[CIG-LONG] Low viscosity material around dike

[Taichi SATO]

Dear Garret and Walter,

First, I tried to simulate M=0.7 spreading ridge model.
I specified half-spreading rate of 3.17e12(=10mm/yr) and dike width of 
500m. Mesh interval is 250m*500m.
I set the divergence force value -8.88e9(= 2*3.17e12/500*0.7). As 
Garrett suggested, I also found to get a divergence rate closer to what 
I wanted I found I had to reduce it by about 50% (so the current value 
of -4.44e-9).
I changed dike upper boundary of 16000 m to 15500~14000 m. I found that 
at the beginning of  simulation, axis is sinking down, but the 
simulation goes on, the small dike hill emerges from the axis.

I agree with Garrett's suggestion about the setting of ShapeVC, but the 
strain at the right side of the dike looks like conjugate fault in pair 
with the initial large fault.

I also tried to run non-newtoinan viscosity model but it does not work. 
but it does not work. Please see attached file.
the error is,
Error in NonNewtonian: T, n, and A must all be non-zero:
     T=0
     n=3
     A=1e+10

I set the parameters for non-newtonian and mantle viscous as below
<struct name="nonnewViscosity"> -->
<param name="Type">NonNewtonian</param> -->
<param name="StrainRateInvariantField">StrainRateInvariantField</param> -->
<param name="TemperatureField">TemperatureField</param> -->
<param name="n">3.0</param> --> <!--power law exponent-->
<param name="T_0">1473</param> --> <!--T_0=Q/(n*R)=20045.37 -->
<param name="A">1.0e10</param> --> <!--power law constant-->
<param name="refStrainRate">1.0e22</param> -->
<param name="minViscosity">1.0e19</param> -->
<param name="maxViscosity">1.0e25</param> -->
</struct> -->

<struct name="mantleViscous">
<param name="Type">RheologyMaterial</param>
<param name="Shape">mantleShape</param>
<param name="density">2300</param> <!--3300-1030 kg/m^3-->
<param name="alpha">3.0e-5</param>
<param name="diffusivity">1.0e-6</param>
<list name="heatingElements">
<struct>
<param name="Q">8.48484848485e-13</param>
<param name="lambda">0.0</param>
</struct>
</list>
<list name="Rheology">
<param>nonnewViscosity</param>
<param>yielding</param> -->
<param>storeViscosity</param>
<param>storeStress</param>
</list>
</struct>

Sincerely,

Taichi Sato

(2010/04/30 5:51), Walter Landry wrote:
> Garrett Ito<gito at hawaii.edu>  wrote:
>    
>> Hi Taichi,
>>
>>    I believe the dike region is staying elevated relative to the material on
>> the sides because vy and vz are constrained to be zero in the "dikeShape".  I
>> did this in order to cause the divergence to only cause the dike to widen
>> sideways (i.e., vx~=0).  Ideally I think you want the dike to be allowed to
>> move up or down with the surrounding material but I dont remember figuring out
>> a way to allow it to do so.  I think this would require setting dvy/dy=0 but
>> not specifying any constraints on vy.
>>
>> Walter, can you see a way to do this?
>>      
> A divergence zone, by definition, creates material out of thin air.
> If material should not be created high up, then there should not be a
> divergence zone there.  It might be better to bury the dike.  You
> would probably not want to specify vy.  Then material could flow up as
> well.
>
> Cheers,
> Walter Landry
> walter at geodynamics.org
> _______________________________________________
> CIG-LONG mailing list
> CIG-LONG at geodynamics.org
> http://geodynamics.org/cgi-bin/mailman/listinfo/cig-long
>
>    

-------------- next part --------------
A non-text attachment was scrubbed...
Name: magma-tecto_new1_nonnewtoian_taichi.xml
Type: text/xml
Size: 36153 bytes
Desc: not available
Url : http://geodynamics.org/pipermail/cig-long/attachments/20100501/31081ddf/attachment-0001.bin