A procedure ontology for advanced diagnosis of process systems

Large-scale complex process plants are safety-critical systems where both the offline hazard analysis and the real-time diagnosis are of great importance. Because of the great complexity and the large amount of information, a combination of analytical and heuristic methods, called hybrid methods are usually applied. In order to properly manage abnormal conditions, risk management procedures, such as HAZard and OPerability (HAZOP) analysis, or Failure Mode and Effects Analysis (FMEA) are applied to most of the process plants in an off-line manner. The results of such an analysis contain a vast amount of diagnostically relevant information that can also be used in a real-time fashion. The overall aim of our work has been to utilize this risk management information in advanced diagnostic systems. For this purpose, an agent-based diagnostic system has been proposed in that uses HAZOP-FMEA information and bidirectional reasoning to derive a diagnostic conclusion. The aim of this paper is to further advance the above hybrid diagnostic methodology by including the information contained in the operating, safety and control procedures and use this for diagnosis of complex process plants