Keywords: safety, failure assessment, fault tree analysis, fuzzy sets, bridge construction.

The number of construction failures or accidents have been increasing with the larger scale and complexity of construction processes and equipment operations, which contain numerous uncertainty factors. Safety is a major requirement in all types of construction operations. Failures in civil or architectural structures during and after construction have resulted in time and cost overrun, injuries, and fatalities. The reputations of the designers and contractors of failed facilities have often been ruined. Therefore, it is essential to prevent or reduce the possible occurrence of failures or accidents from happening. It is also of fundamental importance to utilize or develop useful techniques to assist contractors to effectively analyze system safety and diagnose potential fault origins which may lead to catastrophic consequences.

Conventional fault tree analysis (FTA), a logical diagram technique, is a well recognized procedure used for detecting potential faults and quantifying the failure probability of the system given exact probabilities of fault events [1]. It is often difficult to obtain past failure data from new environments, new components, and imprecise events such as human error. Thus, estimating exact fault probabilities of the system for the use in the conventional FTA is difficult when the prior known probability of each fault cannot obtain or fault events are imprecise.

Probability theory has been a dominant tool to analyze failure uncertainty during the past decades. Probability methods deal with uncertainty which is essentially random in nature such that the occurrence of failure is quantitatively assessed by a particular probability distribution function [2]. However, the assumption of probability distribution of failure can be obtained only when a sufficient amount of failure data is available such that the failure probability can be derived. Also, assessing undesirable fault events is often performed by experts based on their subjective and qualitative judgments. When a fault event is vague, experts usually express the possibility of the occurrence of the event as either "very low", "high", or "very high" rather than by the precise probability of the occurrence. Also, the degree of the uncertainty where a vague or fuzzy event leads to an accident or failure during construction is difficult to directly estimate the probability. To overcome these difficulties, fuzzy set theory and the possibility theory provide a useful tool in overcoming these problems. Based on fuzzy set theory and fuzzy logic, the fuzzy fault tree analysis (FFTA) approach is well suited for handling ambiguous and imprecise information obtained from safety experts [3,4]. However, complicated mathematical operations limit the extensive applications of FFTA models in practice [5].

This paper presents a new approach to integrating fuzzy sets and FTA to assess those fault possibilities that are essentially based on the expert's experience and subjective assessment. Fuzzy possibility scores (FPS) are obtained by using the techniques of max-min aggregation and center of sum defuzzification which are easy to manipulate. Fuzzy fault rates are converted from FPS by using a transformed possibility function. An illustrative example regarding the bridge failures during balanced cantilever construction was exemplified. The results demonstrate that the capability of the approach that can assist safety engineers to better analyze the failure of construction operations.

1

J.B. Fussell, "How to hand-calculate system reliability and safety characteristics", IEEE Transaction on Reliability, R-24, 169-74, 1975.

2

K.L. Cai, "System failure engineering and fuzzy methodology: An introduction overview", Fuzzy Sets and Systems, 83, 113-33, 1996. doi:10.1016/0165-0114(95)00385-1

3

D.I. Blockley, "Predicting the Likelihood of Structured Accidents", Proceedings of Institution of Civil Engineers, 59, 659-68, 1975. doi:10.1680/iicep.1975.3633

4

D. Singer, "A fuzzy set approach to fault tree and reliability analysis", Fuzzy Sets and Systems, 34, 145-155, 1990. doi:10.1016/0165-0114(90)90154-X

5

P.V. Suresh, A.K. Babar, and R.V. Venkat, "Uncertainty in fault tree analysis", A fuzzy approach, Fuzzy Sets and Systems, 83, 135-41, 1996. doi:10.1016/0165-0114(95)00386-X

purchase the full-text of this paper (price £20)

go to the previous paper
go to the next paper
return to the table of contents
return to the book description
purchase this book (price £105 +P&P)