[CIG-SHORT] Issue with Normal Orientation

[Brad Aagaard]

On 02/18/2015 07:37 AM, Romain Jolivet wrote:
>
> On 18 Feb 2015, at 15:30, Matthew Knepley <knepley at mcs.anl.gov> wrote:
>
>> On Wed, Feb 18, 2015 at 9:21 AM, Romain Jolivet <jolivetinsar at gmail.com> wrote:
>> Hello,
>>
>> I am having some more problems with the orientation of the vectors normal to the fault.
>>
>> My problem is in 2D, with a fault slicing the entire medium (see tar ball attached).
>> The fault is FaultCohesiveKin class. I enforce the slip.
>>
>> When I run with 1, 2, 3, 5, 7 or 9 nodes, I have one orientation for the vectors normal to the fault.
>> When I run with 4, 6 or 8 nodes, I have the opposite orientation.
>> I haven’t tried with more nodes.
>>
>> That should not happen, right?
>>
>> This can happen. We have to choose some orientation for the fault, and different partitions
>> produce different starting points, and thus the fault can have the opposite normal.
>>
>>    Thanks,
>>
>>       Matt
>
> Hmmm, that is not very convenient…
> Isn’t it possible to add a constraint on the normal vector for that starting point?
>
> Something like:
> If it points down, take the opposite,
> If it is horizontal and points to the right, make it point to the left.

Romain,

I know I originally setup the fault slip orientation information so that 
it didn't matter which direction the normal pointed in a 2-D simulation. 
Positive tangential slip was left lateral and negative tangential slip 
was right lateral. I can see from your example that this is not 
happening now. I will fix this.

Brad