Optimal Programming of In-situ Tests and Inspections for Existing Buildings

The seismic performance assessment of existing buildings is directly related to the amount of information and knowledge available. Such information can be either extracted from the original design documents and/or from the results of in-situ tests and inspections. The European code part 3 prescribes the minimum requirements for the number of tests that are to be performed for both inspection of structural details and testing of material properties for different knowledge levels. The code-based specifications, however, do not provide specific directions for the spatial distribution of the tests and inspections throughout the structure. This work strives to propose a Bayesian adaptive methodology for optimal programming of in-situ tests and inspections for probabilistic seismic assessment of existing buildings. Prior to performing the tests, a list of structural modeling parameters which are expected to significantly affect the global structural performance is compiled. For each structural modeling parameter in the list, the structural zone in which the parameter is going to be modeled as constant needs to be specified. Based on the information available for the building in question before the tests are performed, a prior probability distribution for the structural modeling parameter is constructed. The programming of tests and inspections is performed in a step-by-step manner in which, at each step, a decision needs to be taken whether to stop the tests or continue. An adaptive sampling criterion for the spatial distribution of the test/inspection points within the zone is established in order to efficiently capture the characteristics of the zone. The Bayesian framework is used in order to obtain a sequence of updated probability distributions based on increasing test results. The structural reliability is then updated as a function of the updated probability distributions. The decision to stop the tests may be reached when the structural reliability becomes invariant with respect to the increasing number of tests. The proposed adaptive methodology for the programming of tests and inspections is applied to an existing case-study reinforced-concrete RC building. The results obtained for the illustrated example are generalized in order to provide some insights for the optimal programming of tests and inspections for existing buildings.