Health assessment of beams - Experimental investigations

Experimental study on the health assessment of defect-free and defective beams recently completed at the University of Arizona is presented in this paper. The method can be extended to consider frame-type structures. The researchers at the University of Arizona proposed a finite-element-based time-domain system identification technique, known as ILS-UI, to detect defects at the local element level of structures. The algorithm does not need excitation information. The theoretical basis of the method is presented in the companion paper. The experimental verification of the method is presented in this paper. Defect-free and defective fixed-ended and simply supported beams are successfully identified by tracking the reduction of stiffness of elements. Initially, the optimal number of finite elements required to capture the dynamic behavior of a beam was studied. The ILS-SI method was then verified using computer generated theoretical response information. Experimental investigations were then carried out to verify the algorithm. Raw data obtained from the experiments failed to identify the elemental stiffness of the beam. Several post-processing techniques were used to remove many undesirable features from the records. The algorithm again failed to identify elemental stiffness of the beam even after the post-processing of the response data. Reasons behind the failure were then investigated. Errors in the amplitude and phase shifts were found to be the most critical factors for the non-convergence of the algorithm. Additional mitigation methods were then proposed to reduce the amplitude and phase shift errors that enabled the algorithm to successfully quantify the changes in the elemental stiffness. It can be concluded that the ILS-UI algorithm is validated using laboratory measured response information and it can be used as a method for detecting defects in structures.