Leveraging 3D CAD Data
by Evan Yares, Senior Editor
In the 2013 State of 3D Collaboration and Interoperability Report, published last May, Chad Jackson, Principal Analyst at Lifecycle Insights, asked the question, “Have we finally realized the vision of fully leveraging the 3D model?”
In the survey on which the report was based, respondents were asked about their use of 3D, both on the critical path (within engineering and manufacturing), and off the critical path (in service, quality, training, technical docs, marketing, sales, and other areas.) The results show that the use of 3D on the critical path is stronger than its use off the critical path. Though the use of 3D in downstream processes is growing, it’s pretty clear that relatively few companies are even close to fully leveraging their 3D CAD models.
To understand why industry is where it is with respect to the use of 3D, it might help to step back a bit, into the history of CAD.
While 3D CAD systems have been around nearly as long as 2D, for the first 25 years of the CAD era, 2D was dominant. It wasn’t until late 1980s, with the introduction of Pro/ENGINEER—the first commercially successful parametric solid modeling CAD system—that 3D really came into its own.
With Pro/E, designers could first create 3D models, then quickly and easily create associative 2D drawings. When they modified the 3D models, the 2D drawings were automatically updated. Pro/E provided the benefits of 3D, without forcing its users to abandon their long entrenched 2D drawing centric product development processes.
Pro/E was incredibly influential, and most competitive CAD systems (with the exception of those designed primarily for aesthetic surface design) adopted a similar approach, coupling feature-based 3D solid modeling with associative 2D drawing creation and annotation.
Throughout the 1990s, it was pretty common for designers to use 3D model data in downstream processes. Yet, it was not convenient, requiring use of 3D models, as well as annotation information from drawings.
As 3D CAD systems matured, their developers started including model annotation tools, so designers could add information needed for manufacturing (such as dimensions, tolerances, assembly notes, and so on) directly to 3D models. CATIA, Unigraphics, I-DES, and Pro/E each had their own proprietary tools of this type. The problem was that they all worked differently, and weren’t compatible with each other. With no recognized standard methods for creating model annotations, most companies continued to use 2D drawings for conveying and maintaining manufacturing information.
In 1997, driven largely by the aerospace and defense industry and the DoD, ASME started work on what would become the Y14.41-2003 standard for Digital Product Definition Data Practices. The objective of the standard was to support the use of either model plus drawing, or model alone, as a complete product specification. Y14.41 was the first of a group of standards that collectively defined what has become known as product and manufacturing information (PMI).
Today, there are at least five different ways that companies can use CAD:
- 2D drawing as authority (full definition in drawing, no 3D model)
- 2.2D drawing as authority + 3D model (full definition in drawing, model not distributed)
- 3.2D drawing + 3D model together as authority (partial definition in both)
- 4.2D drawing as authority + 3D model as authority (full definition in model, either full or partial definition in drawing, automatically generated from model)
- 5.3D model as authority (full definition in model, no 2D drawing)
The first classification here represents traditional 2D computer-aided drafting. The second represents the lowest level of 3D CAD. The third classification is the most common way datasets are structured in many mainstream manufacturing applications today. The fourth and fifth classifications represent how datasets are structured in organizations pursuing model-based definition (MBD) initiatives—particularly in aerospace enterprises.
Model-based definition is one of a number of “model-based” initiatives, all of which have as their foundation the concept of using semantically rich models to represent the functional characteristics of a product. (Among the list of interesting model-based initiatives are model-based development and model-based design, both of which share the MBD acronym with model-based definition, but neither of which are related to model-based definition. They are initiatives related to the design of complex software systems and control systems, respectively.)
The concept of MBD grew out of a big vision concept called model-based engineering, and was itself, at one time, a very big vision concept. Fortunately, sane minds prevailed, and people involved with the MBD initiative focused their energies on making it pragmatic rather than aspirational. They focused on smartening-up 3D model data, to enable downstream usability. An MBD product model is not hard to understand:
It’s a combination of 3D geometry and PMI (including explicitly defined dimensions, tolerances, notes, GD&T, welding symbols, surface texture symbols, and associated data.)
The PMI is semantic (readable by either humans, or computer programs), associative to the 3D geometry, and standards-based. Any number of views of the model can be composed, detailed, and annotated for specific downstream operations.
While MBD is often thought of as mostly being about “getting rid of drawings,” it’s really more about getting rid of the need to use drawings for things they’re not good for. In an MBD context, it’s still possible to create drawings in the traditional way, but MBD offers a better way: it’s exceptionally easy to create drawings that are 2D projections of the views included in the product model. These are generally simpler than traditional drawings, but they have the benefit of being 100% consistent with the model. There is never a question of which document is correct—the drawing or the model.
The process of implementing MBD in an organization starts with one important first step: Getting CAD software that supports PMI. Unfortunately, this first step also introduces one of the issues that’s held the acceptance of MBD back: lack of compatibility. While CATIA, NX, Creo, Solid Edge, SpaceClaim, and SolidWorks all support PMI, they do so in varying and inconsistent ways. And they each use proprietary data formats.
While it’s possible to get past many of the problems with the various implementations of PMI (often by the use of third party software), the experience of implementing MBD is going to be heavily flavored by the CAD vendor whose software you use. Probably.
When the first standards development was done for PMI, an important piece was missing. Though ASME and ISO standards defined PMI, there was no neutral CAD file format that was capable of representing that information. Rather soon, however, the STEP AP203 format was updated to the “E2” version, which included support for PMI data. But, the support didn’t quite cross the threshold of “good enough.”
Soon—likely in January—ISO will publish the STEP AP242 standard. AP242 is designed for long-term archiving of CAD data—which means that it will certainly get substantial support from CAD vendors. AP242 is also designed to provide close to full support for PMI. (There’s no such thing as absolutely complete support. There are always details that slop through the cracks.)
STEP AP242 may be good enough at representing both PMI data and 3D geometry that it becomes widely used as a primary file format in MBD initiatives on that basis alone. But what makes AP242 really interesting is a separate development, from several years ago: Direct editing CAD.
Historically, CAD systems have been great at editing their native files, and terrible at editing non-native files. Yet, to a direct editing CAD program, a STEP AP242 file is, for all practical purposes, a native file.
Over the last couple of years, with the support of the nonprofit 3DPDF Consortium, the 3DPDF file format has made great strides as a visual communication format for 3D CAD and MBD data.
Earlier this year, the US government updated MIL-STD-31000A, Technical Data Packages, to make it 3D MBD friendly. The new revision has a strong preference that physical product data provided to the government be defined and transmitted as 3D authority datasets. Interestingly, 3D PDF seems to hit a sweet spot, combining openness (PDF is an ISO standard), with the ability to accurately represent both 3D model data and PMI.
To date, 3DPDF has seen the most use in applications that involve human viewing. But the PRC 3D format (the preferred 3D representation for 3DPDF) includes an exact representation, which can be read and converted into NURBS curves by other applications (for example, SpaceClaim). Over time, it probably won’t be surprising if 3D PDF is used as an alternative to more traditional interoperability file formats, such as STEP.
Bryan Fisher, of MBD360, one of the best known consultants in the area of MBD, feels that the time is right for leveraging 3D models. “Standards supporting PMI have been, and continue to be, actively developed. But, even though more work needs to be done, we’re well past the threshold of ‘good enough.’” Fisher stated. “Organizations of all sizes, in a variety of industries, have demonstrated significant cost reductions by implementing 3D model-based business processes.”
With the payoff being proven, and all the bits and pieces coming together to make the process of implementing MBD both less difficult, and less risky, we may actually be getting closer to realizing the vision of fully leveraging 3D models.