A Knowledge-Based Framework for Automating HAZOP Analysis.
Laboratory for Intelligent Process Systems, School of Chemical Engineering, Purdue University, West Lafayette, Indiana 1997 Aug:11 pages
A framework was developed for use in automating HAZOP analysis which can be used for a variety of chemical facilities. The study indicated that the knowledge required for HAZOP analysis can be separated into process specific and process general knowledge. The process specific knowledge included the process piping and instrumentation diagram and the process material properties. The mental models used by experts performing HAZOP analysis could be represented using directed graph based qualitative models of process units. Investigation of large scale flowsheets indicated that strict qualitative reasoning could result in ambiguous values for process variables. A novel semiquantitative and order of magnitude reasoning methodology was developed using quantitative information in the form of the design specifications and normal operating conditions of the process units and the quantitative properties of process materials to filter and rank the adverse consequences determined by strict qualitative reasoning. A knowledge based system was developed, called HAZOPExpert. The HAZOPExpert systems was tested using five complex industrial case studies; the results were compared to those of a team of HAZOP experts. The HAZOPExpert identified all the abnormal causes and adverse consequences identified by the team of experts.
NIOSH-Grant; Grants-other; Accident-prevention; Safety-engineering; Hazardous-materials; Chemical-manufacturing-industry;
Chemical Engineering Purdue University West Lafayette, IN 47907-1283
Final Grant Report;
NTIS Accession No.
Other Occupational Concerns; Grants-other;
Laboratory for Intelligent Process Systems, School of Chemical Engineering, Purdue University, West Lafayette, Indiana
Purdue University West Lafayette, West Lafayette, Indiana