We are simulating fluid flows in rectangular 200 x 35 micrometer nanochannels using Fluent. We are looking at the best mesh size for this.
hello Mr. Ismail
Well the best mesh size can not be predicted. you will have to do the grid independent study and for that start with a grid cell size of 1 micrometer and then start reducing it till you find that there are no significant changes in the results.
Yes thats the best way to determine the mesh size which suggested by Brijesh.
for more details you may please read my last post .
https://www.researchgate.net/post/What_is_the_difference_between_Grid_Independent_Test_and_Mesh_Convergence_Study
Thanks for all your inputs. We are wondering whether there are some guidelines (rule of thumb) of the number of mesh per area (for each face/ cross sectional area). Thanks.
hey Ahmad...no there is no such rule of thumb..though there is a limit of computational power available for you. so if your system can support higher no. of elements/cross sectional area. you can go for that. but first do grid independent study.
There must be a theory/theorem that relates the min/max number of elements/mesh size relative to the surface area and/or body volume.
We are wondering what will be the best mesh size to get the fastest result without sacrificing the acccuracy of the simulation? If we are using more elements (i.e. more refined mesh), it will take more computer time.
In Fluent CFD software, we discover that the minimum required orthogonality quality for the mesh is 0.1; hence we can minimize/optimize the mesh element so that we approach this threshold. Operating at this condition will also safe our computational time.
If we start with too refined mesh, it will take very long time to converge. So, how to get reasonably accurate results quickly?
What is the best way to refine the mesh in Fluent? Change element size in Body Sizing? How about type of mesh?
There is also an option whether to use "coarse" or "fine" mesh. I wonder whether the results will be very diferent?
Best way is adaption.. You can adapt it to fine mesh where there is a large change in your gradients.
Results will differ in case of fine or course that's why we do grid independent study.
To get the animation from CFD simulation, we have tried "transient" conditions. How to optimize the computational time for transient mode?
That depends on for how many seconds you want to study for your case and then you can reduce time by using appropriate time step size and max iteration per time step.
Do we need to know how long it takes to attain steady state condition?
I suggest to you to do some calculus before you start to simulate. It is important to define the motion scale in order to understand kind of simulation. Sometime it is no necessary to do a complicated mesh in a laminar flow...What do you espect from your results?
We are trying find the optimum mesh size for the simulation; that is cost-effective and fast convergence?
i suggest to compute the mixing scale from density of fluids, velocities parameters (Kolmogorov scale). Maybe computing the Dean vortex scale if you are working with a laminar flow.
Hi Ahmad. To reduce the computational time and resources consumption after optimizing grid mesh, you can play with the relaxation parameters of the velocity and pressure in the code. Their values may affect the stability of the system. This can drastically affect the no. of iterations to reach the desired level of error residual. you just need to check whether by increasing or decreasing the relaxation parameter individually for velocity and pressure, it increases the rate of convergence. the final values of relaxation parameters can vary from problem to problem and there is no rule of thumb as to how to figure out optimum value or direction of variation of relaxation parameters to ensure stability and fast convergence. good luck.
I guess "relaxation" settings are another art in the CFD work. The question is how to optimize the parameters for the best results.
- Badges
- Science topic
- Similar topics
- Engineering
- Thermal Engineering
- Thermofluid
- Fluid Mechanics
- Fluid Dynamics