Hi everyone,

I waste lots of time in creating my CAD geometry. I'm using CADKEY 99 & Mechanical Desktop3 for solid modelling (they have limited function in converting the multiple surfaces into the single solid surface). Just wonder if you can recommend any other excellent CAD package, which is easy to use & time saving.

Thanks for advice.

Regards, david

Hey David!
Depending on the amount of money you like to spend. Products like UG, Catia or ProE are kind of expensive. I heared of Rhino to offer good functionality at a low price.
Greetings, Marc

If you want surfaes and solids you should try "Thinkdesign" (www.think3.com). It is also possible to do some sort of parametric changes to surface models which they call "Global Shape Modelling". Parametric for solids works as usual.

I can get Pro-E but I'm not quite familiar with it. If you're Pro-E user, could you please tell me if it's possible for Pro-E to first create a set of multiple surfaces and then stitch them together to create a single solid surface?

Anyone know the answer...could you please kindly consult me? Thanks a lot.

David,

We have been going through lots of issues with CAD software at our company in the recent months. To start, there are 3 main classifications of CAD = high end, mid end, and low end. Price and functionality follow these catagories.

ProE and UG fall into the high end. They have great functionality, are high price, and generally require an expert user to make it sing. I know I will cause a bit of a stir here, but generally there does not really seem to be major compelling advantages/disadvantages in the high end market - i.e. they are all good. Without ever using ProE personally, I can still be highly confident that the answer to your question is yes. The problem is that if you have never used ProE, you may struggle with this.

The mid end CAD packages are a factor of 2 to 3 less expensive then high end tools, and also much easier to learn. After a few days with the tutorials, I was modeling very complex turbine blades. The nice thing is that the gap between functionality of mid end and high end is closing rapidly. We currently use Solid Edge - the same company also develops Unigraphics. Be careful here, there is a WIDE variation in capabilities in the mid end market. I will not give you a completely rosey picture of Solid Edge, but it is fairly good (albeit, crash happy and a resource hog). Our problems are associated with free surfaces of our turbomachinery applications. However, we have work arounds. What you have described should be handled with Solid Edge.

I will not even touch the low end market - it is my feeling that generally you get what you pay for here. I.e., not much. Of course, some will strongly argue against this point and I will concede that there are exceptions.

Sorry to be long winded. Maybe you can describe a bit further what you are trying to do?

Steve

David,

Pro/E can do anything you can think of, and then some. That's why it's so expensive. The new interface that they've rolled out in Pro/E Wildfire should go a long way toward making operation more intuitive (but I haven't gotten my upgrade yet, so I can't speak from experience).

My current favorite solid modelling CAD package is ParaLogix from A3DS <www.a3ds.com>. It based on the ParaSolids geometry kernel and can do all the standard parametric solid modelling things that SolidWorks can do. I got the Level 1 version when it was $350 (!!), although it's now selling for $495. In Level 1 I'm limited to 50 procedures (like operations, but some procedures have multiple steps), and 3 drawing sheets per model. So far these limitations have not prevented me from getting the model I want.

I have heard some good things about Alibre's CAD package (also $495), but I have not had an opportunity to try it.

Good luck, Alton

Glad to hear your comments here. I'm working on the turbomachinery as well. Drafttube is the first component that I'm going to model in CFD. I currently use the CADKEY99 to create the solid.

Due to the changing curvature & cross sectional area, it's easier (in my viewpoint) to build the solid by first creating a set of ruled surface(e.g.) and then convert it to the solid. CADKEY does have the heal function to stitch the multiple surfaces into a sheet body and then I can convert the sheet body into a solid.

The problem is CADKEY crashes heavily whenever I try to stitch all the surface together. Not sure if anyone has experience in this CAD program but the error is really consistent. (I'm not CAD expert & could be wrong though)

Seems like you have quite a lots of experience in solid edge. Can you please tell me if the approach I mention above is workable in Solid edge V7? I can get this software as well. Is it really user friendly? I'm quite anxious to get the CAD problem solved.

Thanks for advice.

I will check the Pro-E website to see if it's suitable for me. What I want is to convert the set of multiple surfaces into the solid. I already have a set of curves & surfaces defined in CADKEY...If Pro-E will accept the imported IGES/SAT/STEP file, it will be a lot easier for my work then.

Thanks for your comment here.

I think it will be more proper for me to consider some popular software...This will make sure that I don't have any problem to transfer the file into CFD software later. Thanks for your advice anyway.

David,

I would say your approach should workable in Solid Edge. However, we are up to version 14. I have started using it since version 11 and have noticed enormous leaps in capabilities. I don't want to sound too negative, but based on what my colleagues have told me about these earlier versions, it does not sound promising in version 7.

Any way, my particular method has been to import data points to produce closed curves. Then I use the lofted surface feature to create a surface for each closed curve (in my case, each closed curve represents an airfoil section on a cylindrical cut). Finally, I use the lofted protrusion to create the solid that passes through each surface. Of course accuracy increases with number of sections you import.

Let me ask a couple more questions: * (This one is more curiosity) – you mentioned drafttube, so I am going out on a limb and assuming hydro applications. Where are you located? * Is this a radial machine? Not that it should matter too much, but I have certainly found radial machines more difficult to deal with in Solid Edge compared to my axial blade rows.

Steve

Steve,

Thanks for your information. I will take your word and will try to find out the latest version of Solid Edge. One more question about Solid edge: Is that user friendly? I don't actually want to spend too much time on the CAD software.

I actually try the method that you have mentioned here. I only have information for 12 different cross section (from the old drawing) so far. CADKEY gives me some funny shape between some of the sections when I loft them together. I think the problem could be due to insufficient number of cross sections. I actually want to avoid interpolating the section in between as the old drawing is not actually printed in scale(may be very inaccurate). It would be painful if I have to do so (seems like this is a more realistic solution though). This is the main reason why I put my effort on converting the multiple surfaces into a solid one.

Regarding your questions: I'm working on the Hydraulic Francis turbine and I think you can look at it as a radial machine (certainly not a axial flow machine). I'm not worried too much about the runner part as the blade row is uniformly distributed. Once you create a set, you can duplicate it to the others (Unlike the draftube where the section changes from circular to rectangular and somehow it's a curved duct!)

This makes me crazy. Any fresh idea of how to reduce my workload? Please kindly bring it out. Will be very grateful for that. If you are kind enough to help me, I can also send you the IGES or any format for you to take a look at my geometry (of course I know this chance is very little). Thanks anyway.

HyperMesh from Altair is extremely good at cleaning up CAD data. It can stitch, suppress, modify, trim, and make reasonably complicated new surfaces. It's surface meshing is also very good. I'll then use the surface mesh as input for ProAmm or ICEM.

Regarding user friendliness - that is such a subjective question that is difficult to answer. I will say that after a couple days on the tutorial, I learned the system well. Once you learn their terminology and understand the work flow for the process you need to do, then yes it is user friendly. However, it takes a little to get used to this, so upfront it appears it may not be so user friendly. I.e., using Microsoft as an example (maybe a bad example) you basically know what to expect in the different office products because things are similar and it is easy to poke around. So, I will say this is user friendly. However, for somebody just looking at any MS office product for the very first time, they would probably say it is not user friendly.

It's hard to offer suggestions without seeing your case directly. I am tied up the rest of today, but will send you an e-mail that you can send the IGES file to tomorrow. I will take a quick look and see if I can offer any input.

Thanks a lot, Steve. I'm very grateful if you're happy to look at my case. Not necessary have to come out with solution but at least you could let me know if my approach is hard to tackle in Solid Edge. Thanks god.. you're willing to give me a hand.

Sales people told me that ICEM is quite difficult to use if compared to other mesh generation tool (eventhough it can be quite powerful). I'm not quite sure if ICEM offer a hybrid mesh tool. They do have separate tool for structured & unstructured mesh if I remember correctly.

Regarding your recommendation, I will check it out and get back to you later. Thanks for the information.

Why are you generating surfaces and then stitching them together? This is a tedious and error prone procedure compared to working with a tree of volumes (unless the topology is trivial of course).

You do not mention the analysis/grid generation package you are using which may influence the handling of the geometry (e.g. it may not work with consistent volumes but collecitons of surfaces?).

As a sweeping generalization, speed and reliability come from you adapting the tools to your procedures rather than you adapting to theirs. In practice, this means working more with your parameters and your scripts and less with pointing and clicking in endless windows supplied by the packages. Reasonable support for working in this way is mainly found in the larger integrated packages.

I would echo the observation made earlier that ProEngineer is a good and flexible CAD package.

This is exactly my problem: the topology is trivial.

Regarding your comment 'adapting the tools to your procedures rather than you adapting to theirs', unless you're developing your own software, most of the time you have to adapt to the software that you're using. If it's user friendly, it will let you do the job with whatever method that you want to. It's just my viewpoint & you can have yours. I don't want to argue anymore. This doesn't help solving my problem anyway.

I'm now looking at the Pro-E site as well. Thanks for your reply anyway.

David,

I sent an e-mail to your address regarding helping with your files. Wondering if there was a problem with my transmission?

Hi Steve,

I've send you the IGES file a min ago. Just get your email this morning. Anyway, thanks for spending your time on my problem.

I think you have to consider the entire process not simply the solid modeler. There are many factors that should govern your decision. What solver(s) do you use routinely? How are you planning on meshing the models? What type of mesh do you need? Hybrid? Hex? Some of this drives the type of modeler you want to consider. Are you looking for a complete CAD package that creates 2D drawings as well?

Everyone has laid out the high/mid/low range markets? All of the high end packages are very powerful but they take quite some time to master. The flip side of that is they are used in many large companies, so knowing them is beneficial. Also some high end packages have a built in or add-on meshing module.

Midrange packages are very competent. I am familiar with Solid Edge, SolidWorks, Inventor etc. SE and SW are very similar each has its strenghts. SE has just introduced a new free form surfacing capabilities in V14 (blue point). Think3 may be a nice modeler, but I am not sure how robust they are with their file tranlation. Many CFD, FEA and meshing codes read in native cad files such as SW, SE but none that I know of read think3 native files. You are restricted to neutral files. I would steer away from low end cad. Stick with something that has a reputation and has been proven.

Most CFD codes come with their own preprocessing capabilites, meshing etc. You may want to consider a stand alone package. ICEM CFD is very powerful and if you are considering complex surfaces, you may need to consider something like this. ICEM has a very nice hybrid meshing module if you can afford to work with tets. Their hex mesher is awesome, takes some time to become familiar but that time is well spent. Once you become proficient, you can mesh nearly anything. Their are others on the mkt as well, Hypermesh, etc.

We use ANSYS for FEA and CFD and they also released recently a solid modeler for analysts. Has many of the features of the midrange, extrude, sweep, revolve, multiple bodies etc. No drawings. Writes out parasolid files. Its not bad and is half the price of the midrange cad tools.

Apologize for the ramble but just giving my two cents.

cheers,

derrek