Reliability-based criteria for proof load testing of bridges.

Current AASHTO bridge design and evaluation codes do not specify any criteria for proof load testing of bridges. The objective of this study is to develop a reliability-based rational procedure for determining the optimum proof load. As part of the study, proof load tests were performed on five medium span girder bridges. These bridges are older (60-80 years) and heavily deteriorated. M-60 military tanks were used as proof load. Small stresses and a linear response were considered as an indication of extra safety reserve. The test results are used to update the resistance model of each bridge. The effect of proof load test is to increase reliability index by truncating the lower tail of resistance distribution. The probabilistic model for live load is based on the actual weigh-in-motion measurements. A failure criterion is established for each bridge and then the reliability index is calculated before and after the proof load test. The proof load factors are calculated using iterative procedures so that they provide the post-test target reliability index. Based on this study a target proof load factor of 1.4 is proposed, which corresponds to a post-test target reliability of 3.5. A step-by step procedure to carry out the proof load testing of bridges is presented. A detailed sensitivity analysis is performed to check the effectiveness of the proposed proof load factor to incorporate the influence of several other important factors such as safety of the bridge during the test, reduction of reliability index due to corrosion over design life, and system behavior. The concept of proof sampling of bridges is also demonstrated by using Bayesian updating to show the effect of proof load testing on the reliability indices of similar but untested bridges. Results indicate that proof sampling is effective even for low degree of correlation and only small samples need to be tested.PhDApplied SciencesCivil engineeringUniversity of Michigan, Horace H. Rackham School of Graduate Studieshttp://deepblue.lib.umich.edu/bitstream/2027.42/130822/2/9811181.pd