LEARNING FROM ABNORMAL OCCURRENCES - RESULTS OF A FUZZY ANALYSIS
Aim is at understanding the existence and magnitude of learning effects from past operational experience in nuclear power reactors. Analysis refers to several thousands of abnormal occurrences collected from a sample of 43 nuclear units. Two main issues are addressed. First, is the formal description of abnormal occurrences and their history. The formal tool is provided by stochastic and fuzzy marked-point processes. Second, is the evaluation of the events to state their significance towards risk, compare different plants, and understand the importance of learning phenomena. To this purpose, accidental events are ranked. Owing to their fuzzy character, events are interpreted in terms of necessity and possibility measures of uncertainty. Finally, through repeated application of the rules of intersection and union, meaningful knowledge is extracted, as it is looked for.
Bibliographic Reference: PSA'87 INTERNATIONAL TOPICAL CONFERENCE ON PROBABILISTIC SAFETY ASSESSMENT AND RISK MANAGEMENT, ZUERICH (SWITZERLAND), AUG. 30-SEPT. 4, 1987 WRITE TO CEC LUXEMBOURG, DG XIII/A2, POB 1907 MENTIONING PAPER E 33656 ORA
Availability: Can be ordered online
Record Number: 1989126056700 / Last updated on: 1989-03-01
Available languages: en