Loading...

Developing Hybrid Das-Sr Method for Structural Health Monitoring Using Guided Lamb Waves and Sparse Array of Pzt Transducers with Experimental Validation

Nokhbatolfoghahai, Ali | 2020

431 Viewed
  1. Type of Document: Ph.D. Dissertation
  2. Language: Farsi
  3. Document No: 53666 (45)
  4. University: Sharif University of Technology
  5. Department: Aerospace Engineering
  6. Advisor(s): Mohammad Navazi, Hossein
  7. Abstract:
  8. The main purpose of this dissertation is to investigate and improve the efficiency of Sparse Reconstruction method for structural health monitoring application in order to solve the challenges including; finding multiple simultaneous damages, damage detection and location in anisotropic structures and real engineering complex structures. The main approach of this research is based on presenting theoretical concepts and developing practical methods based on the presented concepts, as well as evaluating and implementing the developed methods using numerical simulations and experimental tests. In this work, the finite element method is used for numerical simulation. In the experimental test, PZT transducers are used as sensor-actuators and the test specimen is a simple aluminum plate at the first step and then quasi-isotropic composite plate, then simple anisotropic composite plate and finally anisotropic composite sheet with stiffener are used at the next steps. Also, the defects are simulated by a glued-on steel rod with 5 mm diameter at the first step and then real defects created by making a 5 mm cut-damage in the aluminum plate and by placing barely visible impact damage on the composite plate with a low energy impact. First, the accuracy and robustness of the recently developed Spars Reconstruction and the well-known Delay-and-Sum methods was evaluated and compared with each other, using numerical simulations under different environmental and operational conditions. For multi-parameter evaluation of structural health monitoring imaging methods, a methodology based on the Taguchi and ANOVA statistical methods was proposed to evaluate the simultaneous effect of several variable parameters. The results showed that the two methods of Sparse Reconstruction and Delay-and-Sum can complement each other in many situations. In the next step, according to the feasibility study performed in the first step, the hybrid DAS-SR method is developed and its application in finding several simultaneous damages as well as using low sampled signals is studied on the aluminum plate using Numerical simulations and experimental tests. In the next step, the Delay-and-Sum and Sparse Reconstruction methods were modified and the hybrid DAS-SR method was further developed to be applied in anisotropic composite structures. At this stage, the efficiency of the modified methods for multi-location damage detection as well as detection and location of barely visible impact damage, was evaluated using experimental implementation. In the last step, to increase the performance of the sparse reconstruction method in real complex engineering structures, a dictionary learning framework proposed to update the dictionary matrix adaptively, which allows the initial estimated dictionary matrix to be adapted to be compatible with the complex structure. The performance evaluation of the proposed dictionary learning framework was performed by experimental implementation on an anisotropic composite pate with a stiffener. The obtained results demonstrated the applicability of the developed hybrid DAS-SR method for finding multiple damages, optimizing the computational cost and health monitoring of anisotropic composite structures. Also, the results confirmed the concept of the proposed dictionary learning frame work to use for health monitoring of complex structures
  9. Keywords:
  10. Structural Health Monitoring ; Dictionary Learning ; Lamb Wave ; Sparse Recovery ; Complex Composite Structures ; Delay-And-Sum (DAS)Method ; Hybrid Delay-And-Sum (DAS) with Sparse Reconstruction (SR)Method

 Digital Object List

 Bookmark

No TOC