Basic Tasks for Knowledge Based Supervision in Process Control

ACOSTA, GERARDO G.; ALONSO GONZÁLEZ, C.; PULIDO, B.

ENGINEERING APPLICATIONS OF ARTIFICIAL INTELLIGENCE

Elsevier

Lugar: Amsterdam, Holanda; Año: 2001 vol. 14 p. 441 - 455

0952-1976

A new tasks taxonomy for knowledge-based global supervision (GS) of continuous industrial processes is introduced in this work. Possible required tasks are specified together with the analysis of their dimensions, which should be useful in the selection of the final capabilities of supervision. Moreover, these dimensions would help end-users and designers when comparing different systems. Several methodologies based on concepts such as generic task, generic operation or heuristic classification have been proposed to transform knowledge-based system (KBS) development in a systematic knowledge engineering activity. These approaches have been quite successful in domains such as medicine or mineral prospecting, identifying a large number of tasks that experts in the domain articulate to solve the problem. However, this was not the case in the process control area. The selection of tasks and their capabilities is the first step to be taken, even before choosing a KBS analysis and design methodology. Authors found a lack of facilities to do this selection in the aforementioned approaches when they tried to develop a global supervision tool in a beet sugar factory in Spain. Hence, this article describes an attempt to fill this gap. Moreover, it shows how this taxonomy supported the analysis and design stages of a supervision tool in the mentioned industrial application.