Damage Identification Based on Curvature Mode Shape using Cubic Polynomial Regression and Chebyshev Filters

Global Congress on Construction, Material and Structural Engineering 2017

材料科学;土木建筑工程

Hasrizam, C.M.^1 ; Fawazi, Noor^1

Mechanical Precision Engineering, Malaysia-Japan International Institute of Technology (MJIIT), Universiti Teknologi Malaysia, Jalan Semarak, Kuala Lumpur

54100, Malaysia^1

关键词: Cubic polynomials;    Damage Identification;    Damaged structures;    Different damages;    False detections;    Free boundary conditions;    Gapped smoothing method;    Structure health monitoring;   

Others  :  https://iopscience.iop.org/article/10.1088/1757-899X/271/1/012091/pdf
DOI  :  10.1088/1757-899X/271/1/012091

来源: IOP

【 摘 要 】

Structure Health Monitoring (SHM) has been applied in various application such as aerospace, machinery and civil structures to maintain structure's safety and integrity. Gapped smoothing method (GSM) is most popular non-destructive identification (NDI) method due to its simplicity and did not require baseline data for comparisons. However, GSM is less accurate to detect wide size of damage in structure and cause false detection. Objective of this study is to propose a method to detect damage in structure using curvature mode shape data estimated from damaged structure and did not require data from undamaged structure. Finite element analysis (FEA) on a free-free boundary condition steel beam was carried out to demonstrate the feasibility of the proposed method that estimate undamaged curvature mode shape data using cubic polynomial regression (CPR) and Chebyshev filters (CF) methods. The results shows proposed method that used Chebyshev filters has better accuracy damage detection on wide notch compared to GSM. Although application of an interpolation and Chebyshev filters showed results with a high potential for overcoming the issue of false detection due to different notch size, however the proposed method still need refinement to better detection of different damage cases.