Computational model updating based on stochastic test data and modelling parameters - a tool for structural health monitoring

摘要

For structural health monitoring purposes model updating techniques are utilized for relating experimentally observed changes of the modal data (natural frequencies and mode shapes) to the changes of the physical parameters of a Finite Element model, e.g. the bending stiffness of a beam element. Such changes should indicate any significant deviation of the healthy state from the actual state of the monitored structure. In the contribution an extension of the classical model updating technique is applied. The extension addresses the assumption that the Finite Element modeling parameters as well as the test data are no longer treated as deterministic but as band limited stochastic variables. Some results are reported from an application to the Gaertnerplatz Bridge in Kassel where the experimental modal data are continuously measured in the framework of a structural health monitoring project.