Daven-

I was not aware that tools allow for thermo-viscoelastic damping analysis.
Please elaborate.

So while I am not a user of ANSYS, but of ABAQUS, my understanding of FEM tools
for MEMS and large-aspect-ratio elements would lead me to suggest selecting one
of the following options:

        * Use six-sided bricks (quadrilateral surfaces) for meshing.  Consider
using either a structured mesh, or an extruded quadrilateral mesh.  You may need
to simplify your meshed geometry to approximate curved surfaces with 90 degrees
corners, etc.

        * Separate your thermal analysis from your piezoelectric analysis from
your damping analysis, or some fraction thereof except where coupling is
essential.  If you absolutely must have all three, then you will likely deal
with convergence problems and more issues with mesh compatability for all three.

        * For your thermal analysis (only) you may be able to use a much thicker
layer of gold (assuming you have large lateral distances compared to thickness)
by reducing gold's bulk thermal conductivity proportionately.  This will give
you the profile of the thermal gradients within the multilayer structure.  You
should not have to change the natural convective surface coefficients.

--
Raj Gupta, Owner
Volant Technologies, Inc.
San Francisco, CA
http://terahz.org


----- Original Message ----
From: daven chou 
To: [email protected]
Sent: Wednesday, November 26, 2008 8:53:16 AM
Subject: [mems-talk] problems about the thermo-piezoelectric ansys of quartz
beam

Dear everyone:

I am using the finite element analysis to calculate the quality factor
of quart resonating beam accerometer. But I've got a puzzle: how to take
piezoelectricity and electrodes into account.  The thermal conductivity
of gold is larger than that of quartz. so,the piezoelectricity and
thermal conduct of the electrodes will affect the temperature gradient
in the beam.

I used the ansys as the FEA tools to do this job. The beam is
quartz,and solid226 is used to mesh the beam. The electrodes deposited
on the beam is about 3000 A, which is beyond the the thickness of the
beam. If i use the 3-D solid to model the electrodes, the elements are
so large that the computer can't resolve it. I haved tried to use
plane223 to model the electrodes,but i don't know how to deal the link
between electrodes and the surface of the beam.

Is there anyone who has tactics to solve the problem. I am looking
forward to your answer. thanks a lot in advance. my english is poor,
and i don't know whether i have described the problem clearly.

                            sincerely yours,
                            jianfa zhou


ps. the error code of ansys hint

*** ERROR ***                           CP =      53.719   TIME= 14:46:01
There are elements with negative electric charge and current conduction
reaction solutions for the VOLT degree of freedom in the model.  This
analysis is not valid.

ps.brief codes of ansys

/prep7
.....
et,1,226,1011
et,2,223,110,
mp
........  !define meterial properties

!creat models
block,.....!model the beam
k,,...
a,.....  !model the electrodes
aovl,....!overlap the electrodes and the surface of the beam

!mesh the models
type,1
mat,1
.......!mesh the beam
.......

type,2
mat,2
........!mesh the electrodes
........

!apply boundary conditions
..........!clamped-clamped beam
..........!apply voltage conditions

fini

!solve the problems
......

.......