SEGMENTAL Box GIRDER : EFFECT OF SPATIAL RANDOM VARIABILITY OF MATERIAL ON DEFLECTIONS

In a preceding paper (Bazant and Kim 1989), a method of probabilistic prediction of the confidence limits on long-time deflections and intemal forces of prestressed concrete segmental box-girder bridges was developed. The sampling method was used. and predictions were improved by Bayesian updating on the basis of short-time measurements. The entire analysis, including its Baye~ian aspect, was made according to the latin hypercube sampling scheme, whIch represents an efficient way to solve the problem. The analysis consIdered the uncertainties of shrinkage, basic creep, drying creep, environmental relative humidity, concrete strength, water-cement ratio, gravelcement ratio and cement content-altogether eight random material parameters of given distributions, assumed to be normal. One aspect of statistical variability, however, was neglected in the preceding work. It was assumed that although the material compliance function has a random multiplicative factor, the value of this factor is constant throughout the entire structure. This is a simplification. In reality, the mulIlplicative factor on the compliance exhibits a spatial variability; it is a random field in space. The purpose of this brief note is to extend the preceding formulation to include, in a simplified manner, the effect of spatial random variability of the compliance multiplicative factor. Development of the basic theory is not repeated from the preceding paper (Bazant and Kim 1989).