Contents     Functions         Previous Next     PDF Index

p_heat

Purpose

Formulation and material support for the heat equation

Syntax

il = p_heat('default') 

Description

This help starts by describing the main commands : p_heat Database and Dbval. Supported p_heat subtypes and their formats are then described.

[Database,Dbval] ...

Element properties are described by the row of an element property matrix or a data structure with an .il field containing this row (see section 7.4). Element property functions such as p_solid support graphical editing of properties and a database of standard properties.

p_heat database

 il=p_heat('database');

Accepted commands for the database are

For fixed values, use p_heat('info').

Example of database property construction

  il=p_heat([100 fe_mat('p_heat','SI',1) 0 -3 3],...
             'dbval 101 d3 -3 2');

Heat equation element properties

Element properties are described by the row of an element property matrix or a data structure with an .il field containing this row (see section 7.4). Element property functions such as p_beam support graphical editing of properties and a database of standard properties.

For a tutorial on material/element property handling see section 4.2.1. For a programmers reference on formats used to describe element properties see section 7.4.

1 : Volume element for heat diffusion (dimension DIM)

  [ProId fe_mat('p_heat','SI',1) CoordM Integ DIM]
ProIDelement property identification number
typeidentifier obtained with fe_mat('p_beam','SI',1)
Integis rule number in integrules
DIMis problem dimension 2 or 3 D

2 : Surface element for heat exchange (dimension DIM-1)

   [ProId fe_mat('p_heat','SI',2) CoordM Integ DIM] 
ProIDelement property identification number
typeidentifier obtained with fe_mat('p_beam','SI',2)
Integis rule number in integrules
DIMis problem dimension 2 or 3 D

SetFace

This command can be used to define a surface exchange and optionally associated load. Surface exchange elements add a stiffness term to the stiffness matrix related to the exchange coefficient alpha defined in corresponding material property. One then should add a load corresponding to the exchange with the source temperature at T0 through a convection coefficient alpha which is alpha.T_0. If not defined, the exchange is done with source at temperature equal to 0.

model=p_heat('SetFace',model,SelElt,pl,il);

Command option -load T can be used to defined associated load, for exchange with fuild at temperature T. Note that if you modify alpha in surface exchange material property you have to update the load.

Following example defines a simple cube that exchanges with thermal source at 55 deg on the bottom face.

 
model=femesh('TestHexa8'); % Build simple cube model
model.pl=m_heat('dbval 100 steel'); % define steel heat diffusion parameter
model.il=p_heat('dbval 111 d3 -3 1'); % volume heat diffusion (1)
model=p_heat('SetFace-load55',... % exchange at 55 deg
    model,...
    'SelFace & InNode{z==0}',... % on the bottom face
    100,... % keep same matid for exchange coef
    p_heat('dbval 1111 d3 -3 2')); % define 3d, integ-3, for surface exchange (2)
cf=feplot(model); fecom colordatapro
def=fe_simul('Static',model); % compute static thermal state
mean(def.def)

2D validation

Consider a bi-dimensional annular thick domain Ω with radii re=1 and ri=0.5. The data are specified on the internal circle Γi ans on the external circle Γe. The solid is made of homogeneous isotropic material, and its conductivity tensor thus reduces to a constant k. The steady state temperature distribution is then given by

   − k Δθ(x,y) = f(x,y)    in    Ω.     (8.15)

The solid is subject to the following boundary conditions

In above expressions, f is an internal heat source, θext an external temperature at r=re, and g a function. All the variables depend on the variable x and y.

The OpenFEM model for this example can be found in ofdemos('AnnularHeat'). Numerical application : assuming k=1, f=0, α=1e−10, θext(x,y) = exp(x) cos(y) and g(x,y)= −exp(x)/ri ( cos(y) x − sin(y) x ), the solution of the problem is given by

 θ(x,y) = exp(x) cos(y)

See also

Section 4.2.1, section 7.4, fe_mat


©1991-2011 by SDTools
Previous Up Next