Thursday, July 1, 2010

What next in 3D CAD? - Ideas for Next Generation Technology

It has been over two decades since a new technology emerged in 3D CAD Design. Parametric representation and B-Rep modelling created a tectonic shift in the CAD industry spawning several codes with the latest generation exploiting the Windows architecture for ease-of-use and reach.

While history based representation has provided innumerable benefits, history free editing, a mere corollary, has been confined to limited areas of geometric manipulations in comparison to creat-modify-update-release nature of drawings for manufacture, that the industry requires.

Common challenges still remain in 3D CAD technology that needs to be addressed. Some of them include:
  • Handling of large data sets
  • Faster regeneration time during edit and update processes
  • Quicker and easier ways of generating/ modifying 2D Drawings
  • Intuitive approach to design using constraint-free environment
  • Operating System Independent environment
  • Freedom to collaborate without data translation
While the debate on choice of parametric vis-a-vis history free approach would continue, a wishlist of user-specific requirements could include the following:
  1. 2D framework that enables 3D data manipulation without losing sight of the 2D drawing - be it part or assembly (because at the end of the day the user requires a 2D Drawing anyway - am I asking for a roll back ? No !)
  2. Hybrid approach with both History-free and History-based environment wherein the user has the choice to decide on the approach in the middle of a design process - going back and forth
  3. Transparent inter-operability - absoluted no barriers !
  4. Ability to handle Large data sets in fractions of time taken currently
  5. Surface or Solid - User does not care or should not be troubled to fix 'closure volume' errors
  6. If one can view HTML pages on any web browser, why not 3D CAD and associated 2D CAD data? e-Drawings has helped - but still more needs to be done here.
  7. Web standards have evolved - why not leverage on this for 3D CAD environment and framework? (a.k.a Google Sketchup with an advanced framework)
  8. OS Free, platform independent approach with distributed cloud computing (this is emerging as a serious alternative with much spoken and written about it)
  9. If Torrents can used a distributed collaborative framework, why not leverage on this for 3D CAD data sets?

Time to re-look at Gregory Patch?

Gregory patch (remember Designbase from Ricoh?) has numerous intrinsic advantages over NURBS in terms of handling large data sets, concise representation and computationally less intensive architecture. Combined with cloud computing, history/history-free hybrid approach, simplified language (CTML - CAD Text Mark Up language?), ability to manipulate 3D data in 2D (yes - I mean this seriously) and 'Torrent'ial collaborative engineering - we will have a deadly cocktail of capabilities that I am sure the users would enjoy and celebrate !

Let us free our minds of constraints and break conventions, now that technology empowers such initiatives. Are we ready for this?


Anonymous said...

Some of the issues stated in this blog has already been addressed by some of the CAD tools.

High end CAD systems like ProEngineer, UG NX and CATIA are supposed to handle large data sets.

Regeneration time is tied to the depth of the feature tree and the complexity of the features in the feature tree.

High end CAD systems have tighter integration with their 3D to 2D models.

Hybrid approach is a dream, which may or may not materialize.History free modeling does not play well in scenarios where user have to go back and modify a parameter of a feature. This requires total re-architecture of the existing CAD system's feature tree regeneration algorithm.

Transparent inter-operability - do you mean between the CAD systems?

In the higher end CAD systems you can mix solid and surfaces together. In ProEngineer and CATIA this is very common.

In near future viewing 3D models in the web will become very predominant. Some of the CAD systems allow exporting their data as VRML or in their proprietary format. With proper plug-in it is possible to view the r CAD models in the web browser. Modifying the 3D models in the web is totally different story.

As CAD is widely used in Auto/Aero/Military/Health industries. The CAD models contain lot of proprietary information which is usually locked into the PLM vaults. Distribution of that kind of data is not at all possible.


Nat said...

1. Concept of Large data sets is expanding, with challenges faced even by stated CAD systems, including, CATIA and UG. Enabling technologies are challenging the management of data sets since users keep breaking barriers.
2. Regeneration time grows exponentially with inter-linked features (equations, excel-driven, top-down relationships or a combination) that could be addressed with hybrid approach using more intelligence that current technological imperatives offer.

3. Though tighter integration of 3D with 2D drawings is available, ease of use and liberty enjoyed by current 2D users is still far from realization - in spite of 3D being there for over 2 decades, 2D standlone designing is still in practice with a considerably larger share of CAD users vis-a-vis 3D.

4. Hybrid approach a dream - hmmm. Are we not familiar with dreams becoming reality many times in our lives? Nothing is constant except change, right?

5. Transparent inter-operability - Why confine only to CAD, though CAD would be a good place to start with?

6. While CAD data contain IPR, we are referring to sharing among eligible groups with collaborative approach and not to all and sundry.

Thanks for the comments... More discussions such as this would lead to interesting developments and ideas?

Anonymous said...

In regard to viewing and modifying 3D CAD models in a web browser, please check out:

Steve said...

Concerning the "hybrid approach with both History-free and History-based environment", what are your thoughts on Inventor Fusion?

Nat said...

Hi Steve:

Inventor Fusion is looking at the Hybrid model albeit in a different way. We have two different interfaces to go back and forth, why? Why cannot this be evolved onto a common interface? Why is the user forced to move between two different applications to get his job done. At the end of the day, whether the history tree gets updated or not, the CAD application should 'facilitate' direct editing irrespective of what happens on the history tree, while protecting parametric representation. I am not confident that Inventor would be able to achieve by decoupling the approach. It looks more of a way to keep the programming complexity away. Also, if the direct editing does not gain ground, they would have a choice to do away with it, right? It is interesting how developers from different companies try to bring out unique approaches to solve a riddle of having history free and history-based CAD together. It is important, however, to keep in mind why direct editing is being discussed. If these benefits can be leveraged in a common environment without user-grief, I am sure the tool-set would be a winner. Until then, any approach giving options to users is always welcome for trial and keep, right?

Steve said...


Yes. From my understanding, the intent is to make Fusion a part of Inventor.

BTW, do you know of any other applications that reverse engineer direct modeling into history?