Lee Meng Yong
Metric No: BK10110135
Topic: TUNING OF PZT GENERATED LAMB WAVE PROPAGATING ON A THIN PLATE STRUCTURE
Abstract: In the field of Structural Health Monitoring (SHM), a reliable and cost effective method is needed to ensure the structural integrity of aging structures. Current SHM methods are mostly passive, which means it does not continuously monitor the health of a structure, instead it performs check on structures based on scheduled time. To continuously monitor health of structures, a small, cheap and reliable method is needed. The use of smart material, Lead Zirconate Titanate (PZT) which is a type of ceramic brings two-fold advantage in SHM, firstly because it is small, cheap and reliable, and secondly it can act as both the actuator as well as detector. Lamb wave which is a type of surface wave which propagates on thin-plate structures has been gaining popularity due to its ability to propagate through a long distance. However, Lamb wave is a very complex wave due to its multimodal and dispersive nature. Thus, in order to perform crack detection, the wave speed of each mode of Lamb wave needs to be determined. The wave speed dispersion curve will first be obtained by my research teammate, Chuah Chong Wei. Using the wave speed obtained, I performed the tuning of Lamb wave to selectively activate only one mode of the Lamb wave (symmetric or anti-symmetric) using Matlab. The tuning of the Lamb wave in this study will be entirely based on the amplitude matching effect between the symmetric and anti-symmetric Lamb wave modes. From the wave speed dispersion curve obtained by my research teammate, below 1000 kHz, only the S0 and A0 mode exist. Thus, analysis will be done at frequencies up to 1000 kHz only. The plot of the strain response vs. frequency was first obtained using the result from the Matlab analysis. The plot shows that the strain response of both the Lamb wave modes will increase and decrease with different frequencies, and thus proving that the ‘sweet spot’ can be obtained at certain frequencies to activate only one Lamb wave mode. It was also found that the strain response of the symmetric mode was higher than that of the anti-symmetric mode, which thus more favorable in crack detection as higher strain response means it can propagate longer distance. From observation, a perfect sweet spot was selected at around 320 kHz where the strain of the symmetric mode peaks whereas the anti-symmetric mode was near zero. To confirm the result, an FEM simulation on a half plate model based on the laboratory setup was carried out by my other research teammate, Lee Sze Shan. The FEM result showed an almost similar strain response but with a slight difference, however the result agreed with the Matlab result whereby at 320 kHz, the strain symmetric mode strain response is high whereas the strain response for the anti-symmetric mode is almost zero. Further studies are needed for the tuning of Lamb wave above 1000 kHz, and to improve the analysis of the laboratory studies to enable analysis above 100 kHz.