how to define such boundary condition.

 

Second thing is, is it possible to calculate what could be the heat transfer coefficient on the surfaces of the slab just by giving geometry and temperature ( and material also).

P.S. Heat is convected from each surface and slab if free to move in each direction. Also the problem is of transient thermal.

 

Thanks in advance! 

[peteroznewman]

You could take advantage of 2 planes of symmetry since the geometry and the loads are symmetric about the two vertical planes that pass through the center of the slab. Slice the geometry on those 2 planes and keep one of the four pieces to use in the analysis. Now you can either apply two Displacement BC, one for each face where the displacement normal to the face is 0 and the other two displacements are Free. That has eliminated 5 of the 6 rigid body motions for the part.

What is the slab of glass resting on?  Do you have Standard Earth Gravity on? 

If you don’t have a Gravity load, you can select any single vertex on the 1/4 model and add a third Displacement BC to make vertical displacement 0 leaving the other two free and that will solve the pivot error or you can add a 3rd plane of symmetry and slice the model in half through the thickness and add a third Displacement BC to make vertical displacement 0 leaving the other two free and scope that to the cut face.

If you have a Gravity load, you can’t do a single vertex as the entire weight of the part would go through that one node and you can’t use a third symmetry plane because the nodes in the center may want to move out of plane depending on the support on the bottom.  Instead, you can use a Remote Displacement, Behavior = Deformable, and set the vertical component to 0, leaving the other 5 rows Free.  Scope that to the bottom face.

Do you have temperature dependent conductivity for the glass?

Do you have temperature dependent Young’s modulus for the glass?

You can solve both the transient thermal temperature results and the stress results in a single Coupled Field Transient model.

Look for published values for the heat transfer coefficient which will be different for a vertical surface, a bottom surface and a top surface.

[ogoz]

You may  check narayanaswamy model for glass relaxation above Tg