# yade-users team mailing list archive

## Re: How to keep mean stress constant

```Hi Sayeed,

```
If s1, s2 and s3 are independently not constant, there is an infinity of values of s1, s2 and s3 leading to same mean pressure (e.g. 100 kPa). Thus you can perform such a control only if you impose additional conditions, for instance other relations to be verified between s1, s2 and s3 (for instance s1 + 5*s3 = 3*s2). But here the problem has nothing to do with Yade. You should precise your physical problem before trying to do something with Yade.
```
Best

Luc

```
```Dear Mr.  Vaclav, Luc and all, thanks for your good co-operation.

```
I want to keep mean stress constant where s1, s2 and s3 are independently not constant but there mean value (*s1+s2+s3)/3 = constant e.g . 100 Kpa*) remains constant though out the triaxial compression test. Actually I am in learning phase of YADE, so I hope your warm assistance to modify the engine if need.
```
I am waiting for your better suggestion.

Thanks all

Sayeed, Saitama University.

--- On *Thu, 10/29/09, Luc SIBILLE /<Luc.Sibille@xxxxxxxxxxxxxx>/* wrote:

You can use the engine ThreeDtriaxialEngine, with this engine you can
choose to keep constant separatly S1, S2 and S3 at different values,
like
that the mean stress will stay constant. Is it enough for you, or do you
want to keep the mean stress constant while S1, S2 and S3 are
independently not constant?

Best,

Luc

>
>> Now I want to
>> do a simulation that I need to keep a mean stress ((S1+S2+S3)/3)
>> constant, and
>> Did anybody handle this before or how is it handled in YADE? can you
>> give
>> any suggestion?
>> If YADE has already includes this kind of algorithm, can you tell me
>> where is it?
>
> Hi Sayeed,
>
> to me it is not clear what are you trying to achieve. Can you be more
> precise?
>
> In particle models, stress tensor is not defined, only forces on
> interactions between particles; you can approximate stresses by
looking
> at interactions around each particle (which I did just recently for
> postprocessing purposes in CpmStateUpdater::update, but only for the
> concrete model).
>
> If you mean something like triaxial test where mean stress in the
whole
> specimen is constant, then you should have a look at
> TriaxialStressController (that gives you ways to control individual
> stress components) and derive your own engine from it.
>
> For periodic compression (simpler, since no walls and other funny
> things, but less mature code otherwise), have a look at the
> PeriIsoCompressor class; you would have to write something very
similar,
> just changing the lines that compute displacements to keep mean
stress
> constant.
>
> For those, you have to touch c++, though -- unless Bruno has some
other
> idea.
>
> Cheers, Vaclav
>
>
> _______________________________________________
>

_______________________________________________

------------------------------------------------------------------------

_______________________________________________
```
```
--
Luc Sibille

Université de Nantes - Laboratoire GeM UMR CNRS

IUT de Saint Nazaire
58, rue Michel-Ange - BP 420
44606 Saint-Nazaire Cedex, France

Tel: +33 (0)2 40 17 81 78

```