Do We Need To Manage CAE Data in PLM?

Do We Need To Manage CAE Data in PLM?

I want to discuss CAE today. CAE is considered as a technology to support engineers in tasks such as analysis, simulation, design, manufacture, planning, diagnosis, and repair. CAE considered as a mainstream technology applied for design and manufacturing. In my view, CAE is one of the major sources of engineering information in the organization.

CAE coverage includes:

  • Stress analyzes
  • FEA
  • CDF
  • Kinematics
  • Event simulations
  • Process simulations

However, the topic I want to discuss today is an intersection of CAE and PLM. To be more specific about what system supposed to manage CAE related data? In my view, this topic is heavily missed in most of PDM and PLM systems. Most of them are completely focused on CAD models, drawings and, lately Bill of Materials. At the same time, CAE, as a massive producer of engineering information about a product is not covered. What I found that high-end CAD systems have a more dedication to this topic, but still not enough, in my view.
What are the aspects of CAE that need to be covered by PLM as a data management platform:

  • Keep references on results of calculations and analyze. It will allow to have all relevant product records in one place.
  • To support variation of product data in case of work on alternative designs, configurations, etc.
  • Keep the historical record of analyzes for audit, regulatory and solving customer’s critical situations.

I found these capabilities undervalued and will be interested to hear about your experience and opinion.

Best, Oleg

Share

Share

Share This Post

  • Jim Stawn

    One challenge is that PLM Systems are all driven on Part Numbers, yet CAE data doesn’t always translate back to a specific part number. Sometimes you are analyzing portions of assemblies, other times you are analyzing portions of single parts or simplified versions of those parts. How do you relate an analysis to geometry that doesn’t really represent a real part in the PLM system?

  • Jim, this is a very good point, thank you! Most of the PDM/PLM systems today, in my view, are driven by model where if you don’t have Doc Id or Part#, you’ll not be able to manage, attach, keep data. And, for CAE, most of the information is not associated with “released drawings”. Those various models, are used for various analyzes but didn’t come to the “single product record”. What I see, as an option, is to expand PLM model to support it. However, I’m interested to learn existing experience in doing so. Best, Oleg

  • Oleg,

    this is a great topic and I’m happy you brought this up. Indeed it’s a complex issue to sync CAD, CAE and Test in a single environment. They all have a different pace but still interact in terms of geometrical, material or assembly information. CAE is often used upfront in design studies or for manufacturing simulation in a later stage. Moreover, there is a huge amount of data to be managed which requires a different architecture.

    I think intelligent use of CAE could even drive the PLM process upfront, the CAD data evolves and information gets syncronized also with test data over the course of a development projec.

    From my point of view, the largest value for customers is an intelligent connection of PDM with special CAE and test data management systems.

  • Dave

    Oleg,
    I believe that the management of Analysis data within the PLM world is actually quite well defined and that DS and Siemens are leading this trend. There are many significant EU funded projects which are beginning to take this to the next level (example – Crescendo Project – Model the behavioural digital aircraft) .. and others .. For me its a simple fact that Analysis data is a huge factor in making decisions on many aspects of the product lifecycle .. for that reason alone it makes complete sense to manage the data and thus simplify the decision making process and provide better traceability.

  • Vladimir

    Oleg,

    what is your plain PLM definition? Why am I asking this? CAE is one of the pillar of PLM, so the question from your headline is a bit confusing for me.

  • Hello

    It is a big topic. I know different initiatives, either departmental pdm apps, either enterprise wide plm apps, both having not the same scope. Otherwise all domains have not the same constraints and regulations.
    One fact to take into account is that the size of files managed by CAE people, which most often are much higher than CAD files or office files, for which plm apps have been designed. This prevent usually to use a mainstream PLM application to manage such data. Otherwise, CAE files comes from very specialized people, and there is usually no need to share a nastran deck or LS-DYNA result file outside the CAE community.
    So most often, there is a study to start with the company, in order to analyse and define precisely what is required, what are the challenges, and so on.

    But you are true regarding product development process, CAE activity is clearly a major facet of PLM, and PLM apps could imagine manage input and output data between departments.

  • The main hurdle I see is you are asking user to do something that has no value in the present time. The value is all based on future time. Therefore you have a massive habit change problem. It MIGHT be great to have all these files (assumes I know what they were looking for and that I can make sense of the files and that the software I have in the future can use these files) in the future in case we need to do it over again or change something… But this means I am working now under the impression that what I am doing is flawed and will need to be done over again. Certainly this is not what we are thinking when we design. So again we have a real habit change hurdle in our way.

    From another perspective, what if we did not have the actual data but we did have a list of what was done? Is this enough? Do we really need all the files? Say a year from now I am changing something and I could see that extensive analysis was done on part A and I could see who did it. Isn’t this enough? What if I could see that the analysis that was done resolved these 5 issues? What if I could see that a year ago the design team had to resolve these 50 issues and had assigned these 100 tasks. Isn’t this better than having the resulting files? PDM/PLM knows allot about A, B and C but has no idea about A-B and B-C, what happened between versions. The effort/work done is more valuable than the results! This same idea is behind all the changes the FDA has driven into their regulations. Knowing how I got from A to B is more important than know A & B and I would argue the files do not really matter.

  • Vladimir

    François,

    I don’t think that it’s necessary to manage all CAE data as they (I agree with you) are pretty huge and even it doesn’t make sense as they are very special by origin i.e not easy “readable” by common consumer of PLM data. As I understood from this blog it’s going about the results and I believe these can be generated by CAE tool (integrated or external) in “report form” and such it should be managed by PLM system if needed.

  • Shekhar B

    Oleg,

    In its present form, the PLM system does not do much help other than just being a central location for keeping the documents for inputs and results linked with the design models/configurations. The CAE process itself creates huge amount of data that cannot be managed in PLM system.

    In future will the PLM system provide the capability to have CAE process and data structures such that one can
    a)Perform analytics and derive intelligence across several versions of products/parts
    b)Extract–>Recreate–>Modify–>Combine–>Solve analysis cases from the data available in the PLM system.
    c)Be the process controller/manager that will drive the job on high performance computing clusters???

    Unless PLM comes out of its “document centric” approach to a “data centric” approach it will be impossible to meet the CAE needs and the question is should PLM really overload itself by doing that?

  • Oleg
    Good Topic!
    In my mind, PLM should be the single location of all product data. be it MRD/PRD/CAE/Design Reviews/Test Data/Reliability and Quality.

    that said, one of the comments was that this data/document may not be related to a part #…The beauty of document management systems within PLM is the ability to cross link documents and part #. this way you can associate documents to part # using a reference link so that all the relevant information is always connected and visible to the enterprise.

    In closing, yes all product relevant data should be in PLM. Why manage Silos and the pain of integration and looking up multiple sites?

  • Thomas, I agree with you. Today’s PDM/PLM architectures are not capable of manage CAE-related information. For most of the cases, I see it is limited to manage files produced by CAE-systems. However, it seems to be a very limited value of such a file-oriented solutions, for the moment. Thanks for commenting! Oleg

  • David, Thanks for the comment. Yes, analysis data started to come to PLM products of DS and SPLM. You are right about decision management. This is the key! Best, Oleg

  • Furthermore, the value is not in the management of data but in the management of content. A managed file silo is of no added value but intelligent process automation, reuse of data, material management etc. immediatally supports the user and rise acceptance. If you manage the content in a smart way you can learn from past experience and improve your future designs.

  • Vladimir, I understand your point. PLM has broad definition and may include CAE. This is a more marketing view. For me PLM is more about the ability to manage data that related to the complete product lifecycle (from requirements and dreams and going to disposal instructions)… Does it make sense? Best, Oleg

  • Jim Stawn

    Besides being part number based, most PLM systems are also Product Structure base. When you start lumping in simulations such as Kinematics and Mechatronics, the product structure that you use is definitely not the same as the manufacturing structure. For example, we install a number of vendor components such as actuators. From a manufacturing perspective, this is one part number. But to simulate it realistically, the CAx sytstem needs to model the actuator as an assembly of several parts (Housing, Pushrod, Rod End, etc). Thus the simulation product structure is not the same as the manufacturing structure.

    As far as file sizes, this is also a problem. As these analysis compute, the user is going to want the files locally. I don’t care how large and fast a server you have, or how fast your network is, it will never be as fast as a SATA drive on your local workstation. Since these analysis operations are constantaly reading and writing to the disk several thousands of times a second, the milliseconds add up very rapidly.

    Finally, as for Chris’s comment that we are assuming that “I am doing is flawed and will need to be done over again.”, I have two words for you – derivative design. We are constantly designing new parts, components, assemblies, and products that are similar to, but different from previous designs. If we have a previous simulation model to start with, it is very easy to adapt to the new configuration. If we didn’t save it, we would have to start over.

  • François, Thank you for clarification. You are right. However, here is the thing. CAE has a tendency to be in the shadow of design and CAD when it comes to the data management. You mentioned right reasons why it is complicated to manage CAE data – file size, various domains, etc. This is all right! However, role of this information in development is hugely undervalued. And, in my view, it is partially because of people. Hiding data is one of the diseases of the enterprises. And this is true for CAE too. If tomorrow, we will have an easier way to connect CAE to development process, everybody will be benefited. Best, Oleg

  • Chirs, Thanks! I think you are bringing a very important point by trying to analyze what is really important? Results of the analyzes? Huge files? Record of what was done? I think ALL is important. Storage is cheap. However, we are not managing it. That’s was my point. Does it make sense? Good to have you on PLM Think Tank, as usual! Best, Oleg

  • Vladimir, You got me right (see also my answer to Chris just above). We need to decide what to manage. However, when storage is cheap, sometimes the decision is more complicated. My main point is that PLM needs to work on better data management solutions to improve the way to manage CAE data, which is different from regular PDM-like file/part/version records. Thanks for comment! Best, Oleg

  • Shekhar, Agree with you. PLM is not capable today to manage CAE data, since this is much beyond just pointing on data files. I think PLM moved to data orientation (from original EDM/PDM approaches). However, this is still in infancy. Thanks for commenting! Best, Oleg.

  • Swati, Agree on the point that all data need to be synchronized. It makes a lot of sense. However, this is not happening. Why? The complexity of the integration is beyond the normal level. Companies are not embarking on “integration projects” and afraid solution is either “not reliable” or “too expensive”. Since PLM has a special capability to touch CAD related data, the integration point into PLM can be preferable. Thanks for your comments! Best, Oleg

  • Thomas, Yes, you are right. There is no sense to manage “just files”. Information (or content) is the way it becomes re-usable. Thanks! Oleg

  • Jim, Thanks! I think you made great clarification on how things work from your perspective. It convinced me repeatedly, that PLM capable of manage CAE data/files need to be different from what we have today and this is a topic to think about. However, with regards to SATA files, I tend to disagree. You don’t need to access them from network. If you have a full history, you can upload and work on them. I assume you don’t need to do it 10 times during the day… right? Best, Oleg

  • Hi Oleg.. Feels like an extension of our conversation at #sww10. Fact is, I don’t believe it has much to do with file size, file types. It has more to do with the perceived value of simulation driven design.

    I am a massive advocate of upfront CAE (CFD,in particular) and have spoken to literally 1000s of engineers over the years. It is still not a “critical path” item universally. Until it is, tracking/capturing it via PLM won’t be mainsteam.

    Engineering orgs tend to lean toward the known path, simulation just isn’t mainstream enough for the data to be a “problem”. Once it is, capturing the info is just mechanics at that point. We will define a std, link to files in the “cloud” and move on.

    Morale here, once its considered vital, the details of tracking the info is simple.

  • Derrek, You are right! Thanks for your comments! What do you think should happen to make usage of CAE in mainstream? Best, Oleg

  • Oleg, probably I am too late to comment on this topic. Nevertheless, I believe if we can figure out the way to abstract the decision points from the huge CAE simulation result data, we would be moving in right direction to manage the real valuable information.
    Just my two cents!

  • Lokendra, Thanks! You cannot be late – thanks for putting your comment. I agree, if we can start to extract a somewhat can be used for decision, it can be very beneficial. Today it is considered like a bulk of data and nobody, except engineers cannot handle it. Best, Oleg

  • Darcy Parker

    From a PLM perspective, CAE models and their results are useful for Design Verification. A part may have a number of requirements allocated to it. During the product’s development life cycle, it may be necessary to test whether or not the part satisfies the requirements allocated to it. In many cases CAE is an appropriate tool for performing this verification. Each requirement will have an associated verification plan. And each verification test will be performed on some version of the part and will follow the verification plan. This scenario is a case where the CAE results data must be associated with the part in PLM. And the CAE results should be associated with the CAE models that were created for the verification plan. And the verification plan should be associated with the requirement(s) the plan intends to verify. The verification plan may also be indirectly associated with the part because the verification plan may require apparatus or CAE models that reference aspects that are unique to the design. (ie a verification plan may require slightly different apparatus for each concept/design developed for the requirements.) There are numerous relationships that need to be tracked in the PLM system when considering the need for requirements management and design verification to demonstrate a product satisfies the requirements.

    Another scenario is when CAE tools are used for optimization/tuning of parameters during Concept Development. The CAE models developed for this optimization effort are associated with the CAD models. In this case, the CAE models may not need to be directly associated with the parts that the CAD models describe. But there is definitely a set of relationships between the CAD models and CAE models that needs to be tracked in the PLM system.

    I agree that PLM systems often neglect the need to track these important relationships to CAE related content (models and results). But I am excited by PTC’s RequirementsLink’s traceability link functionality. Different types of traceability links can be configured to track these relationships in a PLM system.

  • Darcy, Thanks for these interesting examples! I will take a look on PTC Requirements Link. However, I’m sure, situation when such relationships managed and tracked are more exceptional… (unfortunately). Best, Oleg

Subscription Options:

Subscribe via RSS

Popular Posts

Recent Posts

Archives

Blogroll