| Author | Message | ||||
     Mohamed Qasem (m_alodat) New member Username: m_alodat Post Number: 1 Registered: 03-2009 |
Dear Nelson I am modelling the cooling of cylinderically shaped product. The cylinder product is initially bat dimensioless temperature equal 1 and is suddenly fast cooled to temperature zero. I write a flex program but I failed to get reseanable results. I am attaching both the physical problem with the Flex program
| ||||
| Mohamed Qasem (m_alodat) New member Username: m_alodat Post Number: 2 Registered: 03-2009 |
Dear Nelson The loaded file is not PDE script Please find the script that I use
| ||||
| Robert G. Nelson (rgnelson) Moderator Username: rgnelson Post Number: 1224 Registered: 06-2003 |
Your system is overspecified. You really have only one variable, Temp, with the equation: Temp: div(grad(temp)) +(B+A*temp) = dt(temp) The fluxes are precisely the derivatives of the temperature, and boundary fluxes are specified as Natural BC's of Temp. I don't understand your flux equations, anyway. They imply a time lag between temperature change and flux change. What is this weird material? If the funny flux equations came from your attempt to non-dimensionalize the equation, don't do it. Just state the problem in real coordinates. | ||||
| Mohamed Qasem (m_alodat) Junior Member Username: m_alodat Post Number: 3 Registered: 03-2009 |
Dear G. Nelson Thank you very much. You are true there is a time lag between the temperature change and flux change (hyperbolic heat conduction Equation)and I try to solve the system directly with one variable but this need the sedcond derivative of temp. with respect to time and an error massege from flex. Can flex handle an equation with dtt(temp). Also, what is the problem with nonnon-dimensionalizing the equation? I try to modifiy the Flex scripter and the modified program is attched, but still it gives non-reasonable results.
|
*.PDE
*.doc
*.Tex