AnsweredAssumed Answered

Flow Simulation geometry check

Question asked by Brian Robinson on Jan 6, 2009
Latest reply on Jan 26, 2009 by Brian Robinson

Like • Show 0 Likes0
Comment • 14

Hi, I was wondering if there is an easy way for the system to highlight where invalid geometry exists after performing Flow Simulation - Tools - Check Geometry.

I have created an assembly file containing the assembly file of the aircraft along with a solid block enclosing the aircraft. I use the cavity feature to incrementally remove the solid components of the aircraft until the Boolean operation fails. Usually I can find coincident sharp edges or cylindrical faces just touching planar faces. Moving faces to overlap or create a slight gap normally resolves these issues. Is there a more efficient process to find invalid geometry?

Thanks,
Peter Ferreira

No one else has this question

Outcomes

14 Replies

Rich Bayless Jan 15, 2009 8:01 PM
Mark Correct
Correct Answer
Hello,

I'd like to know the answer to that question too. It would be a good feature to add to Flow Sim.

Anyone know if there is a tutorial or other doc discussing this topic? How about samples of problem geometries? Like Brian said, "coincident sharp edges or cylindrical faces just touching planar faces". Show us what 'not' to do.

Rich.
Like • Show 0 Likes0
Actions
- Bill McEachern @ Rich Bayless on Jan 16, 2009 9:20 AM
  Mark Correct
  Correct Answer
  if you are using flow sim 2009 or even floworks 2008 later SP's (I think - might just be 2009) the check geometry command will highlight the issue geometry in the graphics area. When the geometry issues are listed they will either be in black text or gray text. The black text means that floworks/flo sim was not able to deal with the issue and gray means that it could deal with the issue. By deal with the issue I mean it can complete the boolean to produce the fluid volume solid it uses behind hte scenes to mesh and resolve the fluid region. I don't get into this command too much as I am usually building my own geometry so the above might be off a bit but that is how it work last time I tried it which was a lot better that it was before. The other thing one should be aware of is that the check geometry command is not bullet proof. I have had one case where invalid geometry existd in the model and was not detected by the check geometry command. That is once in ten years of use - pretty good but it is good to be aware of the fact that it can happen.
  Like • Show 0 Likes0
  Actions
  - Rich Bayless @ Bill McEachern on Jan 16, 2009 12:35 PM
    Mark Correct
    Correct Answer
    Hi Bill,
    
    I'm using SW2008 SP3.1, and it does not highlight the suspect geometry. I've been waiting for at least SP1 to come out before switching to SW2009, now that SP1 is out I guess it's time to upgrade!
    
    Rich.
    Like • Show 0 Likes0
    Actions
    - Bill McEachern @ Rich Bayless on Jan 16, 2009 12:42 PM
      Mark Correct
      Correct Answer
      2009 SP2 was released the other day. 2008 SP5.0 beghaves the same way if you don't want to upgrade yet.
      Like • Show 0 Likes0
      Actions
      - Rich Bayless @ Bill McEachern on Jan 16, 2009 12:57 PM
        Mark Correct
        Correct Answer
        Hi Bill,
        
        oops, I spoke too soon. SW2008 SP3.1 does have the capability to highlight 'invalid contacts'. Here's the help text:
        
        "In the Invalid Contacts box you will see the list of invalid contacts (if any). Invalid contacts may be selected by clicking on their names in the Invalid Contacts box. The selected contacts (vertexes and curves) are highlighted in the graphics area with bold red dots and curves, respectively."
        
        My problem currently is not 'invalid contact', but instead I have a model that keeps coming up with zero fluid volume. I triple checked that there are no openings. If I uncheck 'Exclude cavities without flow conditions' and check 'Create fluid body assembly', I can see that the fluid body is correctly created.
        
        To be continued.....
        
        Rich.
        
        Like • Show 0 Likes0
        Actions
        Bill McEachern @ Rich Bayless on Jan 16, 2009 12:59 PM
        Mark Correct
        Correct Answer
        sounds like a leak....if I had to guess
        Like • Show 0 Likes0
        Actions
        Rich Bayless @ Bill McEachern on Jan 16, 2009 2:20 PM
        Mark Correct
        Correct Answer
        Hi Bill,
        
        hmmm, does Flow Sim have a leak detector?
        
        I did not try the 'Create Lids' option, maybe that would have found an opening and put a lid on it?
        
        Rich.
        Like • Show 0 Likes0
        Actions
        Bill McEachern @ Rich Bayless on Jan 16, 2009 2:27 PM
        Mark Correct
        Correct Answer
        not sure what your model looks like but I just extrude a box around the problems and just replace the boundary. Sheet metal features on outer boundarys are prone towards hard to detect leakage.
        Like • Show 0 Likes0
        Actions
        Rich Bayless @ Bill McEachern on Jan 16, 2009 2:48 PM
        Mark Correct
        Correct Answer
        Well the 'Create Lids' tool found an 'opening', where we added a chamfer and then shelled inwards to create a thick shell. Sure can't see any opening in the solidworks side of things.
        
        Looks like we'll be sending this model to our VAR for their opinion.
        
        Rich.
        Like • Show 0 Likes0
        Actions
        Brian Robinson @ Bill McEachern on Jan 16, 2009 3:13 PM
        Mark Correct
        Correct Answer
        Hi Bill, it would be nice if I could easily find and correct the leak. Yes, invalid contacts are highlighted in the graphics area, but not the invalid geometry which was what I was trying to locate. Anyway, using the cavity feature helped to diagnose most of the Boolean errors, but it is still a tedious process. SolidWorks tends not to highlight all problem areas to make things easier for the end-user.
        
        Anyway, now the geometry check passes with no errors and shows the correct solid volume for both the aircraft assembly file and also the solid core of the cavity part file. The next issue is that an initial coarse mesh (say 3) shows all solid bodies are included, but a more finer mesh (say 4) shows that many of the solid bodies are ignored. Going back to a coarse mesh (at 3) shows the same solid bodies are ignored now. Even running the mesh twice without changing any parameters shows the same inconsistencies. Meshing the single solid core body of the cavity has the same issues. I even created a solid body output from the check geometry option to analyze and found the same meshing issues.
        
        I am running SolidWorks 2009 SP1 on a 64-bit PC. Does SP2 resolve these disappointing meshing issues?
        Peter Ferreira
        Like • Show 0 Likes0
        Actions
        Bill McEachern @ Brian Robinson on Jan 16, 2009 6:48 PM
        Mark Correct
        Correct Answer
        Hi Peter,
        
        I have never seen those types of meshing problems that you are describing. If you are tyring to do an external flow simulation on the whole aircraft configuration ad you want to use the detailed design model data for the airplane I would try exclude internal cavities and make sure the minimum gap number is left at the default value and turn off thin wall optimization and see how it does. Obviously plug up any big holes through the outer mold line. Ideally you want to creat a single body of the aerodynamic shape. If you need to examine the engine area I would just add detail to that area to get a decent approximation. If you want to get more details you can use the EFD zooming capability or the transfered boundary conditions to get a more refined assessment inthe engine area. Some of this can be don by copying hte outer surfaces into a single part, patchinghte holes and turingit into a solid body by using the knit surface command or thickening a set of surfaces and then patching up holes. Hard tohelpmore without a maybe seeing a picture.
        Bill
        Like • Show 0 Likes0
        Actions
        Brian Robinson @ Bill McEachern on Jan 19, 2009 5:37 PM
        Mark Correct
        Correct Answer
        Hi Bill,
        
        Attached are images describing the issue. I used a symmetry plane at X equals zero. I used a solid core of the cavity of the aircraft assembly, so there are no internal hollow spaces. An initial coarse mesh at 3 shows one result, and repeated again with no changes shows a different result. Check geometry reveals correct volumes and no issues. Am I the first to see this?
        
        Peter Ferreira
        Attachments
        mesh 3 repeat.JPG285.7 KB
        mesh 3.JPG293.0 KB
        Like • Show 0 Likes0
        Actions
        Bill McEachern @ Brian Robinson on Jan 19, 2009 5:45 PM
        Mark Correct
        Correct Answer
        Hi Peter,
        
        No images attached. YOuhave to figure out that tool...you might want to do a test....to get a successful attachment to post..
        Like • Show 0 Likes0
        Actions
        Brian Robinson @ Bill McEachern on Jan 26, 2009 2:12 PM
        Mark Correct
        Correct Answer
        Hi Bill,
        
        I attached the images just as you posted your message. It was another trial and error approach to get it right. Attach File - Upload File - Attach File again was the right combination. One thing I have not been able to figure out yet is how to perform an advanced search of desired words in only one forum category. Even though only one is highlighted, the results always shows the words from all forum categories.
        
        Anyway, I ran another test using SP2 and still had the same issues.
        
        Peter Ferreira
        Like • Show 0 Likes0
        Actions

Flow Simulation geometry check

Outcomes

Attachments

Related Content

Recommended Content