Loading...

Detection of the Location and Severity of Damage in Operational Structures Using the Sensitivity of Transfer Functions Derived from Response-Based Methods

Nourafkan, Mahdi | 2024

0 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 57327 (09)
  4. University: Sharif University of Technology
  5. Department: Civil Engineering
  6. Advisor(s): Bakhshi, Ali
  7. Abstract:
  8. In this research, using sensitivity equations and output-only modal analysis (OMA) methods, a finite element numerical model was updated, and in the updated model, damage identification was performed. The present study, utilizing a sensitivity-based updating method and a validated formulation from the technical literature, conducted the updating process under several different approaches and ultimately verified the fundamental assumptions in the OMA field to demonstrate the validity of the results. The combination of these elements, along with the use of an explicit formulation, establishes a framework for updating finite element models within the OMA field, which is introduced as the innovation of this research. The case study of this research is the structure of a bell tower in Italy, which was monitored under environmental conditions using accelerometer sensors. In this thesis, to identify the dynamic parameters of the studied structure, output-only methods were used, and its modal characteristics were calculated using one of the OMA methods known as Frequency Domain Decomposition (FDD). Then, using the data, tests, and surveys conducted on the bell tower, an initial finite element numerical model of this structure was created in the SAP2000 software. This numerical model was updated based on the results of signal processing recorded by the sensors and coding in MATLAB software. The connection between the SAP2000 and MATLAB software was established using API programming principles. The updating process was carried out using the proposed formulation in four different approaches. The differences between these approaches lie in how the objective function is defined and the number of optimization variables and modes considered for updating. To determine the updating parameters, the most effective parameters in the structural response were selected. The candidate parameters as optimization variables were filtered using sensitivity analysis with the finite difference method, and the most sensitive parameters were considered for participation in the updating and optimization process. The results of the updating show that when four updating variables were used, changing the cost function from natural frequencies and mode shapes to frequency response reduced the average difference between the natural frequencies of the updated model and the real structure by 0.452%, resulting in a more accurate numerical model. This change was 0.306% when six updating variables were used. Additionally, when the cost function was composed of natural frequencies and mode shapes, increasing the optimization variables from four to six reduced the average difference between the natural frequencies of the numerical model and the real structure by 0.556%. If the cost function was composed of frequency response, this reduction would be 0.41%. Finally, assuming damage in the most accurate numerical model prepared at this stage, the outputs of the numerical structure were extracted exactly at the sensor installation locations, and assuming no access to any damaged model and only having the sensor time history outputs, the location and severity of the damage were identified. The results of damage identification using the proposed method demonstrate high accuracy
  9. Keywords:
  10. Structural Health Monitoring ; Finite Element Model Updating ; Operational Modal Analysis ; Output-Only Modal Analysis ; Damage Detection ; Frequency Domain Decomposition

 Digital Object List

 Bookmark

No TOC