Discussion Forum

Hello,

I am trying to model forced convection that occurs when water flows through a pipe. To do this, I am setting up a simple 1D simulation. I don't believe I can do this using any of Comsol's specific physics modes. Therefore, I have set up this simulation using the Coefficient Form of the PDE Mode. I have made the dependent variable F, which is the fractional temperature change ((water temperature-initial subdomain temperature)/initial subdomain temperature). I have assigned coefficients in the subdomain settings so that I can determine how F changes as a function of time and space, dependent on a convection term and a reaction term (defining a [absorption coefficient], da [damping/mass coefficient], and beta [convection coefficient]. I have also defined c (diffusion coefficient) as a small number (0.01) so that I can set up the outlet boundary condition as a no flow boundary. I have defined all of my constants (including defined velocity so I don't have to couple this to a flow model), and I set the initial subdomain to F = 0 ((10 C-10 C)/10 C). I set F at the inlet to 2 ((30 C-10 C)/10 C), and I have set up the outlet boundary condition to a no flow boundary. When I run the simulation water at F~1.9-2 propagates down the pipe. However, there is some scatter that also propagates down the pipe, with values of F ranging from ~1.5-~2.5. This scatter starts at the inlet (location 0 in 1D space) and is centered at ~400 at time step 1000. This occurs when the relative tolerance is set to 0.01 and the absolute tolerance is set to 0.001 in the General tab of the Solver Parameters. If I set the relative tolerance to 0.0001 and the absolute tolerance to 0.00001, then the scatter is minimal but is still there (up to 0.01 C). Does anyone know the reason for this scatter? Is this a problem with my inlet boundary condition that then propagates down the 1D space? I am attaching my Comsol Multiphysics Model file to this post so you can see what I mean by this scatter.

Thank you,
Andrew