Fatigue reliability assessment of an existing steel railroad bridge

Abstract

Systematic procedures to assess the reliability level and remaining fatigue life of an existing steel railroad bridge are demonstrated by examining critical bridge members. Critical members are selected among those members experiencing the maximum stress ranges higher than the corresponding endurance levels over two million cycles, due to the maximum loading condition. Applied loads on the bridge vary randomly, and this results from the uncertain nature of the input loading conditions, such as traffic schedules, passenger volumes, and so on. Only passenger volumes are assumed to be random, which are determined based upon the provided data and on-site visual observations. The processes to evaluate the remaining fatigue life are classified into three categories according to the given assumed loading models: (1) simplified procedure; (2) probabilistic procedure; (3) deterministic procedure. Stress time histories induced by the moving subway trains are obtained by using the Finite Element analysis, which is calibrated based on the field measurements. The coincidence rates of two trains at the same location are also considered and are simulated based upon the provided traffic schedules. Some members with peculiar details due to improper manufacturing have been found. A repairing method to improve the fatigue strength of the members is proposed. Corresponding fatigue tests have been conducted to verify the proper category for the members with peculiar details and the repaired members. Various illustrated results obtained by the proposed procedures are presented for comparison. It is found that the simple deterministic procedure may be adopted to get quite consistent information on the remaining fatigue life of a bridge with acceptable accuracy.