i fkn hit hike before even watching, ths should say how much we appreciate your admirable efforts, Aiden , you have all my gratitudes and support. i wish you all the best in this life now lets go back watching and taking notes
Holy crap. My dissertation is on CFD FSUK aerodynamics, and i wish i found this channel sooner. Been tryint to mesh a front and rear wing on the same geometry but it refused to work. I now understand why. I'll update this if i actually managed to fix it. Thank you for these
Masterful course, Aidan! Plenty of teachings and I gained a different insight on meshing (not only does it allow a gain in accuracy but also in stability)
Nice video! If you included examples of what good looks like (desired aspect ratios for varying mesh problems) for the SS and references for that it might enhance the experience.
This video is so great to explain the implication of aspect ratios in CFD!!! (I usually just care about skewness and non-orthogonality more.) I'm thinking in high aspect ratio mesh, if the long-side of the cell is parallel to the flow in transient flow problem, it can mean Co number is still controllable. Does it also mean to have a good pressure correction to avoid excessive iterations? I'm struggling about the mathematical relationship between aspect ratio and pressure correction. I ask this question it's because I'm thinking in what situation CFD allows to have high aspect ratio. Thanks a lot!!
You will probably be ok, because the pressure field from the previous time step becomes the initial guess for the next time step, so I don't think aspect ratio will have as much of a detriment as the steady solver where you start from nothing!
Thank you for the amazing video! I run the simple case of the flow around a circular cylinder in OpenFOAM using PimpleFOAM, and I realized that when I use structured mesh with higher aspect ratios the number of the outer correctors to converge is much higher than a low aspect ratio case. I can imagine, after watching this video that this is because of the diffusion dominant terms in the pressure equation. My question is, when my Courant number is low(around 0.5), and I use high aspect ratio cells, is it okay to perform just one outer corrector, or should I allow the solver to decide for this (since it needs more to converge)?
You are absolutely correct. High aspect ratio cells needs more outer corrections, due to the diffusion term in the pressure equation. As for the courant number question, I would check your residuals (and monitors) within each time step. If you can convince yourself you are sufficiently converged within a time step, then 1 outer corrector is fine. An easy way to check this is to run 2 simulations, one with 1 outer corrector and another with 2 outer correctors. The compare your drag and lift coefficients. If there is no difference then you are fine to use 1 outer corrector
What about airfoil boundary layer cells in steady state simulations? I guess the pressure correction in not much a problem close to the wall because it's mostly diffused over the large face? And what about the wake?
Yes correct. The pressure is mostly diffused normal to the wall, but the streamwise contributions are important, especially on the suction surface, so the high aspect ratios are always going to slow things down.
@@fluidmechanics101 And doesn't the transient case also include a pressure correction step? What would be your best approach to calculate stall polars on low Mach number airfoils (M
I am not sure what you mean by 'stall polars'. If you want to find where the aerofoil is stalling, you can just run a set of steady state simulations, incrementing the angle of attack by 0.5 degrees each time (changing the inlet velocity components). You can each initialize each solution with the solution from the previous angle of attack, so you won't need many iterations (say 500 per angle of attack). You only really need dynamic or moving mesh if you have a transient inflow (like a pitching aerofoil, or a vertical axis wind turbine)
@@fluidmechanics101 yes that's what I currently do. I was wondering if the transient case would suffer less from the pressure correction in high aspect cells, in your video I get the feeling the transient case is less sensitive to it, but I don't understand why.
Unfortunately there are some questions even I cannot answer 😅 the video shows my current understanding. If I find out anymore I will be sure to let you know!
Hi there it really gave a clear picture, I am working on an aerofoil post-stall aoa so i am running in transient condition, for K omega sst iIam keeping wall y+ 1 for a RE no. 2.2E6 in that first cell height I am getting around 1.1e -5m near the aaerofoil. For this first cell height I need to keep the time step size less than 1e-5 s to keep courant no. less tha 1 at all the sides of an aerofoil but unfortuanetly i don't have computer resources to deal with this time step size. If i go with larger time steps Cl and flow seperation is completley off but Cd is matching within 10% error. If I use Spalart-Allmaras turbulence model, i went with wall y+ 100 the first cell height is 1e-3m so time step size was maintained at 1e-4s in this Cl and flow seperation is matcing even at post stall but Cd is at error of 70%. Kindly advise me on how to proceed further. Thank you.
If you are limited by computer resources then you can just do the best you can and make it really clear to the assessors / whoever is reviewing your work that that is the best you can do. If they want a more accurate result, then they need to get you access to a better computer!
It mostly affects the rate of convergence of the pressure equation, due to the coefficients in the A matrix (which are calculated from the discretised Laplacian)
![[CFD] Aspect Ratio Warnings in CFD](/img/logo.svg){kind=logo}
