# Wrong result Heat Transfer in a Cavity

 Register Blogs Members List Search Today's Posts Mark Forums Read

 February 19, 2014, 21:34 Wrong result Heat Transfer in a Cavity #1 New Member   Abhi Join Date: Feb 2014 Posts: 8 Rep Power: 12 I am trying to simulate one of the cases given in following paper by Metzger Cavity Heat Transfer on a Transverse Grooved Wall in a Narrow Flow Channel. in Fluent Re= 15000. The geometry is a 2D section I am using Standard K-Omega Model I am trying to get the Nusselt number. I first specify Total temperature at Inlet i.e. 320K and T_Wall= 300. I got wrong results. Then I tried opposite way: T_Wall=320K and Total Temp_Inlet=300k. I got some results and tried to plot Nusselt. Attached are the Results. I defined the Nusselt number through Custom Defined function as Nu= (Surface Heat Transfer coeff * Cavity Depth)/ Thermal Conductivity. I used the Thermal conductivity of Fluid i.e. k=0.02 w/m-k. I am new to Heat transfer area so I am not sure what I am doing wrong. Cases converged there was no problem with the case setup. I checked the dimensions of the geometry that's right. Please help. Thanks

 February 20, 2014, 03:55 #2 Super Moderator     Alex Join Date: Jun 2012 Location: Germany Posts: 3,399 Rep Power: 46 Read the manual how Fluent computes heat transfer coefficients. It differs from the usual engineering definition of a heat transfer coefficient.

 February 20, 2014, 17:28 #3 New Member   Abhi Join Date: Feb 2014 Posts: 8 Rep Power: 12 I read the manual. But first of all I wanna know how the temperatures should be specified. Is the Wall Temperature higher or the Inlet Temperature. First of all I wanna make sure that my case setup is right.

 February 21, 2014, 03:49 #4 Super Moderator     Alex Join Date: Jun 2012 Location: Germany Posts: 3,399 Rep Power: 46 It doesnt matter. Both cases should yield the same result when set up correctly, provided you dont account for any temperature dependent material properties.

 February 21, 2014, 10:52 #5 New Member   Abhi Join Date: Feb 2014 Posts: 8 Rep Power: 12 Sorry, did'nt look at the experimental data properly. It was given in the form of string distance. I was getting the right results all the way. But the position of the inlet does affect the Nu distribution and the exit conditions. I also a question about the stability of the discretization schemes. Is it OK to use Third Order MUSCL scheme or Second Order Upwind. I read that second order upwind is stable and reduces numerical dissipation. Thanks for the help

 February 21, 2014, 11:31 #6 Super Moderator     Alex Join Date: Jun 2012 Location: Germany Posts: 3,399 Rep Power: 46 The general rule of thumb for the interpolation schemes for the convective terms is: with higher order schemes you trade in stability for accuracy. To answer your question: The second order upwind scheme is more stable compared to a third order MUSCL scheme and reduces numerical diffusion compared to the first order upwind scheme. Always use the scheme with the highest order that gives stable results.

 Tags cavity flow, fluent 2d, heat transfer, nusselt number