double diffusive convection

Log Out | Topics | Search Moderators | Register | Edit Profile

Author Message   Eduardo Howard (eduardolaplata) New member Username: eduardolaplata Post Number: 1 Registered: 07-2005 Posted on Friday, July 22, 2005 - 07:17 pm:    Hello, I am trying to model a system with double-diffusive instabilities, like salt fingers in ocean water, with Temperature and Salinity gradients. Is there a working example/guide or something solved to follow as a reference about a FlexPDE implementation of this problem? Thanks in advance, Eduardo   Robert G. Nelson (rgnelson) Moderator Username: rgnelson Post Number: 412 Registered: 06-2003 Posted on Friday, July 22, 2005 - 08:51 pm:    We have no example of this in-house, and I am not aware of one elsewhere. Sorry. The example "Samples | Time_Dependent | Chemistry | Melting.pde" may have a few similarities to this problem that would be useful. The moving-mesh capability may also be applicable in this case, but I'm just guessing.   Benjamin Kaplan (kaplan) Member Username: kaplan Post Number: 5 Registered: 08-2003 Posted on Saturday, July 23, 2005 - 12:06 pm:    Dear RG Nelson, I use FLEXpde 5.0.2 and I would like to calculate the integral NORMAL FLUX at a curve surface (2D problem). I actually (cf. attached file 'hemisphere.pde') tried different possibilities that don't work (many functions exists : line_integral, surf_integral... but which one should be used ?...). In the actual file I only "calculate" the flux on the 'hemisphere' surface with different method summarized in the SUMMARY, which give obviously different results and are probably false. Can you help me to find the adequat function of FLEXpde to solve with an accurate precision this problem. Best Regards. Benjamin KAPLAN hemisphere pde file hemisphere.pde (1.7 k)   Robert G. Nelson (rgnelson) Moderator Username: rgnelson Post Number: 413 Registered: 06-2003 Posted on Saturday, July 23, 2005 - 05:47 pm:    As the FlexPDE documentation points out, the Surface_Integral integrates over the surface area of the specified boundary. In the case of 2D cylindrical geometry, this means multiplying by 2*pi*r. Line_Integral is a line integral, and does not contain the 2*pi*r weighting. In your "comparison" integral, you have put the 2*pi*r OUTSIDE the integral, instead of INSIDE the integral, where it belongs. If you conform to the rules of calculus and put the R inside, you will get the same answer for the "comparison" integral as FlexPDE produces for the Surface_Integral.