Knowledge discovery from process operational data for assessment and monitoring of operator's performance

This contribution describes a knowledge discovery system that can be integrated with a modern computer control environment to continuously and automatically capture, characterise and assess the skills and behaviour of operational personnel. The system is developed and tested based on a joint simulation framework of human-process interactions. The operator's performance is modelled using a knowledge-based system, which is a collection of rules representing operator's perception and interpretation of on-line signals as well as subsequent planning and sequence of actions. The process behaviour is represented by dynamic simulators. An important component of the knowledge discovery system is a clustered fuzzy digraph that can be used to qualitatively/quantitatively simulate the temporal behaviour of joint human-process interactions. A case study is also given which demonstrates the feasibility of assessing operator's performance through analysis of operational data. (C) 2000 Elsevier Science Ltd. All rights reserved.