Safety Analysis Technique for System with Limited Data: Case Study of the Multipurpose Research Reactor in Indonesia

Abstract

Fault tree analysis (FTA) is frequently applied to deductively evaluate the safety systems of complex engineering systems such as chemical industries or nuclear facilities. To perform this analysis, generic data are commonly used due to the limitation of historical failure data of the system being evaluated. However, generic data have a degree of uncertainty and hence cannot represent the system’s actual performance. In addition, generic data are not applicable to older components due to the aging process, which obviously degrades the reliability of those components. To deal with this limitation, another safety analysis method, called fuzzy fault tree analysis (FFTA), has been proposed. The purpose of this study is to apply FFTA to evaluate the performance of the primary cooling systems of G.A. Siwabessy Multipurpose Reactor (RSG-GAS). RSG-GAS is a research reactor, which belongs to the National Nuclear Energy Agency of Indonesia (BATAN). Expert justifications were used to evaluate the failure occurrences of basic events in the primary cooling system of the RSG–GAS through questionnaires. The assessment by experts is in the form of qualitative data, which are then converted into quantitative data by applying FFTA. Then, the top event probability generated from FFTA was applied to calculate the event probability using event tree analysis (ETA). It was obtained that the highest event probability was 4.304 × 10−8/year. Since it complies with The International Atomic Energy Agency (IAEA) specified core damage frequency (CDF) limit, i.e., not greater than 10−5/year of reactor operation, the reactor is safe to operate.