Structural damage detection based on stochastic subspace identification and statisticalpattern recognition: II. Experimental validation under varying temperature

Abstract

Although most vibration-based damage detection methods can acquire satisfactoryverification on analytical or numerical structures, most of them may encounter problemswhen applied to real-world structures under varying environments. The damage detectionmethods that directly extract damage features from the periodically sampled dynamic timehistory response measurements are desirable but relevant research and field applicationverification are still lacking. In this second part of a two-part paper, the robustness andperformance of the statistics-based damage index using the forward innovation model bystochastic subspace identification of a vibrating structure proposed in the first part havebeen investigated against two prestressed reinforced concrete (RC) beams tested inthe laboratory and a full-scale RC arch bridge tested in the field under varyingenvironments. Experimental verification is focused on temperature effects. Itis demonstrated that the proposed statistics-based damage index is insensitiveto temperature variations but sensitive to the structural deterioration or statealteration. This makes it possible to detect the structural damage for the real-scalestructures experiencing ambient excitations and varying environmental conditions.