CAD/CAM wises up for the nineties.
`Expert' software that does more of the thinking for you, improved data management and communications, and even animation are among emerging CAD/CAM capabilities
Smarter and more "expert" CAD/CAM systems are on their way for the 1990s. Existing systems are being endowed with features like graphical interfaces, making them easier to use, and capabilities such as intelligent defaults, rule-based modeling, constraints, and associativity, providing designers with a more intuitive environment. CAD/CAM suppliers are striving to provide systems that are more akin to the way engineers actually perform their jobs, rather than forcing the user to adapt to the computer's way of doing things.
Other trends include better management of CAD/CAM data, improved capabilities for data communications, and tighter integration between CAD and CAM. Increasing use of solid modeling, sophisticated 3D graphic rendering, and even animation will enhance designers' ability to understand the form and function of the objects they are representing on-screen. And in the particular case of plastics mold analysis, more intelligent software and faster hardware are fast reducing the role of guesswork in tooling design.
CAD/CAM software in the 1990s will capture much of the knowledge held by experienced design engineers and embed it in the computer's software intelligence.
In such "expert systems," the computer will also store all of an organization's design handbook specifications. Every company has different rules that govern decisions made by engineers. CAD/CAM software will be capable of alerting engineers when they make decisions that violate those guidelines. For executive management, this presents the opportunity to create an accountability system for each engineer. If an engineer chooses to ignore guidelines, the system could capture and track his decision, as well as learn from the subsequent results.
In addition to prompting the engineer for possible design errors, the software could also present alternative decisions from which to choose. By leading the engineer through a series of prompts, menus, and intelligent defaults, the software will assume much of the burden for mundane decision-making, allowing the engineer to progress more quickly to a creative, optimum design.
"Built-in intelligence" is the key feature of some of the newest systems, which offer the full power of traditional CAD, but require many fewer complex commands than most CAD/CAM systems, so they don't require as extensive user training. An indication of where the industry is headed is provided by recent introductions from new startup suppliers. Often it's startups that push technology, because they have no user base to protect, so compatibility with existing systems is less an issue.
One startup, Ashlar, Inc. in Sunnyvale, Calif., if offering a new drafting package called Vellum that is described as "something so superior that you know you're looking at new-generation technology," by George Foundyller, president of the well known CAD/CAM consulting firm Daratech, Inc., Cambridge, Mass. The admittedly jaded Foundyller says this new software changes the way computer-aided drafting is done, by "thinking one step ahead of the user." Ashlar calls the feature Design Assistant, and what it does is use information from the way the user moves the cursor to infer what he wants to do. For example, placing the cursor near a circle might suggest that the user intends to draw a line tangential to the circle. The computer makes an assumption and indicates the construction to the user. If the inference is correct, the user can accept it. If it is incorrect, he simply moves the cursor again and the computer provides another option.
Dr. Martin Newell, Ashlar president and developer of Vellum, understands that users supply a lot of unstated information to a computer merely in the act of moving a cursor around the screen. Conventional CAD software "plays dumb" and ignores this information, forcing users to be more exactingly precise about what they want the computer to understand than comes naturally to most people. Packages such as Vellum and Claris CAD 2.0 for the Apple Macintosh, are beginning to eliminate onscreen guesswork and move toward making computerized drafting as effortless as freehand sketching.
Drafting Assistant's high-end CAD features such as integrated parametrics, associative dimensioning, detail views, and Nurb splines, help shorten the design cycle and increase productivity. Foundyller says, "Using Ashlar's Vellum is the easiest way around to produce technical drawings."
It's likely that such built-in intelligence will find its way into existing CAD/CAM systems through upgrades and enhancements. "One thing is certain," says Foundyller, "Drafting Assistant puts Vellum a generation ahead of other CAD software. Its features are destined to become the most copied item since windows. It's probably safe to predict that in 18 months, no serious CAD software will be competitive without them." Just as importantly, software developers will be stimulated to focus on writing interfaces that try to infer what users want and help them to do it, rather than try to modify user behavior.
GRAPHICAL USER INTERFACES
Ease of use has always been one of the most critical issues in CAD/CAM system development, and will continue to be a driving factor in the decade ahead. Interfaces for the 1990s already incorporate more convenient features such as pop-up menus and icons. Such features are characteristic of so-called graphical interfaces, whereby the user communicates via a "mouse" and on-screen menus of tiny images ("icons"), rather than having to type verbal commands on a keyboard.
For example, Intergraph Corp. in Huntsville, Ala., has adopted Visix Software, Inc.'s Looking Glass user interface as a Unix interface. According to Intergraph v.p. Robert Glasier, "The high-level graphical interface of Looking Glass allows users to concentrate on the engineering processes at hand and shields them from the complexities of the operating system." This icon-driven graphical interface provides not only a visual representation of Unix commands, but also a more intuitive method of managing files and directories, launching applications, and manipulating the Unix environment. It fully conforms with the OSF/Motif look-and-feel standard, based on Microsoft's Presentation Manager.
Other companies with new and improved interfaces include Structural Dynamics Research Corp. (SDRC) of Milford, Ohio, and Graftek, Inc. (formerly Unisys CAD/CAM, Inc.) of Boulder, Colo. While Graftek officials say the latest improvements to its interface go a long way in helping users, their work is far from done. "By the end of the decade, user interfaces will be very flexible," says Rosemary Brody, v.p. of R&D. "Sure, we'll have icons, but the key word will be flexibility. Users will be able to design their own interfaces," she predicts.
Other features that will make things easier in the 1990s are constraint-based modeling and associativity. These capabilities are finding their way into more and more systems. Constraint-based modeling lets engineers tinker with a design's parameters. It enables the building of designs based wholly on parameters such an object's length, width, and height. Change one, and the computer automatically adjusts other parameters to reflect the change. This makes it easier for engineers to try alternate ideas to see early on how well or poorly they work.
With associativity, all versions of a design are linked so that when the master design file is changed, associated versions are also automatically updated. Thus the newest version is always available by all departments, a big help in communicating engineering changes.
OPENING UP COMMUNICATIONS
Design engineers are clamoring for improved data communications, an area where lots of work is still needed. Full-blown CAD/CAM packages like I-DEAS from SDRC and PRO/Engineer from Parametric Technology Corp., Waltham, Mass., provide seamless data communication between different users of the same software. Of course, not everyone uses the same software--even within the same company. Communicating across disparate systems is inefficient at best. Experience shows that when computer files are translated from one data format to another, pieces of information are changed or lost. So for the 1990s, vendors are concentrating more on standardizing communications and on "open" architecture to facilitate communications.
Many users are looking to an emerging data-communication standard called the Product Data Exchange Specification (PDES) to make it easier to transfer engineering data between different applications and computers. PDES doesn't yet exist as an official standard, but a volunteer organization composed of vendors, users, consultants, and others are hard at work to make it a reality. The first version of PDES is expected to become an ISO standard in 1991. CAD/CAM vendors will soon be writing translator products, and shortly thereafter, users will be using PDES for what they hope will be truly intelligent data exchange.
Informally called STEP in international standards circles, PDES is an outgrowth of IGES, a 10-year-old industry standard for basic graphic and geometric data exchange. PDES will go beyond IGES, providing data-exchange capability not available now. With PDES, for example, users will be able to exchange "intelligent objects"--items such as mold components with all their associated attributes, (size, material, and properties)--not just their line-drawing or solid representations, as with IGES currently.
As developers see it, PDES will consist of a dynamic set of objects called the Generic Product Data Model. (GPDM). This will contain the objects designers and other users define as necessary for their applications. By putting objects into GPDM, entities can be shared. Integration--the recognition of similarities and the linking together of similar objects and concepts--is a key part of PDES. A major goal is to develop a minimal set of entities to satisfy all applications.
The goals of PDES are multiple. It's expected to provide a foundation for static and dynamic data exchange as well as for distributed databases and knowledge-based systems. It addresses exchange capabilities not just for basic graphics and geometry, but also for user-level application information such as tolerances, form-feature modeling, finite-element analysis, product structure, and NC machining.
PDES developers are taking a topdown approach in working on these areas, defining what information they require, not just what information currently exists in their CAD systems, as was too often the case with IGES. Therefore, unlike most standards, PDES is expected to drive technology, not just react to it. PDES is aimed at systems of the future, and is developing the appropriate technological tools. The PDES effort involves a significant amount of real research, which is taking place worldwide.
TOMORROW'S DATA MANAGEMENT
Management of engineering data is a big issue for the 1990s. It's becoming increasingly important as manufacturers realize the value of the information stored in their CAD systems' engineering database. Proper management of design data can benefit an entire company. For example, engineering costs can be lowered by reducing the number of redundant drawings. Existing drawings, modules, and analysis tiles can be quickly located and used as a base for new designs.
Engineering data management can also assist in efforts to integrate dissimilar software. Different programs can communicate through a database management system. Design data, for example, can be used in a manufacturing resources planning (MRP II) system without the need for custom software to link the two. This link can also help reduce costs and shorten product-development cycles by getting manufacturing involved in the early design stages.
Another important application for data-management systems is getting a better handle on engineering change orders (ECOs). The system can track drawing releases and changes to help manage ECOs. As a result, everyone who uses engineering data is ensured of having the latest revisions of the engineering design.
Engineering data management as a technology is still in its infancy. A few different approaches have emerged:
* CAD-specific database-management systems can be added to traditional CAD/CAM installations. Several computer manufacturers as well as CAD/CAM vendors offer database-management software (DBMSs) for engineering. Similar to those used for business applications, engineering DBMSs can handle the large amounts of data typical in CAD/CAM and can access multiple software packages for different applications.
* Some CAD packages have built-in data-management features. These programs store "objects" that consist of both graphic and textual data. The modeling module presents one "view" of this data; a spreadsheet or report presents another view. According to developers, this approach permits high levels of integration between dissimilar software programs.
* Scanning, conversion, and documentation control systems can store and track CAD and paper drawings. These systems have grown from CAD conversion peripherals to complete engineering drawing management systems.
It's unclear whether any of these approaches will emerge as an industry standard. But with the plethora of information available from computer systems and the need to keep vestiges of archived information for historical and maintenance purposes, the reality is that data management has become one of the most critical problems facing CAD/CAM users today.
Adding to the problem is the trend toward distributed computing, meaning networks of either personal computers or Unix-based engineering workstations, or multiple mainframes tied together. As the volume of information increases, so do the interrelationships between the various pieces of information. To deal with this problem, manufacturers must either begin building information systems that connect each existing software application to the other applications that must share data with it, or else replace their entire information systems with compatible systems--not a very practical solution.
While CAD/CAM vendors acknowledge paying increasing attention to the issue of data management, strategies are just beginning to emerge. One comes from Intergraph. There, Technical Information Management (TIM) is a new product center. It moves beyond the traditional CAD/CAM/CAE focus to promote information sharing across a distributed network--reaching users who need technical information but were formerly unable to access it effectively.
TIM comprises four major technology areas: Nucleus Data Management (NDM), Computerized Drawing Maintenance (CDM), Document Management and Distribution (DMD), and Technical Data Intergration (TDI). NDM locates, accesses, and controls data on a network. CDM enables users to incorporate paper and microfilm information in the automated CAD environment for management, maintenance, and editing. DMD addresses the management, review, markup, and distribution of documents. And TDI represents the final step, integrating data into multiple document formats, databases, and computer platforms. Intergraph sources say TIM provides the framework for managing the wealth of information resources used by its customers.
Other vendors have other approaches. Sources at Manufacturing and Consulting Services (MCS) in Irvine, Calif., say it is "wrapping around" third-party database software products, rather than emphasizing its own solution. "We're providing connections," says MCS Anvil product manager Scott Owens. "We've chosen not to bundle in any specific database because the people using Anvil have so many different needs," he says. MCS is currently providing interfaces to Ingres, Sherpa, and Oracle.
A SOLID DATABASE
Some industry experts foresee that solid modeling will play a key role in engineering data management. Sources at SDRC, for example, expect that by the end of the 1990s, most CAD/CAM systems will be solids-based. Solid modeling has gotten attention because it offers the most realistic representation of an object. And for any CAD application where analysis is necessary, solids offer the best solution because they're geometrically and mathematically complete.
Traditionally, wireframe and surface-modeling software packages were used to generate geometric models by computer. But wireframe and surface models are ambiguous in representation and interpretation. Also, they are inadequate for such applications as mass property calculations, hidden line removal, generation of shaded images, interference and clearance analysis, and the generation of multi-axis NC machining instructions. Solids on the other hand, allow the creation, storage, and manipulation of complete models. It's like building a physical prototype, but the designer has the advantage of easily changing any aspect of his model. With solid models, a designer can assign a material type and then use analysis software to obtain information about the parts' behavior under various conditions.
And while this is currently the primary application of solids, they can do a lot more. Solids can play a role in both database management and in the integration of design and manufacturing. Solids have the ability to unambiguously define part geometry by mathematically describing the interior and exterior of the part. And that information can be the basis for downstream engineering activities including analysis, testing, NC programming, and drafting.
As solids make their way increasingly into design, the solid-model database is being more often accessed downstream by manufacturing functions, providing a bridge between design, engineering, and manufacturing. In other words, it's a tool for so-called "concurrent engineering" of a product and the manufacturing system to make it. The use of the same solid-model database by design and manufacturing allows important decisions affecting the manufacturing of a product to be made earlier in its development cycle. The solid-model database can serve as a true product structure and contain information about individual part geometries as well as total product configuration data. It can store information ranging from descriptions of component attributes to final assemblies.
Such a solid-model database can also provide an integrated approach to the mechanical design process. Often each software application in mechanical design requires its own unique file format. Though some information about a part is unique and needs to be described at execution, it is just as certain that there will be redundant data and redefinition of data from one application to the next. But a solid model, as the only complete representation of a part, can provide a common geometric database that would enable all applications to share information.
An integrated mechanical design process has further value. With enough information to drive the other applications of mechanical design, solid models could become the basis of a uniform method of mechanical design.
Unlike electronic design, mechanical design lacks a coherent methodology. Currently, mechanical design processes differ from company to company, and even from department to department within the same company. This is inefficient and leads to other problems, especially in light of today's shortage of mechanical engineers.
The development of a standard design methodology would alleviate the manpower crunch somewhat by providing increased transportability. Uniform design methods would enable a company to obtain greater productivity from the engineers it has on hand and reduce the learning curve required for new engineers to learn a company's particular design methodology.
Solid modeling has been available since the 1970s, but only recently has it become practical. Early systems were too expensive; computer power was inadequate; and display technology needed improvement. But as hardware price and performance obstacles have been surmounted, use of solids has begun to grow. Over 30 suppliers now offer solids software, and many of the packages run on low-cost workstations and even PCs.
RENDERING AND ANIMATION
Designers are beginning to see the benefits of photograph-like images and animations created with computer graphics. For one thing, realistic images and animation can assist in verifying that a design is correct. A more obvious use of such renditions is in presentations. Unlike commercial rendering and animation, engineering versions must simulate real-world conditions. And since the rendering and animation capability might be used infrequently, engineers require low-cost, easy-to-use systems.
Rendering is the process of adding visual clues to a 3-D model to make it appear more realistic. These clues include smooth shading, shadowing, and reflections, as well as textures and surface finishes. Many 3-D CAD modelers provide some of these capabilities, but the more advanced features that lead to photorealism are found in specialized systems.
Animation provides moving pictures, typically with images that have been realistically rendered as well. In a few minutes, these animations can lead to in-depth understanding of a problem that might take months with reams of numerical results. Animations can be driven by analysis results. Some dynamic analysis packages, for example, include animation capabilities. Or, animations can be created strictly for viewing. The viewer can "fly through" a proposed factory, for example.
Rendering and animation are coming from a few different directions. CAD/CAM vendors are enhancing modelers to include rendering capabilities. And recently, developers of commercial rendering and animation systems have begun to target engineering users. In any case, the key to integrating these capabilities into engineering design is the ability to use existing CAD models as a starting point.
MOLD ANALYSIS' FUTURE
In the 1990s, mold analysis will be endowed with many of the same capabilities as CAD/CAM systems in general. According to Russell Stay, plastics products manager at SDRC, mold-analysis systems will be provided with intelligent default values to facilitate use. He says the software will have more intelligence, including the ability to learn and remember sequences and commands, so that plastics engineers will be able to move more quickly through the rudimentary steps of analysis and simulation. "In the 1990s," says Stay, "systems for mold analysis will help with quick-and-dirty analysis at the start of the design process to enable engineers to move quickly into the areas that require more time and detail. The software will allow engineers to confirm assumptions or disprove them immediately, thus shortening the time needed to obtain final results. Once an initial design has been created, the engineer of the future will ask the software to recommend changes to correct problems, and the software will suggest alternative geometry."
As we move toward the year 2000, the need for speed will continue to be important in plastics mold analysis. Today, overnight analysis is considered too long. "But by 1995, analysis that takes more than 15 minutes will be considered too long," says Stay.
Recently introduced systems tightly integrate the solid model or 3-D design and the analysis model. In the future, analysis will be performed directly on the solid model, without ever going through the step of creating a finite-element model.
An indication of what's coming can be seen in the latest upgrade provided by Advanced CAE Technology, Inc., Ithaca, N.Y. Its C-Flow software now employs artificial intelligence to explicitly interpret weld lines and air traps. The system will now actually show air entrapment and draw weld lines, rather than leaving them to be deduced. This feature is the result of a systematic algorithm developed by the Cornell Injection Molding Program. Sources at AC Tech say this capability is extremely important to fully realizing the power of CAE technology as applied to injection molding.
Also new from AC Tech is an analysis program to predict the outcome of co-injection and gas-assist injection molding. The simulation traces individual particles of plastic during the entire molding process, allowing the spatial distribution of skin and core materials to be identified at any instant. Differences in material properties, or gas dynamics are accounted for; information is provided on spatial distribution of pressure, temperature, shear-rate and shear-stress profiles, as well as distribution of skin and core materials and skin thickness.
Another instance of new capabilities in mold analysis comes from a new entrant in the field, Techanalysis Inc., Indianapolis (see PT, June `90, p. 14). Its Plastec software provides explicit representation of fiber orientation as a consequence of flow vectors throughout injection or compression molded parts.
PHOTO : Intergraph has enhanced the operation of its CAD/CAM system with Visix's Looking Glass UNIX interface. Other suppliers are also adopting icon-driven graphical user interfaces to provide a more intuitive method of managing files and directories.
PHOTO : Expert' software provides the C-Flow cavity filling simulation software with new capability to predict locations of weld lines and gas traps automatically. In photo 1 (left, top), weld lines are identified in an automobile grill. Color of the lines indicates the order of formation--blue forms first and red last.
PHOTO : Photo 2 (left, bottom) depicts melt front distribution at a filling time of 6 sec. Note correspondence of melt front locations with the red weld line through the grill region.
PHOTO : Photo 3 (above) identifies gas-trap locations to aid in venting.
PHOTO : Photo 4 (right) depicts melt fronts at a filling time of 7.5 sec. Note correspondence of the melt front locations with both the weld lines in photo 1 and gas traps in photo 3.
PHOTO : In an integrated design and analysis system, the solid model can be used as the basis for all analysis.
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||computer-aided design/computer-aided manufacturing software offers improved capability|
|Article Type:||Cover Story|
|Date:||Nov 1, 1990|
|Previous Article:||Where to now for dual-ovenable packaging?|
|Next Article:||CIM begins at home for Texas Instruments.|