June 2007 - Expanding CAD to CAx Interoperability with Geometric Reasoning

June 25, 2007

As CAD applications and practitioners have steadily matured to the extent that 3D modeling is now considered the norm for designers, it is easy to forget that CAE, particularly Finite Element Analysis (FEA), was around before CAD software even existed in any meaningful way. The underlying principles of FEA and indeed the term itself were first established in the 1940s, and by the 1970s commercial code was in use at major organizations involved in the automotive, aerospace, defense and energy industries. In those days, without a 3D CAD system in place to define the geometry under analysis, engineers had to spend significant up front time and energy using whatever rudimentary geometry creation tools were available in the FE pre-processor to build an analysis model. Indeed geometry creation tools were often non-existent and analysis models had to be defined by manually creating nodes and elements

Effectively, real-world geometry, or at least an approximation of it, was being defined digitally specifically for the purposes of analysis, and a fledgling interoperability industry based on creating geometry and pre-processing it for FE applications was born. This was the environment that spawned Cambridge-based FEGS Ltd, the UK forebear of TranscenData Europe Ltd, and its analysis pre- and post-processing and data interoperability software.

Before long, of course, CAD tools began to gain traction, and it made sense to look for ways of re-using the geometry created by designers in the applications being used by analysis engineers. The theme of re-using CAD data in analysis software has persisted to this day, and indeed has been one of the fundamental drivers of recent engineering data interoperability software development. But the challenges interoperability specialists have had to meet over that time have changed steadily and inextricably over the past three decades.

Translation

The initial interoperability hurdle, one that will never really go away completely, is the classic problem of translation from one particular geometry based system to another. This seemingly straightforward process has historically thrown up all manner of problems, typically framed by how different systems go about defining and storing their 3D models, and factors such as the topology of how entities are connected, the mathematics and relative complexity of geometry definitions, and the fundamental modeler tolerances that tie models together. Misalignment in any of these areas and the receiving system, whether it's CAM, CAE or a different flavor of CAD, will make an interpretation based on its own standards, This interpretation is unlikely to preserve design intent and will often cause a conflict that will bring the interoperability process crashing to a halt.

All too often, CAE engineers were forced to spend a disproportionate amount of time recreating the geometry they are supposed to be analyzing, hardly a giant leap forward from the days before CAD had found its feet. Any thoughts of associatively linking CAD and CAE so engineers could proactively make changes to the master model based on their findings would have been met with snorts of derision.

Basic interoperability or what we might loosely call CAD-to-CAD exchange has improved immensely over the past decade or so thanks to a number of factors both technical and practical. Neutral file formats like STEP and IGES have matured and become more widely accepted; the market has consolidated around a limited number of underlying modeling kernels, with systems such as Parasolid being widely adopted. Translation software such as CAD system-embedded tools for automatic native translation and dedicated standalone translation and repair tools have matured to the extent that many technical issues can now be easily solved by non-experts. There has been some success, particularly in larger companies, in sidestepping interoperability issues altogether by implementing single-vendor corporate PLM strategies, in some cases even extending this policy into the engineering supply chain.

Defeatures and benefits

So does the steady reduction in engineering data interoperability issues signal the end of the line for specialists like ITI TranscenData? Far from it. What it has meant instead is a renewed focus not just on the fact that data has to be translated and re-used but also on how it will be used when it gets to the target system. Engineering analysis applications, in particular, have very specific requirements when it comes to the kind of geometric detail that should be included (or excluded) at the meshing and computation stage

Gradually, the process of preparing geometry for analysis has fallen into the realm of interoperability, certainly at ITI TranscenData. CADfix has well established tools for removing both valid CAD features that may be important for manufacturing or cosmetic purposes but play no part in structural performance, and unwanted design artifacts, such as short edges and sliver faces, that will actively hamper CAE. In CADIQ we offer designers the ability to quality check their geometry with CAE in mind, enabling the identification and elimination of problems upstream in the design process long before they come to light in downstream analysis applications and are time consuming and costly to fix.

Here, too, in recent years, the maturing market has to some extent improved CAE interoperability. Computing capacity means meshing small faces and edges is less of an issue, so fewer simplifications need to be made in the interests of analysis run time, and the CAE system developers have developed their own tools to suppress such features and meshing constraints anyway. And again, the emergence of dedicated data exchange tools like our own and to an extent consolidated PLM strategies and corporate acquisitions and product integrations have worked to ease the interoperability burden

Advanced knowledge

So while there is still work to be done on defeaturing, particularly in teaching software what geometry can and cannot legitimately be suppressed prior to meshing, the real focus now is on more advanced CAD to CAE integration considerations and the geometric requirements of high end specialist analysis applications. We are now looking at the automatic generation of complex flow volumes and "shrink-wraps" of complex assemblies of components. These are all new interpretations of geometry that focus on or suppress internal details (key considerations for CFD, electromagnetics, acoustics and other CAE disciplines) and in some cases dimensional simplifications, perhaps automatically reducing certain regions to thin plates prior to meshing for FEA. Such transformations are a step beyond standard modeling tools and access to the native CAD system or the history tree governing the model's construction may be of little or no use. For more advanced and radical defeaturing and adaptation of the CAD model for analysis, a new level of understanding, or Geometry Reasoning (GR), of the structure and make-up of the CAD model is needed. GR will underpin future developments in interoperability and CAD to CAE integration

GR recognizes that the traditional B-REP nature of 3D modeling, whilst serving its purpose for general CAD, has its limitations when the aim is to extract a deeper understanding about the nature of complex geometry. Beyond the simplest shapes it is ill equipped, for instance, when it comes to delivering properties such as proximity, aspect ratio and thinnest and thickest points, factors fundamental to making the kind of transformations outlined above. The GR of a CAD model extracts exactly these properties by using a technique that "walks around" inside an object, capturing its underlying characteristics at a new level of abstraction. GR is a key area of functionality that is being actively researched and developed today at ITI TranscenData, with a view to bringing new GR based product functionality to market in the coming months.

Advances in interoperability and CAE technology have always gone hand-in-hand. We have seen the interoperability focus shift from translation and repair of geometry to proactive support of the downstream application. With new forms of CAE emerging and demanding more and more geometric insight, it seems certain the interoperability journey will continue for many years to come, and ITI TranscenData aims to be at the forefront with developments in its product suite to meet these new and exciting requirements.

About ITI TranscenData

ITI TranscenData is the product data interoperability business within International TechneGroup Incorporated (ITI). ITI TranscenData provides interoperability solutions that enable companies to effectively exchange, reuse, integrate and share engineering product data among diverse software applications used throughout a manufacturing enterprise and its supply chain. ITI TranscenData solutions are aimed directly at reducing and eliminating non-value-added time and costs associated with reusing product data. ITI TranscenData solutions are distributed worldwide through a combination of direct sales, resellers, and strategic vendor/OEM partners. Included among ITI TranscenData’s customers are ABB, Airbus, BAE Systems, Boeing, EADS, Ford, Honeywell, Lego, Lockheed Martin, MAHLE Powertrain, Pratt & Whitney, Raytheon, Renault, Rockwell Collins, Rolls Royce, Samsung Electronics, Siemens Automotive, Snecma, Spirit AeroSystems, Thales, ThyssenKrupp, US Air Force, US Army, Xerox and others.

www.transcendata.com

For further information, please use the ITI TranscenData Contact Page or the below information:

ITI TranscenData Americas and Corporate

Tel: +1-888-783-9199

ITI TranscenData Europe, Middle East and Africa

Tel: +44 (0) 1954-234300

ITI TranscenData Asia-Pacific

Tel: +1-888-783-9199

Privacy Statement