Printer Friendly

TI develops expert systems.

TI develops expert systems The Process Automation Center of Texas Instruments, Dallas, TX, has developed two new application programs incorporating expert-system technology, according to Marvin Aardal, the Center's manager of business development.

One program is a rule-based system that automatically analyzes signals captured during a plasma-etch process. Purpose of the application is to determine whether or not the process is in control.

Before this system was installed, an experienced engineer had to manually examine traces on a strip chart. Since this was time-consuming, only spot checks were affordable. Using a rule-based system has enabled embedded human judgment to be applied continuously.

TI also has implemented an expert system to improve control of the epitaxial process in semiconductor manufacturing. This process requires continuous adjustment of control parameters to adhere to specification limits.

Before the system was installed, operators with over 10 years of experience often were needed to maintain consistent quality. The expert system replaced the judgment of these operators to keep the process within control limits.

Key to success in this project, says Aardal, is continual maintenance of the knowledge base. By updating the system with new techniques learned by the process engineer and operators, the process progressively gets better.

As a result of applying this second system, the company is achieving tighter conformance to thickness and resistivity targets. Quality of the end-product is thus improved and maintained at high levels.

"These applications are exciting because they emulate very complex functions previously performed only by humans," says Aardal. "They are, however, the hardest types of applications to develop. Each system took over two years to develop and implement.

"Quicker payback has been achieved by using expert systems for equipment diagnostics and repair. Maintaining complex equipment used in the semiconductor industry often requires highly trained technicians with years of experience. In some cases, the manufacturer's rep must be called in to get equipment back on line after a failure.

"Expert systems can be built rather quickly to advise less knowledgeable technicians and even operators in repair of equipment," he adds. "This reduces the mean time to repair, and also results in a higher mean time between failures."

One key to success in implementing expert systems for equipment diagnostics is integrating the expert-system portion with other computing systems, he says. "The most useful systems have automated access to PLCs, databases, and graphic libraries. These multiple forms of computer systems work together to make the over-all system very user-friendly.

"Even though today's expert systems cannot be made to learn, the effective integration of techniques gives the appearance to the user that the system is gathering new knowledge. By retaining a record of maintenance performed on an individual machine, the system determines the probable cause of an equipment failure. This is unique to each machine.

"Then, by way of interfaces to PLCs," Aardal continues, "an expert system can be programmed to cycle the equipment through a diagnostic journey. The system then monitors results without human intervention. In some cases, it can invoke changes in equipment setup to adjust an anomaly--again without assistance from an operator or technician."

Other types of applications for expert systems in manufacturing are emerging, he says. Applications being pursued include product failure analysis, intelligent data analysis, factory simulation, and scheduling. "All future applications will require integration with exisiting factory information systems," he concludes.
COPYRIGHT 1989 Nelson Publishing
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 1989 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Publication:Tooling & Production
Date:Sep 1, 1989
Previous Article:Monitoring quality welds at GM.
Next Article:Shop cuts time for prototyping.

Related Articles
CIM begins at home for Texas Instruments.
TI, Symbian & Nokia Team Up For 3G WID Dominance.
TI to Buy Libit for $365m to Boost VOIP for Cable Modems.
Texas Instruments teams with Silicon Logic Engineering to strengthen ASIC offering.
TI handhelds and peripherals: for more information, visit (Special Advertising Section).
Pixology launches in-camera red-eye solution, developed in collaboration with TI.

Terms of use | Copyright © 2016 Farlex, Inc. | Feedback | For webmasters