18 Replies Latest reply on Sep 26, 2009 3:58 PM by Mauricio Martinez-Saez

    Solidworks Simulation Power & Stability

      Hello,

       

           I'm new to the forums, and I have a few questions on the power and stability of Soildworks Simulation. I'm currently a Mechanical Engineer working for an aerospace company. I've been using Unigrapghics Nx and ANSYS (Classic and Workbench) for six years now. Several of my suppliers have seen my work and have requested that I perform some FEA analysis for them. Consequently I've been looking into purchasing a CAD program and a FEA program for my personnel use. I've heard a great deal of good things about Solidworks, and Solidworks simulation. I joined the forum to learn more about these programs. I was hoping to get some of your opinions and possibly get in touch with Vince Adams. I read Vince's book "Building Better Products with Finite Element Analysis" and I've also read one of his papers from Impact Engineering. If anyone knows Vince could you please pass my e-mail along to him. As for my other questions, I've been using ANSYS for quite sometime and I am very happy with the product however the software is cost prohibitive. I was wondering if people are using SW simulation for large non-linear dynamic, non-linear static, snap through buckling, forced response, or transient thermal problems? I was also wondering if SW Simulation can take advantage of multicore processors and high capacity memory workstation 24 GB and higher? One other concern I have on this software is I continue to read on these forums it appears that Soildworks Simulation may not be the most stable program. If you do a quick search it appears the forums are littered with issues relating to the stability of the product. Are these post exaggerations or is this a product that is being released too early. Thanks for everyone's help, I'm not sure if this was the right forum to post on or if there was a better way to reach out to existing Solidworks Simulation users. Also if anyone knows Vince Adams please forward my e-mail to him repars@aol.com.

       

      Thanks again

       

      JP

        • Re: Solidworks Simulation Power & Stability
          Derek Bishop

          I'm new to the forums, and I have a few questions on the power and stability of Solidworks Simulation. I'm currently a Mechanical Engineer working for an aerospace company. I've been using Unigrapghics Nx and ANSYS (Classic and Workbench) for six years now. Several of my suppliers have seen my work and have requested that I perform some FEA analysis for them. Consequently I've been looking into purchasing a CAD program and a FEA program for my personnel use. I've heard a great deal of good things about Solidworks, and Solidworks simulation. I joined the forum to learn more about these programs. I was hoping to get some of your opinions and possibly get in touch with Vince Adams. I read Vince's book "Building Better Products with " and I've also read one of his papers from Impact Engineering. If anyone knows Vince could you please pass my e-mail along to him. As for my other questions, I've been using ANSYS for quite sometime and I am very happy with the product however the software is cost prohibitive. Finite Element Analysis

           

          Vince sometimes comes onto the forums. Haven't seen him for  while.

           

          I was wondering if people are using SW simulation for large non-linear dynamic, non-linear static, snap through buckling, forced response, or transient thermal problems?

           

          The main areas of use for me are in static analysis with some frequency, buckling, thermal and pressure vessel analysis. I suspect the more complicated the software is and the less people there are that use the software the worse the problems become.

           

          I was also wondering if SW Simulation can take advantage of multicore processors and high capacity memory workstation 24 GB and higher?

           

          I have a dual core processor so cannot speak from experience. I believe that quad core processors will be utilized in many of the FEA processes. Have a hunt around in the search of this site and you will find some discussion and documentation on the issue.

           

          One other concern I have on this software is I continue to read on these forums it appears that Soildworks Simulation may not be the most stable program. If you do a quick search it appears the forums are littered with issues relating to the stability of the product. Are these post exaggerations or is this a product that is being released too early. Thanks for everyone's help,

           

          I am also a mechanical engineer. I've been involved with doing FEA work for a couple of years using SolidWorks professional and have had all sorts of problems. I've done the training and while not an expert consider myself to have reasonable competence in using the software.

           

          As I've become more used to the software, stability problems have become less of an issue. It seems to me that it is often that you have done something not quite right in setting up the simulation. The software lets you do this but then bombs when you try to run the study. The approach that I've found works best is to start from the known and simple and gradually add in complexity. You sound like you've had a bit of experience so this may be second nature to you. I'm interested to hear about your experiences with other FEA software and the kind of problems that occur and how often.

           

          For me at least the issues now tend to be more software problems than crashing. A job that may take a couple of days can string out for weeks when trying to problem solve and trouble shoot the bugs in the software. This is not so bad if you have a lot of repeat work. It quickly becomes a waste of time and effort if you are doing a lot of one off type projects. Unless of course you have a very understanding client who is quite happy to keep paying you to troubleshoot SolidWorks software. I  personally do not find a lot of these clients.

           

          An example of a problem that I've been having over the past few weeks. I use SW 2009 SP3. A few weeks ago I was doing FEA on a plastic assembly. I defined a custom material property data base and then went back to change the properties of the materials data base. I found that I had a fatal error. I could no longer access the data base. This was sent off to SolidWorks to fix which they kindly did. I then had to do the same thing again and found the same problem happened. I sent them off an SPR and basically their response was that they could not replicate the problem in 2009 SP4 and I should upgrade. I've seen people having a few problems with SP4 so I'm reluctant. On top of this there are other problems changing and editing the material data base. This causes confusion and may result in materials having the wrong properties applied or even being changed unknowingly. I've been going backwards and forwards with my VAR and SolidWorks for several weeks on these matters. This was for a job that took a couple of days. I can expect at least two or three of these issues on a job. Not all problems are this drawn out.

           

          I suspect the problem stems from several areas. Lack of expertise and industrial experience in both SolidWorks and the VAR support is a factor. What may seem to be a trivial problem to the programmers can often be cumbersome, inefficient and even cause fatal errors for a user. The feedback from the customer is a slow time-consuming and laborious exercise and done at significant cost to the customer and or client. QA/QC seems at times to take a back seat to commercial constraints and marketing forces in the SolidWorks machine. In the past at least, some problems I've seen in the software are so blatant you are left shaking your head. SolidWorks seems to rely extensively on customer feedback and problem solving to identify problems in the software. That  works okay if you've got thousands of people using the package and doing the same thing. With the more complicated and less used addins like Simulation it seems to me that often I am one of a few people who are actually using the software and finding the problems. Even with an RX (video of the problem) it becomes hard for SolidWorks identify and replicate anything other than relatively simple repeatable problems. Unfortunately when doing complicated analysis and models this is rarely the case.

           

          At the same time I am Beta testing 2010 with some tank design and have been very impressed with some of the latest enhancements and the capability of the software. I have found it both strange and frustrating to use a program with so much potential and yet which keeps getting tripped up by a lack of attention to some fundamental problems.

           

          In the long term I believe that the extra money spent on getting reliable software is worth the investment. If you have found other FEA packages that do this then I'd suggest stick with them. If you choose to invest in SW Simulation be prepared for some frustration. I'd get the trial version and try redoing some of your past case studies or see if the local VAR will do these to understand its capabilities.

           

          I have a friend who works with a company that has both SW Simulation and Ansys and he prefers to use SW Simulation because of its ease of use and in some situations has greater capabilities. He has the SW Simulation Premium package.

           

          Stay away from the early releases of the software. It is normally not until around the third service pack release that things start to settle down. Had a lot of problems with the first releases of Simulation 2009.

           

          I'm not sure if this was the right forum to post on or if there was a better way to reach out to existing Solidworks Simulation users. Also if anyone knows Vince Adams please forward my e-mail to him repars@aol.com.

           

          Simulation seems a suitable place for these questions.

           

          Well you asked for people's opinions, and I've given you both barrels. Hope it helps. As mentioned above, I would like to hear of your experiences with the other packages you've used.

           

          Kind regards,

           

          Derek Bishop

            • Re: Solidworks Simulation Power & Stability

              Thanks for taking the time to respond to my questions. I'm sorry to hear about the in-stability and quantity of bugs in the SW software. I was hoping this was the best FEA software for my needs. As I mentioned in my original post I've been using ANSYS and ANSYS workbench, stability is never a problem and I've only found 3 bugs in the software during the last 6 years. Two quick follow up questions. Is there any action plan for SW to address the stability issues. Also with every copy of ANSYS it includes an Accuracy Verification Manual to ensure your software is working. You can run through examples and see how the FEM compares to traditional calculations. Is this something that SW supplies when you purchase the Simulation package?

               

               

              Thanks

               

              John

                • Re: Solidworks Simulation Power & Stability
                  Derek Bishop

                  Each year I've heard them say the software will be more stable. Each year the dramas continue.  You have to understand the software is constantly in a state of flux. With every new development there are new bugs and instabilities. I have given my assessment of the situation here

                   

                  https://forum.solidworks.com/message/113294#113294.

                   

                  In summary,  at the Beta testing stage the software is still full of bugs. From what I can gather, from this stage and onward SolidWorks are almost totally reliant on the goodwill of users to identify bugs in the  the software. I would like clarifcation from SolidWorks on this point. There are relatively few users of the software actually doing the Beta testing so a lot of the bugs fall through the cracks. You discover them when you start using the software on a job. Also there is no guarantee that Solidworks will actually fix the bugs once they have been identified. It depends to some extent on their marketing preferences and on what they think is a significant problem. Sometimes the bugs are fixed in the next release which may be several monts away. Sometime the bugs are fixed in the next version. That is the case with the problems I'm having with entering the material properties.

                   

                  If your designs are relatively simple and you are just trying to get a rough appreciation of the design you may find the software acceptable. For more complicated and serious engineering work it is unacceptable. If you've had three bugs using Ansys in six years of usage my advice would be to stick with the product.

                   

                  There are no bench mark examples that give a comparison of results at present as far as I'm aware. There is a white paper that gives a comparison of results which you should be able to find through the search facility. My assessment is that the actual FEA solver is robust and accurate. I've done manual checks that have demnstrated this time and again. The software falls over in what should be the easy part, ie. the user interface. These occur in both the pre-processing and post-processing interfaces.

                   

                  I'd be intersted to know the types of products you are designing. Mine tend to be multibodied parts and some assemblies. Lots of platework,  beams or weldments and pipes on occasions.

                  • Re: Solidworks Simulation Power & Stability
                    Brian Zias

                    I didn't see any replies regarding the Verification studies that the OP asked for.  Once installed, there are dozens of online tutorials.  There are also dozens of verification problems.  They actually give you the hand solution and the file set so you can work through them.  They have verification for all the different study types, as well as NAFEMS benchmark problems.

                     

                    The Flow Simulation comes with a technical reference PDF that includes the verification examples, about 2 dozen of them.

                     

                    I also have several additional documents with more verification problems, please send me an email if you'd like copies.

                     

                    Brian

                  • Re: Solidworks Simulation Power & Stability

                    Has any one bumped into Vince Adams?

                  • Re: Solidworks Simulation Power & Stability
                    Mauricio Martinez-Saez

                    John,

                     

                    The following are my comments:

                     

                    I used ANSYS for a long time on very complex analysis and while I do not use it for the last 4 years and therefore I am not familiar with any last version, I can say that SolidWorks Simulation (former COSMOS) is quite more easy and friendly to use.  If you get the Simulation Premium (when I purchased was COSMOS Advanced Professional) you can do all ANSYS can do and some more things.

                     

                    Over 90% of the problems people report on this forum is caused by lack of experience working with FEA systems, bad modeling, etc.  I encounter a BUG from time to time but no more that with any other application,  on the other hand you need to understand that Simulation is presently going thru some code changes in the process of being integrated to SolidWorks (as mention the original application is COSMOS), if you purchase the premium package you will also get the COSMOS GeoStar package (which is the old but solid COSMOS), however as I mentioned Simulation works OK if you know how the application should be used.

                     

                    In our case, we use SolidWorks, because we where users of COSMOS and when SolidWorks purchase COSMOS and integrate it to the CAD we decide to move to SW from Pro-E (since the integration with ANSYS cost a bundle of money).  In my oppinion, for the money COSMOS (Simulation) is the best you can get.

                     

                    On any event, I will advise you to get the 30 Day demo and test the application on some of the work you plan to do, also get a VAR that have an engineer that know FEA on the staff so he can explain you any issues you may need to start using the application (since the "help" and the "tutorials" are not that good), in the old times, when we deal directly with COSMOS we get support form excellent people with knowledge on the application, now I assume those people work for SolidWorks.

                     

                    Take a look on the forums and you will see that there are people that know how to use a FEA system, but also you will see people that are CAD operators, that are trying to use Simulation for some weird things...

                     

                    We use SW simulation for linear static and non linear dynamic, buckling, forced response, fatigue analysis, and thermal and so far is OK.  We also use other systems for large frame structures (such as Risa 3D and ROBOT) as well as Piping Network Flow Analysis Systems, etc.

                     

                    Regarding the multi CPU question... SW do not use multiple CPU's on more that 80% of the process, Simulation work un multiple CPU's while doing the SOLVERS, but while getting data from SW (the CAD model) is limited by SW.  We use DELL Precision T7500 with dual XEON Quad Cores 3.2 Ghz and 18Gb of RAM (x64 OS) and the application work with good performance.

                      • Re: Solidworks Simulation Power & Stability

                        Thanks for your reply. The comments you've made make me feel a little more comfortable with the SW simulation software. I have been looking over the forums and it appears that there are a great deal of people that don't have a basic understanding of FEA and it's various applications. I'm definitely going to take your advice and give the simulation a 30 day trial before I decide to purchase anything.

                        If you have any other experiences don't hesitate to drop me an e-mail. Thanks Again.

                         

                        John

                          • Re: Solidworks Simulation Power & Stability
                            Mauricio Martinez-Saez

                            34830 wrote:

                             

                            Thanks for your reply. The comments you've made make me feel a little more comfortable with the SW simulation software. I have been looking over the forums and it appears that there are a great deal of people that don't have a basic understanding of FEA and it's various applications. I'm definitely going to take your advice and give the simulation a 30 day trial before I decide to purchase anything.

                            If you have any other experiences don't hesitate to drop me an e-mail. Thanks Again.

                             

                            John

                             

                            John,

                             

                            The problem with bugs are mainly on the interphase since COSMOS is going thru the process of being fully integrated into SolidWorks.  The main solvers of COSMOS are very solid and produce valid results (if you do your modeling and define parameters correctly).  When you get Cosmos Simulation Premiun (former COSMOS Advanced Professional) you also get the COSMOS DesignStar and the COSMOSM which are the FEA systems that works OUTSIDE SolidWorks (you can load models from several CAD applications) and which do not have some of the bugs you will encounter while SW people finish the integration of COSMOS into SW (what now is SIMULATION).  COSMOS is an old solid FEA system developed and refined over the years by excellent engineers, COSMOS was developed by Structural Research & Analysis Corporation (SRAC) in 1982, the SRAC was acquired by SolidWorks which initiate the integration of COSMOS into SolidWorks,  the weackness of COSMOS before was lack of modeling functionality (it was a good FEA but you need to model on a separate application such as SW, Pro-E, etc.) the integration into SW resolved (or will resolve) that issue.  As menioned I used COSMOS for years (as well as other FEA systems) and I have no complain.  However to use a FEA system you must have the propper engineering knowledge (normally you need to know how to do the analysis with paper and pencil), I do not see how a  CAD user without formal engineering education can use a FEA system and how can interpret the result of the analysis.

                             

                            M. G. Martinez-Saez

                            Director of Engineering and Product Development

                            IMSA Group.   

                        • Re: Solidworks Simulation Power & Stability
                          Joe Galliera

                          Upfront the SW forums here may seem like a direct place to find out about the software, but where people have issues, you don't know all the background involved with it.  The forums are about solving issues, and another resource that we at SolidWorks offer to help our customers.  I don't recall ever seeing anyone writing on the forums specifically about how great and enjoyable they find using SolidWorks & SW Simulation software.  You might only get bits here and there from people like Mauricio, who have been using the software for a long time, which speaks volumes to its integrity.

                           

                          There are validation problems available directly in the software through the Tutorials dialog.  Also, please find the attached validation guide.

                            • Re: Solidworks Simulation Power & Stability
                              Derek Bishop

                              Thanks Joe,


                              I note you are one of the technical managers working for SolidWorks Simulation. Please advise:


                              1. The extent of testing that is initiated and undertaken by SolidWorks on their Simulation software at the Beta phase?
                              2. Give an approximate percent of the number of problems discovered by users and those discovered by SolidWorks at the Beta testing phase.
                              3. At what point in the development does SolidWorks initiated testing of the software discontinue.

                              When speaking of testing I am referring to attemps to troubleshoot and isolate problems in the software.


                              Thanks in advance.

                                • Re: Solidworks Simulation Power & Stability
                                  Joe Galliera

                                  Derek, I don't know the answers to your questions, because I do not work in the QA department.  But I can offer the attached document that talks about our QA process; although older much if it is unchanged.  I skimmed through it again just now quickly, and the only change I would update to it is the following: instead of "a beta release goes to a limited number of customers who have the COSMOS subscription service," I would say that Beta is available to everyone on active subscription.

                                   

                                  Our QA department is very very good and efficient; we (SRAC and then SolidWorks) have been in business for over 25 years.  I can assure you that our customers are not doing our bug finding for us, most reported bugs  are known issues.  Fixes are not immediate and may require work from several pieces of the code across many developers.  (aside: The question is if you fix one bug do you create more elsewhere in the code?)  Sometimes workarounds are made available and published in the knowledge base.  SolidWorks offers many outlets, such as our VARs and these forums, to help our valuable customers.

                                   

                                  It is my belief that the software testing is continuous and carried out round the clock, as I have seen it running on computers in the office.  While a customer only sees a software build that is a major or minor release, there are many many more builds created internally.

                                    • Re: Solidworks Simulation Power & Stability
                                      Derek Bishop

                                      Thanks Joe. I was involved with Beta testing and sent through around 90 SR's. Many problems were considered critical and blatant  and should have been picked up by SolidWorks if they had a thorought testing regime. That does not seem to be the case. If a Beta tester did not identify these problems who would? This gave rise to the questions I posted above. Could you please relay these to the QA/QC department. I'm seeking direct responses to all of the questions. The article you posted was interesting but did not answer these questions directly. I spoke with my VAR yesterday about the same issue. He was to send the questions to SolidWorks. I would prefer to do this through yourself if possible. Please confirm either way. This will give greater insight into the problems we are encountering with the software and help identify ways around these problems. Thanks in advance.

                                        • Re: Solidworks Simulation Power & Stability
                                          Joe Galliera

                                          You do not understand what a Beta program is then.  Dialog boxes might not work right, because they are considering the layout; help is not active because the document is not completed; or icons don't look right because they are just placeholders  There is also a difference between bugs being fixed and caught.  During the Beta program, it is OK to have a software build that has bugs.  Blantant bugs, as you pointed out, would be high priority bugs and garner the most attention in development.  If these bugs don't get fixed by Pre-release, then they will pull functionality before FCS (First Customer Ship).  A good example is that throughout the 2010 Beta, the automatic scaling of result plots has not worked properly... they are fully aware of this and will be fixed.

                                            • Re: Solidworks Simulation Power & Stability
                                              Derek Bishop

                                              Joe,

                                               

                                              Please confirm that you will forward the above questions to the SolidWorks QA/QC and that they will provide a response to the questions.

                                                • Re: Solidworks Simulation Power & Stability
                                                  Derek Bishop

                                                  Looks like Joe has gone quiet. I'll try again.

                                                   

                                                  Joe,

                                                   

                                                  Please confirm that you will forward the above questions to the SolidWorks QA/QC and that they will provide a response to the questions.

                                                   

                                                  1. The extent of testing that is initiated and undertaken by SolidWorks on their Simulation software at the Beta phase?
                                                  2. Give an approximate percent of the number of problems discovered by users and those discovered by SolidWorks at the Beta testing phase.
                                                  3. At what point in the development does SolidWorks initiated testing of the software discontinue.

                                                    • Re: Solidworks Simulation Power & Stability
                                                      Joe Galliera

                                                      The forum doesn't have a out of office automated reply, so sorry for my delay as we had a holiday here in the US.

                                                       

                                                      The responses come from QA.

                                                      1. There is internal testing going on constantly, both by automated software robots and manual testing.  Testing is done round the clock by robots and people located in Santa Monica, Concord, Romania, Pune, Chennai and Japan.  There are around 1,000 test cases developed for Simulation QA, and 4 to 5,000 for COSMOSM, which has been developed for over 20 years.  Not all combinations used by customers can be tested because there are too many parameters, such as: OS, hardware, language, program options/features, etc.  Only 40% of all SR's (service requests) get converted into unique SPR's (software problem reports), and then those get filed by color depending on severity.

                                                       

                                                      2. The number of SPR's found internally is significantly higher than the number of customer issues.  The Beta program, which is done in parallel to internal testing, does look for feedback from customers regarding new functionality.  Looking for a percentage, it's about 85% found internally and 15% by users during the Beta program.

                                                       

                                                      3. Testing does not stop.  There are multiple builds on a daily basis that need to be tested.  When there is a release (major or service pack), then testing stops for that release only and immediately continues to the next release.

                                                       

                                                      I hope that this answers your questions.

                                                        • Re: Solidworks Simulation Power & Stability
                                                          Derek Bishop

                                                          Yes and no. I appreciate the responses and am heartened that SolidWorks does undertake testing and debugging at the Beta phase. However the question that remains unanswered is the reason for the large number of bugs in the software. I found a  large number of significant bugs during Beta testing. Many remain unresolved. In the first instance I ask why there are so many bugs. In the second instant I wonder how long these will take to be resolved. Time will tell.

                                                           

                                                          In the example discussed above regarding defining material properties I noted that there were a number of bugs that were fixed in one of the service packs. In spite of that I still found others relating to the same issue that remained unresolved. That could have lead to serious error in a calculation. Why wasn't this discovered and fixed at the same time as the others. Where there are that many problems in an important feature of the software then SolidWorks needs to do a very thorough review of the whole feature and get it working properly. That evidently does not happen. Instead it took me several weeks of trying to pin point the problem and dragging SolidWorks by the nose to get it to agree that there was a problem.


                                                          • Re: Solidworks Simulation Power & Stability
                                                            Mauricio Martinez-Saez

                                                            Joe,

                                                             

                                                            After all that testing....

                                                             

                                                            "...internal testing going on constantly, both by automated software robots and manual testing.  Testing is done round the clock by robots and people located in Santa Monica, Concord, Romania, Pune, Chennai and Japan.  There are around 1,000 test cases developed for Simulation QA, and 4 to 5,000 for COSMOSM,"

                                                             

                                                            How it is possible that in several ocassions SW released complete versions that fail to install correctly (do to "bugs" or errors on the installer program),  I fail to understand how it is possible to burn a large quantity of DVD's, pack them and send it to the field without detecting that the programs on the DVD do not install correctly.  Do enyone of the people on those many testing locations ever tried to install the software on a computer?  The following document this statements:

                                                             

                                                            1.- Look at Solution ID S-015632 related to installartion problem on 2007 version

                                                             

                                                            2.- Perform a serach on the knowledge base for "failed to install" and / or for "installation errors"

                                                             

                                                            How it is possible that SPR# 480083 was only fixed on 2009 SP 4.1, when from SP 0 up to SP 4.1 SW crash everytime the user try to open 3D ContentCentral from within the application if Internet Explorer v 8 is installed on the same computer.  So some one test the 2009 version and tryed to access 3D ContentCentral (a site promoted by SW) with a computer having updated versions of Windows before releasing the 2009 version.

                                                             

                                                            The aforementioned are way more basic "bugs" that problems on calculations of very complex solvers inside the Simulation package, I can understand that bugs on the code of Simulation will take time to be detected and fixed, but it is difficult to understant why basic issues on installation and simple functionality get undetected and not fixed for several SP's.

                                                             

                                                            I really like SW, it is one of the better softwares I ever used, but it will be a lot better if QA do a better job (or maybe if SW do not try to add so many new "toys" to the application before being sure that the existing functionality is Solid and Working properly, perhaps this is what makes the job of QA to be a "mission impossible").