Discussion Forum

Hi

well I haven't seen that one ;) but if you analyse the sentence you should look at the way you define the stress, could it be that you are entering a complex value when a only real was expected, i.e. you are missing an abs() (+ sign) or a real() or perhaps even an imag() for a variable transfer ?

What are the values of this complex value ? is the complex part negligible ? ...

From thereon you should be able to decide how to continue

--
Good luck
Ivar

I'm having the same problem. I'm using version 3.0a and trying to model the effects of thermal expansion of a wire (elasto-plastic material) inside a ceramic material. I keep getting the following error when its finished:

"Evaluation of post processing expression failed. Complex stress function".

The details on my model follow..

====================================
====================================
Model:
Axial Symmetry (2D) -> Structural Mechanics Module -> Static Analysis Elasto-Plastic Material model

Constants:
Th=1900+273
Tl=298
Rm=200e-6
Ra=2e-2
ht=1e-2
Em=314.5e9
sy0=414e6
Etm=68.9e9

Scalar Expressions:
currT=Tl+(Th-Tl)*(param*(param<=1)+(2-param)*(param>1))

Draw:
Rectangle 1:
corner at origin; width=Rm, height=ht/2
Rectangle 2:
corner at origin; width=Ra, height=ht/2


Subdomain Settings:

(1 - Inner Material):
elastoplastic material
E=Em
nu=0.380
alpha=5.35e-6
rho=10220
Hardening model: isotropic
yield fn: von Mises
sigma_ys=sy0
E_T_iso=Etm

(2- Main material):
isotropic
E=315e9
nu=.239
alpha=7.8e-6
rho=3690

(1&2):
Load->Include thermal expansion is checked
Tempref=Tl
Temp=currT

Boundary Settings:
(2&5 - 'bottom') Rz=0 (symmetry)

Mesh:
Initialized then refined twice (~8000 elements)

Solve:
Static elastoplastic -> Parametric Nonlinear
param = 0:0.01:2
(ramps temperature from room Temp to 1900°C then back down to room temperature).

Nothing else was changed from the default settings.

=====================================
=====================================

It's the combination of elasto-plastic material and thermal expansion that seems to be failing.

The following simpler model receives the same "Complex stress function" error

=============

Axial Symmetry (2D) -> Structural Mechanics Module -> Static Analysis Elasto-Plastic Material model

Draw a 1x1 square

subdomain:
Use the default material, select elasto-plastic, keep default elastoplastic settings (isostatic w/ tangent modulus).
check "include thermal expansion."
Tempref=298
Temp=298+1000*param

no boundary conditions imposed

solver: static elastoplastic -> nonlinear parameter

param=0:0.2:1

=================

It's the default von Mises yield function (at least in my case) that is causing the error. Using tresca_axi or s1_axi or real(mises_axi) as the yield function runs through successfully (at least for the simple unrestrained geometry).

mises_axi=sqrt(sx^2+sy^2-sx*sy-sy*sz-sx*sz+3*sxy^2+3*syz^2+3*sxz^2+4*Sys^2*eps)
tresca_axi=max(max(abs(s1-s2),abs(s2-s3)),abs(s1-s3))

Your equation for mises_axi is missing an sz^2 term. That’s why it get negative values.

Nagi Elabbasi
Veryst Engineering

Your equation for mises_axi is missing an sz^2 term. That’s why it get negative values.

Nagi Elabbasi
Veryst Engineering

I apparently forgot it when typing it up above. The mises_axi is a preset variable (so I didn't actually define it within comsol..I was just manually copying it into the forum post from a copy from comsol's variable definition dialog box).