Abstract
The theory of the filtered Poisson process is applied to analyze the response of highway bridges to stochastic live loads. For each lane of the bridge, the live loads are idealized as a train of uniformly moving vehicles for which a statistical data base exists so that the probability distribution functions of the vehicular headways and weights can be established. In particular, the headways are assumed to have exponential distributions. The cumulants, and hence the mean value and variance, of any desired response quantity at any bridge location can then be obtained with the aid of the associated influence lines. On the basis of the cumulants thus obtained, the characteristic function of the response quantity can be constructed and used to obtain the corresponding probability density function by means of the Fast Fourier Transform (FFT) technique. Examples are given for the cases of a simply supported girder bridge and a three-span continuous girder bridge subjected to such stochastic live loads.
