when surfaces meet solids1 Jan, 2003 By: Don LaCourse Cadalyst
With little fanfare, EDS PLM Solutions and the folks down in Huntsville, Alabama, added basic surface modeling functionality to Solid Edge v12 earlier this year. Though meant only to provide users with basic tools needed to import and repair 3D models that are less than perfect solids, this addition is a significant development. When viewed from the broader MCAD industry perspective, it signals one more nail in the coffin of the solids-only midrange MCAD application. SolidWorks Corp. with its flagship product broke similar ground in recent releases.
These MCAD developers recognize the important role that surface modeling can play in building complex solid models. They are joining the ranks of veteran surface/solid applications such as VX CAD/CAM from VX Corp. and thinkdesign from think3. Surfaces and solids are finally converging in midrange MCAD applications, to the benefit of all users.
IN THE BEGINNING
When solid modeling made its debut in the mid-1980s, MCAD vendors employed two distinct database structures (CSG and B-rep). CSG (constructive solid geometry) used mathematical Boolean operations (union, subtraction, and intersection) to implicitly combine simple solid shapes to form more complex shapes. B-rep (boundary representation) explicitly positioned and related surface elements to form an airtight boundary that enclosed the volume defined by the part.
Surface and solid modelers today share the same roots within their database structures. CSG was fine for the early solid modelers, but soon went by the wayside as users demanded the ability to model more complex shapes that CSG alone could not handle. In stepped hybrid modelers that took advantage of the strengths of both CSG and B-rep.
As MCAD developers migrated toward their second generation of modelers, programmers discovered how to employ CSG-like operations using B-rep. You still find Boolean operations in today's solid modelers, even though virtually all of today's MCAD applications (surface and solids) use B-rep as their sole database structure.
WHAT IS A SOLID?
Surface modeling is not governed by these rules. You are free to create any number of surfaces, in any order. They may connect cleanly, or they may not. It's up to you to employ clean modeling techniques. This freedom attracts many users and is what makes surface modeling so powerful. Once a surface model is complete, you can use a simple "thickness" command to convert it to a solid.
TODAY'S HYBRID MODELERS
You can apply traditional solid feature operations, such as holes, pockets, protrusions, fillets, and chamfers, to open sets of surfaces. At the same time, you can explode a solid to gain access to its underlying set of surfaces, then apply surface modeling operations. You can add, remove, or replace surfaces, and perform local surface modifications such as control point editing (tweaking). When the required surface modeling is complete, sew the surfaces back into a solid (provided they form a closed volume).
A CONVERGING INDUSTRY
Developers of traditional solids-only applications add surface creation tools for a number of reasons, and some vendors are further along than others. Other than the obvious need for complex surface handling, the solids-only roots of many modelers cause problems for
To solve this problem, solids-only developers are starting to provide the surface creation and editing tools needed to repair those less-than-perfect models everyone seems to get once in a while. The hybrid solid/surface modelers never faced this problem, but did need to provide better-than-average import and export translators to handle the more complex surface models out there.
All users will benefit when the industry unites on support for both surfaces and solids. The expanded design envelope that surfaces bring to users of traditional solids-only applications is only one benefit. Import and export translators should improve because developers will have to export the new surface entities their programs can now create. With improved export, improved import must follow.
WHAT SHOULD YOU BUY?
Even with the convergence of surfaces and solids well underway, MCAD applications differ in their strengths and weaknesses in each area. You can count on all applications to offer the basics of feature-based solid modeling. They differ in their integration, approach, and support for surfacing functionality. If you are wondering, "What should I buy?," you should ask some additional questions as well.
Do I work with large assemblies? If so, an application with in-context assembly design (the ability to edit a part within the context of the assembly) and assembly management capabilities is a plus. Also look for an application that allows surface models to be interchangeable with solid models within the assembly. Find out what the restrictions are, if any.
Do I frequently import models from neutral formats (IGES and STEP)? If the answer is yes, then consider an application with strong support for surfaces. For example, tool and die designers need to repeatedly import multiple versions of customer parts that may come in many different formats. The CAD application should not only support surfaces but also provide extensive topology healing tools that help repair those occasional problematic parts.
As surface and solid modeling functions continue to converge, expect the line between traditional midrange solids-only applications and hybrid modelers to blur considerably. Though hybrid modelers have a leg up when it comes to developing the functionality and integration between surfaces and solids, this gap will continue to close as well.
From the user's point of view, midrange applications can only get better. Developers of hybrid modelers can't afford to sit around and simply tout that they offer the best of both breeds. This convergence will spur them to develop more advanced surface functionality and possibly new areas to set their products apart. Developers of solids-only applications will finally break free of those bonds and look to expand their user bases into more complex applications. The future of midrange MCAD application development looks bright indeed!