Pointwise Pointers: C-Grid by Extrusion

In this video I describe the process for generating a c-grid topology for the NACA 0012 airfoil geometry by extrusion. A normal hyperbolic extrusion coupled with a translational extrusion in the wake were used to define the c-type topology. The benefit of this approach is that by using the hyperbolic extrusion utility in Pointwise, no post smoothing operations using the elliptic solver are necessary as the grid is highly orthogonal and normal to the surface.
Y+ Calculator: www.pointwise.c...
AirfoilGenerator Glyph Script: github.com/poi...
SU2 NACA 2412 Case: github.com/tra...

[@firdausmohamad9605]

Hello Travis, Thank you for this video. I have a difficulty in preparing a C grid using extrusion where i need the wake is started exactly at the trailing edge of the airfoil. I want to prepare the C-grid like your previous video (C-grid tutorial) but using extrusion. I tried to use method that is the same like your previous tutorial, but my code give the error due to the high aspect ratio. Thank You

[@xiaowenzhao8968]

thanks for your share! now I have a problem:I put 467 points on the airfoil, 700 in the wake , with the same farfield size. i use your method to generate the mesh, after i run 40 steps there appears an apophysis at the joint pozition of trailing edge and the wake line, but a dont't know why? hope for your advice! thank you

[@AeroDZ]

Hello , I have a question Hope you can help me it , i have seen both your videos about C-grid and O-grid and i am wondering about which approache is best to calculate Aerodynamic coiefficients for both CL And CD for an Airfoil For differente angles of Attack -6

[@princeofpersia9826]

Hi Tnx for your video. How can we revolve this mesh axi-symmetrical with a nice and smooth mesh?

[@78RRem]

Really helpful video, thanks ;) I'd like to know how you would proceed to mesh a multi element airfoil, one having a sharp trailing edge and one having a blunt trailing edge?

[@TravisCarrigan]

I'm glad you enjoyed the video. A multi-element airfoil would make for a great video. I may have to put one together. But to answer your question, it's really not that different than this video or my other video for handling airfoils with blunt trailing edges and creating an o-grid. All you need to do is joint the elements together. For example, you'd create a connector that joins the trailing edge of the slat with the leading edge of the main element and another connector joining the trailing edge of the main element to the leading edge of the flap. You would then simply select the connectors defining the slat, main element, flap, and the branches to create a single closed loop...topologically speaking. By adjusting your growth rate and some of the smoothing parameters, particularly Kinsey Barth, you can extrude off everything at once. And, if you'd rather have a c-topology than an o-topology, just create another branch connector from the trailing edge of the flap to some downstream location like I did in this video.

[@78RRem]

Travis Carrigan Thanks for the info . Pointwise really seems to be THE software when it comes to multi element meshing, as blocking around is often quite long. I understood your linking/closing strategy, but I wonder if it would still hold if, say, the leading edge of your flap is upstream of the trailing edge of the main element. The connector would have to be really curved: would the extrusion handle it with suitable Smoothing parameters ?

[@TravisCarrigan]

Remi Roncen Depending on the severity of the connector's curvature, the extrusion may be possible with the right combination of suitable smoothing parameters. However, this might be a case where you'd have better success manually blocking that portion of the flow and then use the extrusion elsewhere.

[@arturocajal4111]

Hello, Thanks for the video. How do I change the offset of the grid before running the extrusion? In other words how can I set the further point of the c-grid to be a special coordinate? Thanks

[@TravisCarrigan]

If you'd like to specify an exact location for the farfield, you'll want to follow this example instead, kzbin.info/www/bejne/p5rJdXWCl7Cpra8

[@arturocajal4111]

Thanks for the quick reply. I followed that tutorial before, but I found this one faster method to generate grids around airfoils, that why I was wondering if there is any chance to do it using this extrusion method. For the kindly reply I am guessing that to achieve the exact coordinates at the edge of the grid I must follow the other procedure rather than this one, am I correct? Thank you again.

[@TravisCarrigan]

Arturo Cajal correct

[@arturocajal4111]

Thank you!

[@liangyuxu7891]

Thanks very much for the share. For higher angle of attack, should the same procedure be used? Should the mesh at the trailing edge follow the direction of foil shape? Thanks.

[@TravisCarrigan]

+Liangyu Xu Yes, the same procedure can be used. That said, the grid above can be used for a handful of angles of attack successfully by adjusting your inlet velocity components. The best technique to use when adjusting angle of attack is an o-grid. With an o-grid all you need to do is assemble one mesh and then adjust the velocity components. There is a video describing how to create an o-grid on my channel as well.

[@liangyuxu7891]

+Travis Carrigan Thanks very much!

[@liangyuxu7891]

+Travis Carrigan It seems O grid cannot work for a sharp trailing edge. What would be the best practice if I want to adjust the velocity for a sharp trailing edge foil? Thank you.

[@TravisCarrigan]

+Liangyu Xu Actually, an o-grid will work for an airfoil with a sharp trailing edge, but will generate poor quality elements at the trailing edge. Some solvers can handle the poor quality robustly. In reality, airfoils have blunt trailing edges and should be modeled that way. I'm not certain why sharp trailing edges are even used any more when performing CFD. But, if you must use a sharp trailing edge you will need to use a c-grid topology. Again, with a c-grid you can set all the outer boundaries to a freestream BC and change the velocity components. If your farfield is sufficiently far from the airfoil you can explore many angles of attack without having to regenerate your grid. Plus or minus 10 degrees, possibly even 15 degrees or higher. I would encourage you to experiment. Play with it and see how the solution changes when you simply change the velocity components versus rotating the airfoil and regenerating the grid. Also, another fun experiment is to change the farfield distance and watch the solution change. Be sure you are looking at integrated values like lift, drag, and pitching moment. Good luck!

[@liangyuxu7891]

+Travis Carrigan Thanks. Another question: I tried to increase the mesh number to study mesh convergence. However, when the mesh becomes very dense, unsteadiness occured and the result seemed to be worse. Have you encountered such problem?

[@alihaider82]

Hi, How can I do Grid Convergence Study with "Constant Grid Refinement Ratio" using "gridRefine" Glyph script for a similar airfoil study?

[@TravisCarrigan]

+A Haider Simply change the refinement factor and that will adjust all the spacings and dimensions.

[@alihaider82]

+Travis Carrigan Thank you for the reply. I am following the link: www.grc.nasa.gov/WWW/wind/valid/tutorial/spatconv.html for grid convergence study. So, i want to obtain three grids : Coarse, Medium and Fine. The ratio of Medium/Coarse = 2; Fine/Coarse = 4.... but when I change the refinement factor 2, 3, 4,...times over a baseline Coarse mesh, the refinement ratio is no more constant.

[@alihaider82]

The grid in the given vidoe has very high aspect ratio and highly stretched cells that may cause instabilities in the solution. How to get a good aspect ratio of cells?

[@TravisCarrigan]

+A Haider It will be constant if you are using a structured grid. If you look through the script it simply takes that factor and multiplies all the connector dimensions by that amount and divides all the spacings by that amount. If you are using an unstructured grid then it will not be constant.

[@TravisCarrigan]

+A Haider High aspect ratio cells are not a problem for cases such as this. I've run aspect ratios >1000 in in some cases close to 10,000 without issues so long as there are no large gradients in the streamwise direction.

[@manuelsanchez4029]

Ey you are a CHAD