REASSESSING RELIABILITY BASED ON SURVIVED LOADS

Abstract

The reliability of flood defenses is often dictated by large uncertainties in the hydraulic loading and the structural resistance. Additional information decreases uncertainty, however, acquiring it is often costly. One source of information, even though in many cases readily available, is hardly used - survived loads. This article shows how data on survived load conditions can be incorporated in reliability analysis by means of Bayesian techniques. The theory is illustrated by simple and realistic examples. In contrast to other sources of information, reassessing structures using survived load data always leads to higher reliability or lower probability of failure. Furthermore, attention is given to the expected development of failure in time. This may be relevant for situations, where the safety requirements of a structure are stated in terms of a design or inspection period. For both, re-assessing reliability based on (one time) survived loads as well as the expected increase of reliability in time, the examples show significant impact. Use of this knowledge, may, consequently, safe cost of construction or reinforcement.