A comparison of automatic fault tree construction with hazard and operability study

Hazard and operability study (HAZOP) and fault tree analysis are the methods most often used in the risk analysis in the process industry. Considerable efforts have been made in order to reduce the amount of work needed in a risk analysis and to improve the quality of the study at the same time. Automation of the fault tree construction process is one of the recent developments of greater interest. In the study manual HAZOP-analyses and automatically constructed fault trees were compared in the risk analysis of a black liquor feeding system and a subsystem in a phenol plant. The results show that the automatically constructed fault trees have not revealed all the hazardous deviations identified by HAZOP. However, the fault tree did identify some of the causes of deviations more accurately than the HAZOP-analysis. The main reason for unidentified deviations was incompleteness observed in the component models. This means that one still cannot trust the automatic fault tree construction as the only method. Detailed manual analyses are still needed in order to guarantee a high level of risk identification. In spite of the problems and defects in the automation fault tree construction, the method is already an interesting way to complet the results of other methods and to secure the quality of risk analyses.