YYT:

You can use thermal stress to impose an initial stress.  However, your e-mail
implied that you wanted to apply separate discrete initial stress values on two
different materials.  It will be difficult to achieve this type of initial
stress distribution using the thermal stress approach.  Also, the thermal stress
approach can be difficult to implement, if you have temperature dependent
material properties.

To create an initial thermal stress you just need to run a stress analysis with
thermal loading in the first load step and any additional analyses in later load
steps.  Thermal stress is created by uneven thermal growth [(T-Tref)*CTE].  To
create uneven thermal growth, you can either specify an uniform temperature for
the entire model and different coefficients of thermal expansion for different
areas or you can specify the same CTE for the entire model and different
temperatures.  Either approach creates uneven thermal growth, and thus,
thermally induced stresses.

Regards.

Dan Shaw
ANSYS

-----Original Message-----
From: mems-talk-bounces@memsnet.org [mailto:mems-talk-bounces@memsnet.org]On
Behalf Of yyt@mail.nwpu.edu.cn
Sent: Thursday, January 11, 2007 9:15 PM
To: mems-talk@memsnet.org
Subject: RE: [mems-talk] About initial stress in ANSYS (Yiting Yu)


Daniel

Then how about introducing initial stress by the temperature change, now we call
it as an equivalent thermal stress method? Can you explain the principal
operating
steps? What's the difference between this method and by ISTRESS command?