Autodesk Manufacturing Tech Day – Digital Simulation

Although I am a Mechanical Engineer by profession and understand the geek speak involved in simulation and analysis, I almost never write about that side of the design process. Somehow I have always concentrated more on the geometry side of things. Probably because that is linked directly to my day job of writing software that creates and edits geometry. However, I intend to change that. Going forward, I would like to spend some time with analysis and simulation software. Although simulation and analysis are sciences in their own right for which you need to receive formal training in, the software these days does make things a bit easier for people not trained the art to do simulation and analysis. This was the subject of one of the presentations called “Digital Simulation” given by Jeff Wymer, Senior Product Line Manager, Digital Simulation and Hilde Sevens, Director of Product Management, Digital Factory and who were assisted by a couple of people in technical marketing.

According to Jeff, the main reasons why CAD users shy away from adopting Digital Simulation are:

1) Ease of Use: Due to the complex nature of simulation itself simulation software is difficult to use. For example, in the case of Finite Element Analysis, you need to bother yourself with meshing and the parameters that you need to adjust to control it.

2) Interoperability: Since simulation software is complex, it usually sits in a specialized application and transferring data back and forth with a CAD system while maintaining associativity can sometimes become a problem.

3) Cost of Ownership: Needless to say, due to its sheer nature, simulation software is expensive.

The real effect of Autodesk’s acquisitions of Algor, MoldFlow and other similar analysis and simulation software vendors is quite evident in this years releases of products. The whole point of the presentation was to highlight how Autodesk was trying to address the exact three barriers to adoption mentioned above. Here are a few pictures showing how Autodesk customers now have a wide range of options to choose from whether they want to do initial low level simulation right inside Inventor or accurate (as accurate simulation can be, that is) high level simulation in Algor. The key point to note is the efficient way by which rich data flows between the various products involved. Click the images for larger views.

This is an assembly being analyzed right inside of Autodesk Inventor Professional. Meshing happens automatically depending upon the various factors.

This assembly was simply “pushed” into Autodesk Algor Professional. There was no need to set up the model once again in Algor since Inventor now shares the same material library as Algor.

The frame analysis tool has videos embedded right into the ribbon that shows the user how the apply constraints and set various parameters. Also there a number of simulation guides built into Inventor that act like wizards which help new users set up constraints, boundary conditions, loads, etc.

Direct Manipulation has found its place in Simulation as well. Here you see a load being positioned and oriented in the graphics window while at the same time setting up its parameters as well. Autodesk has gone to lengths to offer a superior user experience across their product line.

Even Inventor Fusion has been invited to the party. This is a part imported into Fusion and set up to be analyzed in Algor (see below).

The Fusion part being analyzed in Algor after which a change in geometry was easily carried out back in Fusion using a simple pusl/pull command.

I find the last two pictures particularly interesting because although Autodesk is being tight lipped about the plans that they have for the standalone Fusion application after the technology has been embedded into Inventor, it is not very difficult to see that they may do exactly what Spaceclaim is doing with ANSYS. After all Fusion is a direct modeler like SpaceClaim and can be pitted directly against the ANSYS SpaceClaim Direct Modeler.

  • Those are most definitely not the main reasons why CAD users shy away from adopting Digital Simulation.

    Ease of use is the most important of those 3, and a real factor.

    Interoperability is important, but easily trumped by ease-of-use. And please note that embedding a CAE tool directly into a CAD tool is not necessarily better than a high-quality linkage between two best-in-class tools. (Or vice versa, to be fair.)

    Cost of Ownership: This means nothing to the end users of CAD software. It might give management pause, but not if they are given a clear value proposition– a recent product failure that caused 6 months of redesign and testing (and $500k in lost profits due to the delay), for example. These guys would spend $300k on a new physical lab for prototype testing… $5, $10, or $30k per license for a CAE solution isn't shocking.

    Nope… your two biggest barriers to Designers adopting CAE are:
    – Lack of CAE milestones on the left side of project GANTT charts
    – Infrequent usage because the designer/engineer has many different responsibilities and may only have occasion to do CAE at the beginning of a project… maybe with 3-9 months of non-use in the interim… therefore losing confidence in their skills at it… or forgetting it's an option entirely.

  • Deelip,

    As someone said on twitter a few months ago “FEA is the new CAD”. I do think it would be good to start looking into it mainly because a lot of CAD folks are doing this.

    Al Dean had a similar post in Develop3D also talking about the barriers to entry:

    I think ease-of-use WAS a problem and may still be for high end problems. However, I think most people doing low to mid-range FEA with the modern tools would not flag this as an issue.

    Kind of like I commented on Al's post, I think the real problem is “if I do this, will I understand the results?”


  • Deelip,

    some few comments too…
    Easy-of-use: I agree with Jeff this is very important point. It's not only about special parameters, mesh or boundary conditions, it's about the whole thinking. FE analyst needs to convert the real (physical or CAD model) to FE model and sometimes to get the good results it's necessary to make some tricks as the FE model is more abstract to compare with real one.
    Interoperability or integration to CAD: For designer this also is important because CAE modeling capabilities are a bit different than designer is used to be in CAD. On the other hand integrated FE package might be weaker in solver part or postprocessing (but not necessarily).
    And finally costs: already mentioned… these packages are quite expensive, but generally CAE tools are able to save a lot of money in early stage of design when properly used.


  • MC

    I agree with some of Al Dean’s comments from the link provided by “burhop”. Although price is probably a barrier to some, I believe the biggest barrier is a lack of education and training.

    CAD and CAE are design and engineering tools. In the United States, I believe most Mechanical Engineering programs are excellent at teaching design and engineering theory, but they are terribly deficient at teaching future Mechanical Engineers how to apply the theory through the use of these tools (I experienced this first hand with the Mechanical Engineers I worked with and whom I supported on the phone). Most of these Mechanical Engineering programs require one basic design class where the students are introduced to 2D and 3D CAD. They are then expected to apply their very limited modeling/drafting knowledge in many other classes. (In my experience, it usually takes a person one to two years of extensive training and practice to gain the skill to be able to create almost any type of model). Therefore, many U.S. mechanical engineering students don’t even gain a very solid foundation in modeling. On top of that, many colleges and universities provide FEA analysis classes, but these classes are usually not required for graduation. So, with few exceptions, many United States BSME graduates are sent into the work force with very little modeling knowledge and usually no analysis knowledge. The only exception I have seen to this in the U.S. is in the Mechanical Engineering Technology programs. These students don’t receive as much theory education, but are required to complete a number of CAD classes and usually at least one FEA analysis class.

    Most of the MEs I have encountered would have no problem using the modern analysis software. They already know the theory, they just need the training.

  • Decisions made during the early stages of design have the most impact on the final cost and performance of any product.

    CAE that simulates and validates critical engineering parameters, assumptions, and design strategy should best be done before CAD.

  • Sustainable Mind

    FEA is the new CAD, just like heteroscedasticity is the new calculus, and just like Toyota is the new Yugo.

  • Pingback: Inventor Fusion For Engineers |