Virtual Manufacturing and Knowledge-Based Design: Society of Automotive Engineers Topical Technical Conference

Ellen Domb, Editor
The TRIZ Journal

Introduction

The SAE Topical Technical Conferences focus on specific issues, and bring together diverse experts in the field with people active in implementation of the methods being discussed. TRIZ was included in this meeting on virtual manufacturing and knowledge-based design because TRIZ’s history of development from the knowledge base of the world-wide patent collection, and because TRIZ is unique in its capability to accelerate the conceptual design and problem resolution process. The other tools presented either

• Automate existing processes (the expert system process presented generates 1.36 million solutions, which then have to be sorted to find the 4 that are practical!) or
• Enhance existing proceses through early human interaction. Many of the virtual reality tools make it easier for people to see environments in 3 dimensions (such as a workcell, or a car interior) for problem identification before the commitment to the design is made.

This review will cover only those papers that had direct or possible relationships to the use of TRIZ, with the editor’s observations on those relationships.

Technical Papers Related to TRIZ

George Hrydziuszko is Director of the EDS Detroit Virtual Reality Center. He presented an overview, "The Virtual Revolution in Design and Manufacturing" that predicted that extensive, integrated tools of virtual reality (VR) in the next decade will change design, manufacturing, and assembly radically. VR is now being used for visualization of design—in the future it will be used for production as techniques now used for rapid prototyping become the manufacturing tools for "mass customization."

He noted 3 barriers to the implementation of the family of virtual reality tools::

1. Cultural: The comfort of the use of paper design, physical prototypes. Traditions of specific decision making authority based on models. The cultural impact of making mega-dollar decisions through the use of "toys."
2. Organizational structure and antiquated methodologies. Tools are enablers. Organizations can be enablers or roadblocks. Improvements tend to start in remote corners of the organization, protected from the roadblocks of the existing organizational infrastructure. Until the improvements are used organization-wide, only a small fraction of the benefits will be realized.
3. Perception of cost: There is always a cost of change, and benefits may not be immediate. Long-term rewards of change are priceless. Cost analysis frequently does not includes such issues as

• What is now being done that could stop being done with VR technology? (Physical prototypes, rework, redesign, etc.)
• What are the losses due to product introduction delay?
• Competitive advantage to early adopters.

Accounting for these factors will overcome any objections to the cost of VR systems.

Bernie Nadel of IntelliGineering Corp., presented an excellent tutorial on knowledge-based engineering. He started with the history of expert systems. Since expert systems are computer programs that reason like the experts the benefit is the merged value of consulting with many experts. For creative situations, there is no deterministic algorithm, so the methods used combine knowledge with inference. This results in gigantic searches, and amplifies the trial-and-error opportunities many-fold. His example of a constraint satisfaction model for the design of a transmission using an expert system was an excellent teaching tool for demonstrating the method. Possible combined use of artificial intelligence techniques with virtual reality were discussed, and possible combined use with TRIZ became obvious—TRIZ analysis and concept development could limit the number of trial-and-error pathways that are explored by the artificial intelligence system, and make the results much more focused.

Sylvia Karmanoff of General Motors Truck Group presented "Workcell Visualization." She described a system that uses 3-D modeling of people, machine, and parts to simulate manufacturing assembly. The technique lets engineers detect problems before designs are finalized, and is an important step in simultaneous engineering. Involvement of suppliers in simulation of the final process has started.

William Regli of Drexel University presented "Internet Computing for Agent-Based Engineering." Networked suites of tools for design, acquisition, test, production and maintenance will be available to engineers, buyers, testers, manufacturers and marketing professionals. "Agents" are appearing for many applications. Three examples were presented

1. Partner and subcontractor selection
2. Design-for-manufacture
3. Design knowledge-bases and repositories

TRIZ-related tools, such as searches for scientific effects that enable breakthrough solutions are in the third category of the tools that Dr. Regli described. ( Invention Machine’s Phenomenon™ is an example of this kind of enhancement of a basic TRIZ tool by adding Internet search capabilities with intelligent guidance of the search to find means of accomplishing desired functions.) The research now underway with academic-industry-government collaboration is to develop very fast means of identifying appropriate search objects by many characteristics so that the searches return a limited number of very useful results.

Shuh-Yuan Liou of the Ford Motor Company presented the Ford Design Advisor system, that is a principal tool for the Ford 2000 integrated system. The Voice of the Customer and the Voice of the Manufacturing Process are integrated with the detailed CAD-CAM systems and specialty engineering databases (injection molding, stamping, forging, etc.) Tools identified as "Design Evaluation Reports" and "Tooling Feasibility Analysis" are probable points of entry for TRIZ, since these Design Advisors alert the designer to conflict situations. He presented case studies for a variety of parts and subsystems that demonstrated the potential for 10-15% tooling cost reduction and 15-25% redesign time reduction and improvement of first pass success.

Likewise, Doug Fish of Tecnotatix Technologies showed a proprietary software implementation of a virtual manufacturing tool set. The feasibility analysis step is the injection point for TRIZ application to parts design. This system deals only with parts and subsystems, so there does not appear to be an opportunity for system-level breakthrough design.

Fascinating papers were presented by many other speakers, showing the integration of robotic software with VR data to simulate the behavior of robots in the the assembly process, the integration of human performance with VR for ergonomic design, the use of VR for training and development, and the use of VR in conjunction with the World Wide Web, and the technical issues of managing proprietary data on detailed designs when using VR for esthetic feedback. The repeated theme, used by all speakers was that VR "catches problems sooner" thereby saving money.

Ellen Domb tied all these sophisticated tools back to the product and process development method with her tutorial on TRIZ. The conclusion was that VR may help you find problems, but TRIZ provides the means to develop innovative concepts that remove the problems, and that prevent future problems. The audience provided TRIZ demographic information—in a sophisticated engineering audience that uses many of the most powerful CAE tools and that also develops new tools, only 12% had heard of TRIZ (or TIPS, or SIT) and only 3% had any experience with TRIZ.