[Getdp] User defined source

Kubicek Bernhard Bernhard.Kubicek at arsenal.ac.at
Tue Oct 17 10:42:52 CEST 2006


Hi Amit,

maybe I can help you there in some way. We have the problem that we need to import a three dimensional scalar field into GetDP, to do further calculations with it in an formulation (i.E. some for_v). Considering that one already has the temperature nodal values in the sorting that the mesh in GetDP uses, one can do the following:
Using the additions at the end of this mail to the normal formulation "for_v", one will find a hell lot of values 12345678 in the .pre file after an initial sole pre-processing step.
One moves the file to a new filename .ori.pre
Now, using a self programmed tool, one can replace the "funny" 12345678 values with the corresponding temperature values, as they have the same sorting as the mesh nodes, and recreates a .pre file. Then one invokes the calculation and dances happily around his computer.
I understand that this can be considered as ugly as ugly can be, however it works..
Otherwise, there exist an undocumented function ReadSolution (to be found in the source-code), where one would need to supply a self-created solution file. 

In your case, I would propose you create a mesh file without gmesh, directly from you hexahedral mesh (being very careful about the volume orientation, and the nodal numbering) using self programmed software. 

Although this methods require a lot of hand-work, and although GetDP has a quite steep learning curve, I still recommend using it. We are very happy about it, and have experienced calculation times about 70 times shorter than a similar calculation using FEMLab (transient 2D simulation o electron convection/dissipation within electric fields)...

Nice greetings from Vienna,
Bernhard




Constraint {

	{ Name Constr_T ; 
		Case {
			{ Region Vol; Value 12345678;}
		}
	}
}

FunctionSpace {

	{ Name fsT ; Type Form0 ; 
		BasisFunction {
			{ Name sn ; NameOfCoef vn ; Function BF_Node ;
				Support Vol ; Entity NodesOf[ All ] ; }
		}
		Constraint {
			{ NameOfCoef vn; EntityType NodesOf ; NameOfConstraint Constr_T; }
		}
	}
}

Formulation {
	{ Name dummyT; Type FemEquation ;
		Quantity { 
			{ Name T ; Type Local ; NameOfSpace fsT ; }
		}
		Equation {
			Galerkin { [ Dof{ T} , {T} ] ; In Vol ; Jacobian Vol ; Integration Int ; } //Never Executed
		}
	}
	{ Name for_v; Type FemEquation ;
		Quantity { 
			{ Name T ; Type Local ; NameOfSpace fsT ; }
			...
		}
		Equation {
			... //use {T}
		}
	}
}

Resolution 
{
	{	Name all;System 
		{
				{ Name T ; NameOfFormulation dummyT ; DestinationSystem A1;}
				{ Name A1 ; NameOfFormulation for_v ; }
		}
		Operation 
		{ 
			Generate[A1] ; Solve[A1] ; SaveSolution[A1] ;
		}
	}
}

-----Ursprüngliche Nachricht-----
Von: getdp-bounces at geuz.org [mailto:getdp-bounces at geuz.org] Im Auftrag von Amit.Itagi at seagate.com
Gesendet: Dienstag, 17. Oktober 2006 00:02
An: getdp at geuz.org
Betreff: [Getdp] User defined source



Hi,

I am trying to solve the heat equation in a 3D geometry. My thermal source is a result of optical power dissipation in a lossy material. This source term is calculated on a Cartesian grid using a different program. How can an import this source into my getdp problem ?

Thanks

Amit

_______________________________________________
getdp mailing list
getdp at geuz.org
http://www.geuz.org/mailman/listinfo/getdp




More information about the getdp mailing list