The application of Fuzzy logic to determine the failure probability in Fault Tree Risk Analysis

FTA is an appropriate tool for failure analysis and determining the failure rate. Sometimes, the lack of information on the process and main events, it is difficult to accurately determine the probability rate of the failures. The application of Fuzzy tool can be help-ful in such cases. In present study the Fuzzy logic was applied to determine the failures probability of a distillation unit in an oil refinery in its risk assessment using FTA. The method and the results of the reactor risk analysis are discussed in present paper. In this descriptive study, HAZOP was applied to determine the process failures. Its fault tree was determined then. The probability rate of those failures which had the required input data for fault tree was determined using their equations. Otherwise, Fuzzy logic was used. The Fuzzy process began with the selection of an expertise team and ended with determination of the probability rate for the route events. The equation recommended by Onisawa was used to determine the probability rate. Then the probability of top events was determined using the appropriate equation. The critical Minimal Cut Sets (MCS) were determined using Fussell-Vesely equation. In the end, the influence of deleting three critical MCSs alone and simultaneously was studied. The application of HAZOP led to determine 4 main and 13 end failures. From 13 end failures, 7 failures were Fuzzy and 6 were probabilistic. The results revealed that cutting of the most critical MCS (e.g. A.3.2) may increase the reliability of top event by 5.9% from 0.914 to 0.968. Also cutting of all three main critical MCSs (e.g. A.2.2, A.3.1 & A.3.2) may increase the reliability of top event by 8.4% from 0.914 to 0.991. The failure probability can be determined using Fuzzy logic when there is not enough data on process and top events.