I am doing a numerical study of a new model of a heat exchanger and there are no previous studies on the same geometry that the study was conducted on. Is there a way other than conducting experiments to ensure the validity of the numerical study?
Dear Ahmad Aboulkhail ,
Please mention that the computational method is the one to solve problems approximately. So, the results are not exact. To make the solutions more accurate, at least we have to check the following:
1) The solution method is right
2) The models involved in the computation are suitable and appropriate
3) The solution is not depend on the grid number or size
4) The solution is not depend on the number of iteration
5) the solution is not depend on the convergence criteria
5) The results are close enough to any valid results, such as experiment or analytical results.
Have a nice job....
If there is no experimental data there check some other similar date that closely resembles that model. this is one way. another method is the mathematical modeling and than verify the modeling in MATLAB
First, if your problem differs only for the geometry of the model from the others, try first if you are able to use your CFD code and your setting (physical model, BCs., etc) for obtaining those well assessed results.
Then, your problem may need to be analysed properly for the suitable grid generation. Depending on the use of the formulation and turbulence model you can work using grids of different resolutions.
These are only general guidelines but there are much more things to check. In such cases, experience in CFD is required.
Hello Ahmad Aboulkhail ,
I'm also dedicated to simulating a wide variety of heat exchangers. Based on the experience of my current work, You need to define or be more specific on what kind of simulations are performed, I mean: Are you analyzing the thermal performance, the pressure drop, etc.?. Once, the study is defined, there are several ways to validated your results, The most trustworthy way to do this, is to carry out a physical test for the same geometry.
However, You must ensure that your numerical method and simulation setup is ok, I mean, in terms of boundary conditions, coherent physics to analyze, grid dependence (Mesh convergence), etc, as Filippo Maria Denaro and Nazaruddin Sinaga indicated, as you may know.
I highly recommend to validate your setup against a physical test.
Please let me know your feedback.
Thanks to all of the colleagues who have reported me. As for the analysis I do, thermal performance analysis (pressure drop - heat transfer) within a new geometry of heat exchanger. The numerical study is about to be completed. The independence of the network was studied by increasing the size of the network by 40%, during which both the pressure drop and the heat transfer coefficient 2% and 1.7% were changed respectively (is this acceptable or should it be taken at least three cases ?!). Can the results be validated by applying the basic laws on heat transfer. (Q = M.CP.DT = U.A.LMTD).
Thank you in advance
Oscar Alcocer Filippo Maria Denaro Naseem Ahmad Nazaruddin Sinaga
Dear Ahmad Aboulkhail ,
Could you tell me the meaning of network you mean, please? Did you mean it was the size of the geometri, for instance, you enlarge it 40%?
Hi dear Prof. Filippo Maria Denaro ,
Nice to see you again. I am looking for you in our other discussion. Hope you always fine and happy.
Best wishes.
Dear Nazaruddin Sinaga
I'm fine, thanks for your interest.
Best regards
Sorry I missed typing, I mean the network is the grid (Mesh)@ Nazaruddin Sinaga
- Badges
- Science topic