An investigation of system identification and damage estimation using modal plots, count plots and a damage indicator

Abstract

Structural Health Monitoring (SHM) aims to detect slow and/or sudden changes in a structure, and the possible damage associated with such changes, using system identification and damage detection methodologies. It eventually also aims to provide an estimate for the remaining useful life of a structure and to provide guidance to structural design. In this thesis, new approaches are presented for system identification, damage detection and sensor deployment under operational loads. In the introduction chapter, the importance of the subject, four crucial questions for SHM system design, and a short literature review are discussed. In the second chapter of the thesis, a sensitivity based damage detection algorithm is developed. The performance of the proposed system identification and damage detection methods are investigated numerically. The numerical demonstration of the proposed damage detection method is investigated for the full measurement and restricted measurement cases at the end of the chapter. The proposed method is tested on a numerical six-story structure model for damage at one story and multiple story scenarios in the fifth chapter. In the third chapter which deals with system identification, a new interpretation and a new algorithm called the ‘modal plot’ are provided for ‘stabilization diagrams’; it is shown that, it is possible to automatically identify the natural frequencies and the mode shapes of a system by converting its modal plot to a ‘count plot’. The count plot approach could be viewed as an alternative to power spectrum analysis. A case study is given at the end of the chapter and the proposed method is used to investigate natural frequency changes due to damage for data taken from a real structure in Chapter 6. Chapter 4 presents detailed discussions on the performance of the proposed methods via three numerical examples, which show that using the proposed methodology, it is possible to detect damage using the first three modes. In the experimental study, it is proposed that modal zones are sensitive to presence of the damage and the damage indicator is used to estimate the damage location. In the concluding chapter, a general discussion of the methods proposed in this thesis is provided.