Parameterized models for prediction of lifetime bearing demands

A good replacement and maintenance policy for bridge bearings is essential for bridge integrity and functionality. This requires proper estimation of the bearing life expectancy which in turn is dependent on the working condition demands. The lifetime travel and peak displacement demands are highly sensitive to the loading, bridge geometry, and bearing properties. Hence, predicting when bearings should be replaced is difficult. To facilitate the decision making process, this study proposes prediction models for the annual demands of elastomeric bridge bearings. First, random bridge configurations (e.g. elements geometry, deck type, and number of spans) and random loading conditions (e.g. temperature profiles, earthquake records, and traffic loading scenarios) are generated using Monte Carlo simulation combined with Latin hypercube sampling, and the bearing demands are computed. Bearing demand prediction models are then developed via regression analysis and used to create a general fatigue loading protocol. This protocol can be used for testing and rating sample bearings. This would aid in predicting the bearing life expectancy, allowing for better replacement scheduling, and budget estimation. The application of the proposed demand prediction models for generating fatigue loading protocols is demonstrated through a case study.