Development, Validation, and Limiations of an Expert System for Safety and Health Concerns in Technology Development

• Published in: Proceedings of the Human Factors and Ergonomics Society Annual Meeting Vol. 44; no. 27; pp. 306 - 309
• Main Authors: McMullin, Dianne L., Akladios, Magdy, Gopalakrishnan, B., Myers, Warren R., Becker, Paul E.
• Format: Journal Article
• Language: English
• Published: Los Angeles, CA SAGE Publications 01.07.2000
• ISSN: 1541-9312; 2169-5067; 2169-5067; 1071-1813
• DOI: 10.1177/154193120004402706

Hazardous waste remediation projects, like many other technologies, are experiencing a rapid change in technology triggered by the continuous demand for better, faster, and cheaper cleanup technologies. TEXPERT, a web-based expert system for safety and health, is beginning its third year of development. The goal of the project was to provide appropriate worker safety and health information to technology designers during the design process. A group of ergonomists, safety engineers, and industrial hygienists and a group of computer systems specialists first produced a prototype of an evaluation system for a single, simple technology. Armed with this success and lessons learned, the developers expanded the system to allow evaluation of an almost endless number of technologies. The developers determined that the majority of technologies contain a finite number of components (e.g. compressors, storage tanks, etc.). Rules and knowledge bases were then developed for each component. The components act as building blocks that can be assembled by the designer into a representation of his/her technology. Each component has design questions that result in a list of probable safety/health events, recommendations and standard operating procedures to lessen the risk of the events, and probabilities and severities associated with the events. Considerations undertaken by the TEXPERT development team were the human-computer interface, the safety/health knowledge level of designers, the interaction between components, the interface between databases and the expert system shell, the time required vs. detail level of the expert system, and the vast array of technologies that could be evaluated.