Application of Time-Frequency Analysis for Impulse Response of Melons
Date Issued
2007
Date
2007
Author(s)
Tang, Yi-Kai
DOI
zh-TW
Abstract
Acoustic impulse response method for measuring natural frequency of an impacted fruit is a nondestructive textural quality evaluation method extensively studied in the literature. The method used various types of pendulum to hit the studied object and to record the acoustic or vibrating responses. The recorded data were then processed by means of digital Fourier transform to obtain the natural frequency of the object. The main theory behind is assuming fruit to be an elastic object that response to the impaction, although fruit fresh is viscoelastic in reality. The objective of this thesis was to use time-frequency method to analysis impulse response for a melon in an impaction test. Both acoustic and vibrating responses were recorded and processed with a short time Fourier transform program developed in this study to give spectrographs for each impulse response data obtained in experiment. The spectrograph gave a better resolution for the natural frequencies also with the time history of the power spectrum. By this improved analysis tool the changes of mechanical properties of melon during short storage were easily depicted in decreasing of major nature frequencies and variations of power in different wave bands. Those were two main parameters separately discussed in the literature. Abstracting information from the spectrograph, the time history of each natural frequency was obtained. Skew curves were observed for low frequency signal, the graphics were similar to those in stress relaxation curves for the viscoelastic materials. Along with other evidences that suggested time-frequency method might be a new tool for studying viscoelastic properties of fruit.
Subjects
時頻域分析
聲譜圖
水果
敲擊檢測法
Time-Frequency Analysis
spectrogram
fruit
acoustic impulse response
Type
thesis
File(s)![Thumbnail Image]()
Loading...
Name
ntu-96-R94631038-1.pdf
Size
23.31 KB
Format
Adobe PDF
Checksum
(MD5):fa1d840c7d84a33dacb61f2c9fde7971
