simulation flow over sailing yacht appendage
Hello! I'm using the free-trial version of caedium professional and I have to simulate the flown over some sailing yacht appendages for my thesis in order to calculate the resistance and the polar of every appendage (I need it for a vpp program).
I'm importing the appendages in .stl format (from rhino) but I have some problems with the surfaces and the simulation doesn't start... I think because I cannot set correctly all the parameters.
I've also read the example "flown over an airfoil" and "NACA 4415 airfoil calculation" but I don't want to recreate the appendage in caedium, I just want to import it, and when I import the .stl file the program doesn't recognise the leading and trailing edge.
In "NACA 4415 airfoil calculation" example I see that is possible to calculate the airfoil lift and drag...How can I do that? and what kind of resistance the program can calculate? Only form drag, form and induced drag or also the friction resistance of the foil?
Someome can help me or can tell me where I can find a caedium guide?
Thanks.
Feedback
Questions? Ideas? Problems?
Prefer STEP Over STL
STL is not a good format to exchange geometry, if you have the option it is better to use STEP.
After importing STL (or any other surface-based facet format, e.g., OBJ) into Caedium you can only perform surface based CFD simulations using the Panel Flow add-on, by setting the Substance->State properties as follows:
The panel method assumes idealized inviscid (non-viscous) flow, which is good for lift and pressure (form) drag results. For more details see "Why Use a Panel Method?"
Unfortunately STL doesn't preserve feature edges, so you won't be able to attach a wake and then the panel method won't be able to calculate lift.
If you want to account for viscosity (i.e., skin friction) and separation then you'll need to use the RANS Flow add-on (default Substance->State settings). However, it uses volume-based solvers, so you'll need to provide a solid model representation of the flow domain by either creating geometry with Caedium, importing true geometry (e.g., STEP), or importing an existing volume mesh (e.g., a FLUENT msh/cas file).
The Caedium Essentials Manual is a must read and then take your pick of relevant tutorials. You might also find the forum post "Setting up an external case" helpful.
stp format and simulation
Hello! After reading the forum post "setting up an external case" I've imported the appendage in .stp format but it doesn't have only 4 surfaces (in rhino the appendage has only the tip srf, the root srf and the 2 sides of the wing) but a lot of surfaces that reproduce the leading edge and so, when I start the simulation with the panel method, the program calculates all the quantities on every surface.
I've tried to fix all the surfaces but I don't know how to obtain a single surface on every side of the wing (I hope that you understand what I mean).
For the same reason, when I try to launch a RANS simulation, after a lot of time, the program shows me an error message without an results [In this case I've created a domain (a box) that contains the wing and, for the simulation, I use the volume that results from the subtraction of wing and domain.]
How can I resolve this trouble?
Thanks.
Require a Solid Volume
You need to create a solid volume of the wing, where each edge is connected to 2 faces. Then subtract that wing volume from the large box - analogous to the wheel and the box in the tutorial "Wheel in a Box."
My understanding of Rhino is that it is a surface modeler. So once you have your surface geometry in Caedium you will need to use the Volume->Stitch tool on the Geometry Tool Palette to stitch the faces to form a solid volume. Keep an eye on the log as you perform these operations.
It's fine to have multiple faces represent the upper and lower surfaces of your wing, but they must be joined as described above.
I'd also recommend taking a look at the recent forum post on validating your mesh, "Volume won't mesh," as a precursor to trying to run a simulation.
If you could quote the errors you are seeing it would help to better diagnose the problems you are having.
polar curve
hello! now the simulation with panel method and rans method works well and the results that I obtain are similar to the result for the same appendage that I obtained with openfoam... but I have to ask if it's possible to plot, instead of the drag and lift forces, the drag and lift coefficient with the panel method and if it's possible to plot the forces respect the angle of attack of the appendage (now I can plot only respect to the "time"). thanks.
Lift and Drag Coefficients
Good to hear you're making progress.
You can plot lift and drag coefficients by creating new Results. I suggest you review:
It is possible to perform an automated angle of attack sweep with the panel method, for an example see "NACA 4415 Airfoil Calculation". The project .sym file is available towards the end of the article.
just a question!
Hi I'm a new Caedium user,
I've just download it trial to try calculate the dynamic coefficient on a mainsail with a panel method calculation.
I make a benchmark test on NACA 0012 wing, and I appreciate the results vs Time calculation, so I decide to continue to use this solver, but now I've some questions:
1) as your explained before to introduce a wake, and use a lift theory, is necessary to have two surfaces connected on the same egde, so I can't use the only one surface to simulate the sail?
2) If I utilized two surface connected ( at the Leading edge and trailing Edge) separed by a distance of 2[mm], the Panel Method can't converge even if I use custom accurancy 75 and GR:1.05.
to obtain some results, I have to icrease the distance till 10 [mm] and rounded off the LE by CATIA, but I'm not able to understanding if these are really reasonable...
have you any Idea how is better approch on this matter, considering that actually I need to have just an idea on the configuration geometry to use on a RANS analysis?!
thank you in advance ,
Franco
Wake needs separate faces, gap is too small
In answer to your questions:
A Caedium RANS Flow simulation doesn't have the wake constraint, but you are still better splitting up the airfoil in separate faces so you can attach an Accuracy constraint to the leading edge.
Thanks for your answers, but
Thanks for your answers,
but probably there was a misunderstanding for the point 2, following I try to explain better the matter:
I'm studying a main sail, before to perform a RANS analysis on the model I have to evaluate the different behaviour of my design parameters, so I need acceptable quick results given by panel method.
the geometries under evaluation are characterized by Croot is 2 [m], Ctip 0,5[m] and the span is 5 [m]. The airfoil is obviusly thin. I simulated this sail, in two way (see my previous post). Second way, I stitch lower and upper surfaces in the L.E. and T.E.. To avoid big difference with real beaviours I try to give as small as possible airfoil thickness ( that correspond at surfaces distance), using meh parameters described in my previous post, but the method didn't converge untill I reach a thikness equal to 10[mm].
I'm choose a wrong way to simulate the sail or is the method not able to solve this?
Thanks in advance,
Franco
Airfoil Thickness Constraint
OK I see, you mean the the thickness of the airfoil is 2 mm. Yes, that is still a problem, because the panel method assumes a thick body and I guess you've discovered that in your case that means at least 10 mm.
For quick turnaround you might be better looking into "Thin Airfoil Theory". Although, as I'm sure you are aware, these linear theories are not accurate near stall.
Hi, just a last question...
Hi,
just a last question... In your opinion, considering the solver method used by CAEDIUM and taking in account that the upper and lower face will be create by estrusion of middle surface and so with the identical shape,
how much and in which way airfoil thickness (10 [mm] instead of 2/3 [mm]) can influence the aerodynamics coefficient analysis?
thank so much
Franco
Drag will be wrong
I would imagine the difference to be significant, especially the drag coefficient. However, you will not predict accurate drag with a panel method - it assumes inviscid flow. For more details see "Why Use a Panel Method?".