>> I encountered a strange problem with the use of quadrature points. In a
>> function I interpolate values of a given variable onto the quadrature points.
>> The following code worked fine for elements of type TRI:
> 
> [code snipped]
> 
> Here is a sample output
>> 
>> -0.0101701<x<-0.010009
>> .0100975<y<0.0102585
>> Quadrature point 0: x=-0.00839908 ; y=0.00843638
>> Quadrature point 0 is in element!
>> Quadrature point 1: x=-0.00958875 ; y=0.00972795
>> Quadrature point 1 is in element!
>> Quadrature point 2: x=-0.00381006 ; y=0.00384751
>> Quadrature point 2 is in element!
>> Quadrature point 3: x=-0.00838955 ; y=0.00852933
>> Quadrature point 3 is in element!
>> 
>> As you can see, although the x-values for the element vary between -0.0101701
>> and -0.010009, the x-value of the first quadrature point is -0.00839908. I
>> guess that I do something wrong with the initialization of the finite element
>> or the quadrature rule, but I can not figure out what. The values of the
>> initialization variable are
>> dim=2
>> VAL_ORDER_ES=1
>> VAL_FAMILY=0
>> VAL_ORDER_GAUSS=3
> 
> 
> Hi there,
> 
> Are you confusing points in reference space (which is where the
> quadrature points are defined) and points in physical space (which is
> where the element is defined)?  If you want to know the physical space
> locations of your quadrature points, just catch the reference returned
> by
> 
> fe->get_xyz();
> 
> I only gave your code a cursory examination, if this is not the real
> issue I'm sorry.

Also, by default we are using a quadrature rule with nonnegative weights,
and that ends up with a 4-point rule.  the distribution thus "squishes" two
points toward a single corner and are not generally symmetric - is that what
is bothering you?  The Gaussian quadrature points points are located
*within* the element, and thus will be inside the bounds of the element...

-Ben


------------------------------------------------------------------------------
Come build with us! The BlackBerry&reg; Developer Conference in SF, CA
is the only developer event you need to attend this year. Jumpstart your
developing skills, take BlackBerry mobile applications to market and stay 
ahead of the curve. Join us from November 9&#45;12, 2009. Register now&#33;
http://p.sf.net/sfu/devconf
_______________________________________________
Libmesh-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/libmesh-users

Reply via email to