Solving River Pollution Problems by Means of Fuzzy Fault Tree Analysis

摘要

In this work, an algorithmic procedure, based on Fault Tree Analysis (FTA) in its fuzzy version (to count for uncertainty), has been developed for solving river pollution problems. The main steps followed are: (a) determination of metrological requirements, necessary to maintain the present ecological status of the river, (b) combinatorial relaxation to obtain acceptable tolerance intervals for each ecological sub-system, (c) sensitivity analysis to quantify the impact of the parameter(s) values deviation from the normally expected values, (d) synthesis of the corresponding fault tree and assignment of weights, (e) FTA by using fuzzy input, (f) formation of representative alternatives by combining the most influential final events acting as ultimate causes that contribute to the occurrence of the 'top event', and fuzzy multicriteria ranking of them, and (g) sensitivity /robustness analysis, of the ranked first alternative and suggestions for corrective action. A case example is presented, where the top event is "high BOD at site G", downstream of site F where wastewater is discharged into the river. The input independent/explanatory variables are given as fuzzy trapezoid numbers and the results (obtained as crisp numbers, after defuzzification) are discussed.