CIMdata PLM Industry Summary Online Archive
February 2012
CIMdata News
“Simulation-based Design: PLM’s New Foundation”, By Keith Meintjes, Simulation & Analysis Practice Manager at CIMdata Inc.
The current capability for simulation-based design (SBD) presents great opportunities to quickly engineer new products. SBD enables designers to evaluate the performance and function of new products directly from digital expressions of the physics and of functional requirements. For the first time, engineers can start a design—points, lines, angles, numbers, and all the rest of it—working directly from specifications and requirements, without having to resort to trial-and-error prototype testing and development.
This is computer-aided engineering (CAE), which complements computer-aided design (CAD). It is giving a powerful push to collaborative product development but there are new issues to be managed. SBD can result in an explosion of data with an unprecedented variety and complexity.
SBD is rapidly becoming embedded in new-product engineering from concept development to detailed design, then on through product integration to manufacturing, “mechatronics” (embedded controls), ergonomics, and assembly. SBD has also penetrated purchasing, field service, marketing, safety, sustainability (economic as well as ecological), regulatory compliance, traceability and much else.
Dassault Systèmes’ SIMULIA Version 6 DesignSight product suite guides designers through a series of actions to prepare the model, run the simulation, and view the results. The designer interacts directly with the design to define load location, orientation. Courtesy Dassault Systèmes’
In the words of Louis Sullivan, “Form ever follows function. This is the law.” In mechanical design, CAD describes the form. CAE simulation evaluates (or predicts) the function. SBD is a cornerstone of the model-based enterprise or MBE, and we see that, indeed, form follows function. Simulation requires (and generates) far more data than is simply required to describe the product’s shape and geometry in a CAD system.
Peter Bilello, president of CIMdata, frequently points out that, “thanks to SBD, far more information about the development of new products is available now than ever before. But compared to CAD and the very basic information-handling approach of product data management or PDM,” he cautions, “what is happening now presents significant challenges to engineering.”
He adds, “The complexity of SBD and its volume of data have the potential to stymie collaboration among engineers in new-product development. The good news is that collaboration itself offers solutions to SBD’s challenges. This is yet another reason why collaborative product development is the most important of all sustainable competitive advantages,” Bilello says. He adds that SBD is a vital component of product lifecycle management (PLM).
The auto-focus lens undergoing CAE analysis, in this case for shock loading and vibration. Courtesy SpaceClaim Corp.
The word “simulate” is of primary interest here. Simulation determines how well a new product will meet its design criteria, how well the capabilities will match the demands of the intended customers. Because SBD starts with simulation, it offers unprecedented opportunities to quickly engineer new products. SBD does this by changing the traditional CAD-first, then-CAE paradigm for new-product development. In the traditional approach, CAD operators define geometry for release to manufacturing. Prototype testing and simulation then evaluate the (completed) design’s performance, concurrently with its preparation for manufacturing. If a performance issue is found it results in disruptive design changes, which can be very expensive if manufacturing commitments have been made.
There are two key technology enablers for SBD. One is the ever-increasing speed of computers, combined with very capable software. Four decades ago, it would have taken 41 months to complete the engineering calculations that today can be done in one second. The other enabler is technologies to easily and rapidly create and modify geometry suitable for simulation. In particular, “direct modeling,” a CAD technology, frees simulation from dependence on CAD specialists to provide geometry. This potentially brings geometry inside the process where simulation is used to find optimum designs.
SBD enables better, more robust designs in less time and at less cost. In the engineering of manufactured goods, this part of the digital revolution is as compelling as Moore’s Law in electronics. Direct modeling systems are now offered by various providers, and are often integrated or coupled with simulation applications.
Creo Simulate- Applying Loads and Constraints. Courtesy of PTC
Every revolution, of course, has its “however.” To name just three for SBD:
- Many different product representations are required for simulation, depending on the physics involved; this makes the data management problem for CAE much more complex than for CAD.
- Using simulation to optimize designs and to analyze robustness to variation—in manufacturing tolerances or material properties, for example—can result in thousands of simulations rather than just one.
- MBE is an exercise in optimization across multiple functions, for example the cost, weight, strength, and durability tradeoffs involved in material choice.
CIMdata strongly believes Systems Engineering principles must be used to manage SBD, and that systems engineering activities like requirements management must be much more closely coupled to design and simulation than they have been in the past.
All the foregoing has big implications for collaborative product development, which is a key element in CIMdata’s business. CIMdata is the leader in market analysis and consulting to implement enterprise-level strategies, PLM in particular.
Visualization of CAE results using lightweight formats such as JT directly within the data management environment enables teams to collaborate more effectively for better decision making. Courtesy of Siemens PLM Software
Silos: Frustrations and Opportunities
Those who have been frustrated with implementing enterprise-level strategies often blame organization charts. “Org charts” are said to be littered with silos of expertise that do not or cannot or simply will not “talk to each other.” Blaming high-level systems failures on a supposed failure to communicate between specialized engineering units is unfair to the engineers in those silos, and unduly harsh to the managers who run the silos.
This lack of comprehension is the real cause of so many organizations’ inability to agree on how best to handle the explosion of data that they all share, sometimes painfully. Working with the silos—giving them the tools, systems and strategies they need— rather than approaching them as if they were the problems will ensure that silos are part of the solution. Resistance to change is real. Knowledgeable, persistent pushback by innovators is the only effective antidote.
The real challenge is the near total lack of standardization in the explosion in digital information that cascades through, and between, the silos. This calls for an in-depth understanding of the multitude of CAE formats and their specific purposes plus an intense focus on the flows of data within and between engineering units (silos).
Proven approaches include the venerable practice of data modeling and, in particular, “commonizing”—focusing on what SBD information and data have in common. Focusing on the differences, the usual approach, only leads to frustration. As a part of data modeling, commonizing and related technologies can mature into enablers for collaboration.
It must be considered that different functions or silos require different information, and that only some of their information is required to be shared to another silo. An automobile windshield provides a good example. The information required to manufacture the windshield is generally different than that required to ship it or to assemble it on to a vehicle, though the data required certainly has overlaps.
Different performance simulations require different representations of the windshield. For the aerodynamics of the car, the exterior surface is required. For occupant comfort, the interior surface is required. The windshield is also an important element in vehicle structural stiffness, where it is modeled as a sheet, meaning that the average of the interior and exterior surfaces is required. Other simulations for occupant safety, ergonomics, solar load, etc., all have unique requirements to describe the windshield.
There are proven approaches to overcome potential snags:
- Understanding why data is created, not just how, and who needs to reuse it.
- Using metadata, or data about data, to keep track of databases scattered across most of the planet’s time zones. Automakers have estimated that the engineering data for any new vehicle resides on as many as 20,000 CPUs, which run as many as 200 different CAE solvers.
- Determining whether FEA test and simulation data is to be kept in its “raw” form, as uncertainties may be introduced by pre-processing and post-processing.
- Deciding how best to handle specific file types. A file containing a first step in an analysis is reused very differently from an (finite element analysis) FEA results file.
- Recognizing why materials data is a source of variability equal in importance to part geometry. IT experts estimate that engineering-oriented companies have databases of hundreds of materials. Some have hundreds of attributes that are lifecycle-managed through dozens of versions.
- Determining levels of granularity, which embodies the tradeoffs between file size and relevant detail.
- Incorporating any custom translators developed among adjacent silos and departments, provided their users understand them.
- Acknowledging that software vendors contribute to the collaboration problem by continually segmenting their offerings for new organizational needs structures.
- Using lightweight file formats such as 3D PDFs for data redistribution.
- Managing engineering workflows with flexible templates.
- Stifling extraneous “phantom” requirements that complicate workflows, slow down data transfers, or deter nervous users.
These challenges are not trivial and, taken altogether, can be daunting. But they are not insurmountable, either. Successful implementations are regularly achieved by determined, focused teams. The best teams contain users, software vendors, and experts from outside the company as well as from IT.
From our experience, such a team would start with what we call the shared-drives issue. Engineering information is scattered across hundreds of disk drives in dozens of locations. Effective collaboration requires that these drives and their contents are known. Nearly all new-product information is created on drives “shared” by engineering units and, yes, silos.
In CIMdata’s opinion, it is unrealistic to assume or plan that all engineering data will be consolidated in a single database. The data is distributed, much of it in functional silos. It is important to understand which data needs to be shared, and to come up with strategies to share information as required. OSLC (http://open-services.net/) is one such initiative.
The need for, and success of, SBD is driven by tough global competition, increasing regulatory scrutiny, concerns about environmental sustainability and “greenness,” excessive energy use, and transparency in supply networks. Nothing new there. Of the many impacts these drivers have on collaborative product development, two stand out:
- More people than ever have a hand in product development but many have little or no technical background, engineering or otherwise. The presence of all these people and their organizations/silos of expertise is often misrepresented as the big org-chart speed bump in the road to implementing enterprise-level solutions. These people are in fact the opportunity.
- Collaboration in product development is more important than ever—even as it becomes more difficult, partly as a consequence of widespread SBD adoption. This difficulty is addressed by commonizing data to minimize redrawing and re-creating as data is reused across the enterprise.
To come back to Peter Bilello, “meeting the data commonizing challenges can resolve many of the interoperability problems among engineers’ tools,” he points out. “This can provide incentives for engineers anywhere in the organization to buy into enterprise-level solutions. The benefits to future products, both near-term and long-range, will be well worth the effort.”
Clarifying the terminology:
S&A is a core process of new-product development and manufacturing, and CAE applications are its tools.
SBD is a goal toward which most engineering organizations are striving. PLM builds upon SBD (and CAD, CAE, and PDM) to create enterprise-level strategies for managing an organization’s intellectual assets. PLM unites people, processes and technology.
Keith Meintjes
Keith Meintjes is the Simulation & Analysis Practice Manager at CIMdata. He has over 30 years of experience in the development of simulation tools and in their application to transform product development. His career spans academia, industry, and consulting. He holds B.Sc. and M.Sc. degrees in Mechanical Engineering / Aeronautics from the University of the Witwatersrand, Johannesburg, South Africa. He also holds an M.A. and a Ph.D. from Princeton University.
Become a member of the CIMdata PLM Community to receive your daily PLM news and much more.
Tell us what you think of the CIMdata Newsletter. Send your feedback.
CIMdata is committed to your privacy. Your personal information will never be sold or shared outside of CIMdata without your express permission.