INTERACTION OF SIZE EFFECT AND RELIABILITY OF DESIGN: CASE STUDY OF FLEXURAL STRENGTH OF CONCRETE

In spite of the fact that size effect phenomena have been extenSively studied and their importance has been amply demonstrated, the size effect is still not considered in most design specifications. The aim of the paper is to demonstrate the influence of size effect on the reliability of design. This is done by comparing the effect of the dead and live load factors as specified in the design codes. The so-called hidden size effect implied by the currently used dead load factor is reviewed. With the help of a simple example of flexural strength of a concrete beam failing at crack initiation, it is shown that the structure size may have a drastic influence on the failure probability, and particularly that even large load factors cannot play the role of a "safety bell". Further it is explained that the size effect at fracture initiation must have been a significant contributing factor in many catastrophic structural failures, e.g., the collapse of Schoharie Creek Bridge caused by fracture of an unreinforced foundation plinth. Because of its simplicity, the example of flexure of an unreinforced beam is helpful for showing the general methodology. The method of realistic prediction of the interaction of size effect and reliability using computational modeling based on nonlinear fracture mechanics combined with structural reliability approaches is briefly discussed. Finally, critical comments are made on the AASHTO LRFD Bridge Design Specifications l , particularly on the limit state Strength IV which characterizes the case of very high dead-to-live load effect ratios, which is typical for large-span bridges.