Advanced Nondestructive Evaluation I

Abstract: The purpose of this study is to develop a portable electronic nose system to measure volatile components of agricultural and food products. Also, operating software to control the electronic nose system through the Internet was developed. Various experiments to find optimum operating conditions of the system were performed. An array of commercial metal oxide gas sensors was used to detect various gas components. For kochujang experiment, flavour signal patterns were different to the type of kochujang. Transient and steady state signals were analyzed. Transient signal analysis was more useful in PCA. Cluster analysis showed the possibility of reduction of the number of sensors. It is feasible to optimise the kochujang making process.

Abstract: This study was performed to show the feasibility of nuclear magnetic resonance (NMR) techniques for quality evaluation of various agricultural and food products. A real-time in-line NMR quality evaluation sensor was designed, constructed and tested. The device consists of an NMR spectrometer coupled to a conveyor system and a data acquisition system. The conveyor was run at speeds ranging from 0 to 300 mm/s. An NMR signal can be detected when a sample is within ±50 mm of the NMR coil center. The response of NMR sensor was tested using several fruits. The results showed a feasibility of an NMR sensor for evaluating internal quality of various fruits.

Abstract: A real-time white ginseng quality evaluation system based on a machine vision technique and artificial neural networks was developed to replace the current manual grading and its efficiency was tested. The system consisted of conveyor, image acquisition system synchronized with a sample-detecting sensor, and image processing and decision-making system. Software running under Windows system was developed. The algorithm included three consecutive stages of (a) image acquisition and preprocessing, (b) mathematical feature extraction, and (c) grade decision using artificial neural networks. Mathematical features such as area ratio, mean and standard deviation of gray level, skewness of gray level histogram, and the number of run segment, were extracted from five equally divided parts of a specimen. An artificial neural network model was used to classify samples into three grading categories. The grading error of the system was about 26%, which is comparable to the 30% in case of manual grading. The grading rate was one sample per a second.

Abstract: Compaction is becoming a great concern in crop production and the environment. Recently, three has been a need of field management based on site-specific conditions to improve sustainability of agriculture and reduce environmental damage. In the study, soil management or tillage depth was recommended nondestructively based on cone index profiles for typical Korean rice paddy fields. Field variables related to tillage, soil strength, rice growth, and other soil physical properties showed considerable spatial and vertical variations as well as significant (α<0.1) correlations among them. Cone index profiles observed also varied by field sites, and maximum cone index and depth to the maximum cone index showed significant (α<0.1) correlation with tilled depth as well as rice growth and other field variables. When soil management was recommended based on CI measurements, 13.4, 16.8, and 95.3% of the total surveyed areas, and 10.6, 18.9, and 51.6% of the total soil volume were chosen for management depth of 10, 20, 40 cm, respectively, indicating that soils of many field sites would not restrict rice growth. It was concluded that the concept of site-specific soil management based on soil conditions could save labor, time, machine use, and energy.

Abstract: Impulsive sound and vibration signals in gear system are often associated with their faults. Thus these impulsive sound and vibration signals can be used as indicators in condition monitoring of gear system. The traditional continuous wavelet transform has been used for detection of impulsive signals. However, it is often difficult for the continuous wavelet transform to identify spikes at high frequency and meshing frequencies at low frequency simultaneously since the continuous wavelet transform is to apply the linear scaling (a-dilation) to the mother wavelet. In this paper, the spike wavelet transform is developed to extract these impulsive sound and vibration signals. Since the spike wavelet transform is to apply the non-linear scaling, it has better time resolution at high frequency and frequency resolution at low frequency than that of the continuous wavelet transform respectively. The spike wavelet transform can be, therefore, used to detect fault position clearly without the loss of information for the damage of a gear system. The spike wavelet transform is successfully is applied to detection of the gear fault with tip breakage.

Abstract: Impulsive vibration signals in gearbox are often associated with faults, which lead to due to irregular impacting. Thus these impulsive vibration signals can be used as indicators of machinery faults. However it is often difficult to make objective measurement of impulsive signals because of background noise signals. In order to ease the measurement of impulsive signal embedded in background noise, we enhance the impulsive signals using adaptive signal processing and then analyze them in time and frequency domain by using time-frequency representation. This technique is applied to the diagnosis of faults within laboratory gearbox.

Abstract: In general, dynamic friction coefficient in a dry surface condition is dependent on the relative velocity between two sliding bodies. A conventional method of estimating the velocity dependent dynamic friction coefficient is time consuming and requires a special jig to measure the friction forces. In this paper, a new technique called the state space mapping method is proposed based on the nonlinear dynamics of a 1-DOF friction oscillator. One body is constructed as a single degree of freedom system and another body is formed as a moving base that may be built as a rotating disk or a horizontally oscillating plate. The resulting friction induced vibration system is used to estimate the dynamic friction coefficient. The measured vibration signals are utilized to construct a three dimensional plot of acceleration versus state variables. Then, the velocity dependent dynamic friction coefficient can be estimated from the plot. The natural frequency can also be estimated from the plot, which can be used to verify the quality of the results.

Abstract: The sensitivity of the microphone is determined absolutely by primary calibration technique based on the reciprocity principle. The free-field calibration has the inherent problem due to the weak signal received in the microphone. Multi-path noises such as cross-talk and reflections corrupt the received signal, and thus the sensitivity evaluation is poor. The time selective method has been widely used as the useful tool to get rid of such noises. However FFT based signal processing invokes unwanted processing error such as leakage. A well designed frequency window which would be a kind of low-pass filter would reduce the leakage problem, and it helps to separate multi-path noises from the directly propagated signal well. The result well shows the advantages in separating the noises and the availability in the free-field calibration with the enhanced sensitivity evaluation.

Abstract: Multi-reference, scan-based Nearfield Acoustical Holography (NAH) is a useful measurement tool that can be applied when an insufficient number of microphones are available to make measurements on a complete hologram surface simultaneously. The scan-based procedure can be used to construct a complete hologram by joining together sub-holograms captured using a relatively small, roving scan array and a fixed reference array. For the procedure to be successful, the source levels must remain stationary for the time taken to record the complete hologram: that is unlikely to be the case in practice, however. Usually, the reference signal levels measured during each scan differ from each other with the result that spatial noise is added to the hologram. A non-stationarity compensation procedure that is based on the acoustical transfer functions between the sources and both the reference and scanning, field microphones are invariable is introduced. Numerical and experimental results show well the availability of the introduced procedures to suppress the spatially distributed noise and to get better sound fields partially separated.

Abstract: The goal of system monitoring and diagnostics is to minimize economic losses, to increase stability, to maximize productivity, and to maintain product quality in manufacturing. The feature extraction from the signals acquired in rotating machine is required for performance evaluation, condition monitoring, and fault diagnostics. In this paper, we extracted the distinctive features from vibration signals gathered in rotor-bearing system during acceleration in order to monitor an abnormal condition by using various kinds of signal processing methods such as the Fast Fourier Transform, Short-Time Fourier Transform, Wigner-Ville Distribution, and Discrete Wavelet Transform.