Defect Size Estimation Research Articles

Signal processing techniques for rolling element bearing spall size estimation

Abstract This paper proposes signal processing techniques to extract entry and exit points from the bearing vibration signal for spall size estimation. The entry point is the start point of a low frequency response when a rolling element enters a localized defect on the raceway, which is contaminated with background noises and difficult to identify. An empirical model based signal processing method is proposed to effectively identify the entry point. The Variational Mode Decomposition (VMD) is applied to the bearing entry signal for more accurate estimation. Differentiation technique is used to identify the high frequency exit point with more reliable threshold values for automatic diagnostics. Then, based on defect size estimation models for both inner and outer race defects, the spall size can be estimated. The proposed methodology is validated on a machine tool spindle’s bearing system. Experimental results show that the proposed signal processing techniques provide less biased results with respect to spindle speed and more precise estimation.

Fatigue limit of a cast Al-Si-Mg alloy (A357-T6) with natural casting shrinkages using ASTM standard X-ray inspection

This work aims to quantify the effect of natural shrinkages on the fatigue limit of A357-T6 cast aluminum alloy. The main objective is to characterize the fatigue strength of the alloy following the ASTM X-ray standard defect identification, by correlating a different ASTM indication to a fatigue limit reduction. The analysis has been conducted on six customized castings with two defect types (cavity and sponge shrinkages) and a different harmfulness ASTM grade as defined by Non Destructive Testing (NDT) standard. The effect of the defect size has been correlated with the ASTM grade and a Kitagawa-Takahashi diagram has been generated for surface defects. A micro Computed-Tomography (μ-CT) analysis has been performed in order to compare the defect size estimation using two equivalent parameters AREA1/2 and VOLUME1/3. A Defect Stress Gradient (DSG) approach has been proposed to simulate the Kitagawa-Diagram obtained for surface defects. Finally, the effect of the specimen surface has been investigated for the reference material (ASTM grade < 1) by comparing the “as cast” surface to a machined surface finish.

Detection of air leakage into vacuum packages using acoustic measurements and estimation of defect size

Abstract Air leakages in food and ingredient packages which are sealed in vacuum environments may cause a marked deterioration of the product, leading to a loss of functionality. Manufacturers of such products have very stringent but rather costly quality control procedures and there is a pressing need for developing more economical ways of automated quality control techniques to test the vacuum packages reliably. However, due to the fact that the defect size of a typical package with a leakage problem could be micro- or nano-scale, such faults are not detectable using conventional techniques. In this paper, the performance of a proposed acoustic method is assessed for the detection of air leakage in instant dry yeast packages sealed in a vacuum environment, which are typical of food and ingredients packaged under vacuum conditions. The investigation is carried out in both laboratory and in-situ environments. The acoustic pressure created by leaking air into the faulty packages is measured using a low-noise microphone in an acoustic chamber. Faulty packages are then identified using the changes in measured sound pressure levels within a certain frequency band. A mathematical model is also proposed to predict the pressure inside a yeast package with certain defect size as a function of time. The mathematical model is then used to determine the size of a defect causing the leakage, using the time required for the pressure inside a faulty yeast package to reach to a threshold level. The results of this investigation show that, using the state of the art measurement techniques, it is possible to detect packages with leakage problem if the diameter of the defect is greater than a few tens of micrometres.

A new model for rolling element bearing defect size estimation

A new model based on the Hertzian contact theorem is proposed to estimate the size of a line spall defect located on the bearing’s outer race. The entry point can be determined from the ball-race geometry relation, while the exit point can be identified from the time domain signal. Therefore, the defect size can be estimated from the vibration signal without requiring additional bearing load and stiffness measurements. Experiments were performed on a 3-axis CNC machine tool at speeds ranging from 500 to 3000rpm and three line spall defects were estimated using the proposed method. The proposed model is demonstrated to estimate defect size with minimal speed related error, offset error, and standard deviation.

Enhancing Weak Defect Features Using Undecimated and Adaptive Wavelet Transform for Estimation of Roller Defect Size in a Bearing

ABSTRACTDefects in bearings affect the vibration level, resulting in and increase in temperature and decomposition of lubricant. Estimation of roller defect size is a complex task because it revolves as well as rotates during the motion. Signals from a defective roller of a bearing are superimposed by the signal from races, cage, and background noise. In this communication, a signal processing scheme is proposed that makes the signal suitable for estimating the size of the defect in the rolling element of a tapered roller bearing. To achieve this, in the first stage of processing, shift-invariant soft thresholding is applied to denoise the signal. It suppresses the noise without affecting defect-related features. Further, in the second stage of processing, continuous wavelet transform (CWT) using adaptive wavelet is applied. The adaptive wavelet is designed from the impulse extracted from the signal using the least squares fitting method. It results in higher coefficients in the region of impulse produced due to the defect. Finally, time marginal integration (TMI) of CWT coefficients is carried out for estimation of defect width. A study was performed for six different cases in which the size of the defect and orientation varies. Results of measurements of roller defect widths estimated using the proposed scheme were compared with defect widths calculated using image examination. For the nonoverlapping signature of defects (such as defects at 0° and 90° orientations), the maximum deviation in the width measurement using the proposed scheme is 6.52%. The error may increase when signature of two defects are overlapped.

Defect position and size estimation in billet by ultrasonic transmission method using a small number of transducers

For a steel billets inspection, we have proposed defect detection and size estimation from time-of-flight (TOF) profile by ultrasonic transmission method with linear scanning. In this method, a defect is detected by deviation of TOF of transmitted wave caused by diffraction at the defect. In addition to defect detection, defect size can be estimated from deviation of TOF because TOF is affected by defect size. However, deviation of TOF is also affected by the depth of the defect, which cannot be precisely estimated only by linear scanning of a pair of transducers. In this study, we propose defect position and size estimation using multiple pairs of transducers and validity of the method was evaluated. By using multiple pairs of transducers, TOF of inclined paths, where transducers is not facing each other are also obtained as well as parallel paths, so that the position of defect can be estimated. As a result, the defect position could be estimated by using two pairs of transducers and defect size could also be estimated correctly compared with conventional linear scanning method.

A defect size estimation method based on operational speed and path of rolling elements in defective bearings

This paper describes the effect of inertia and centrifugal force that act on a rotating rolling element in a defective bearing, on the measured vibration signature. These effects are more pronounced as the speed of components increases. Significant speed-dependency of the characteristic events that are generated at the angular extents of the defect are shown by simulation and experimental measurements. The sources of inaccuracy and the speed-dependency in the existing defect size estimation algorithms are explained. The analyses presented in this study are essential to develop accurate and reliable defect size estimation algorithms. A complete defect size estimation algorithm is proposed that is more accurate and less biased by shaft speed when compared with existing methods.

Effect of mode conversion on defect detection and size estimation in billet from time-of-flight profile by ultrasonic transmission method

The effect of mode conversion on defect detection and size estimation in a billet from the time-of-flight (TOF) profile is evaluated by comparing the results of wave propagation simulations in an elastic wave field and a scalar wave field, in which only longitudinal waves propagate. As a result, the effect of mode conversion appears in the TOF profile Δτ and received waveform. TOF deviation increases when mode conversion is considered compared with considering only longitudinal waves. This tendency compensates the error in the previous experiment and simulation results. The consideration of the shear components will make simulations and defect size estimations more accurate.

Assessment of Ventricular Septal Defect Size and Morphology by Three-Dimensional Transthoracic Echocardiography

Morphologic description of ventricular septal defect (VSD) is mandatory before performing the newly developed transcatheter closure procedure. Inaccurate estimation of defect size has been reported using conventional two-dimensional (2D) transthoracic echocardiography (TTE). The aim of this study was to assess VSD morphology and size using three-dimensional (3D) TTE compared with 2D TTE and surgery. Forty-eight children aged 21.4±29.3months with isolated muscular (n=11 [22.9%]) and membranous (n=37 [77.1%]) VSDs were prospectively included. Three-dimensional images were acquired using full-volume single-beat mode. Minimal diameter, maximal diameter, and systolic and diastolic VSD areas were measured from 3D data sets using multiplanar reconstruction mode (QLAB 9). Maximal-to-minimal VSD diameter ratio was used to assess VSD geometry. Linear regression analysis and the Bland-Altman method were used to compare 3D measurements with 2D and surgical measurements in a subgroup of 15 patients who underwent surgical VSD closure. VSD 3D diameters and areas were measured in all patients (100%; 95% CI, 92.6%-100%). Maximal diameter was lower on 2D TTE compared with 3D TTE (7.3 vs 11.3mm, P<.0001). Mean bias was 4mm, with 95% of values ranging from -1.76 to 9.75mm. Correlation between 3D maximal diameter and surgical diameter was strong (r(2)=0.97, P<.0001), while correlation between maximal 2D diameter and surgical diameter was moderate (r(2)=0.63, P<.0001). VSDs had an oval shape when assessed by 3D TTE. Maximal-to-minimal diameter ratio assessed by 3D TTE was significantly higher in muscular VSDs compared with membranous VSDs (3.20±1.51 vs 2.13±1.28, respectively, P=.01). VSD area variation throughout the cardiac cycle was 32% and was higher in muscular compared with membranous VSDs (49% vs 26%, P=.0001). Three-dimensional TTE allows better VSD morphologic and maximal diameter assessment compared with 2D TTE. VSD shape and its changes during the cardiac cycle can be visually and quantitatively displayed. Three-dimensional echocardiography may thus be particularly useful before and during percutaneous VSD closure.

Detection and recovery of fault impulses via improved harmonic product spectrum and its application in defect size estimation of train bearings

Abstract The detection and recovery of impulsive signature play a vital role in the diagnosis and prognosis of rolling element bearings. Though different approaches have been proposed to deal with this problem so far, challenges still exist when they are applied to the bearings operating under harsh working conditions. The difficulties mainly come from the multi-resonance and multi-modulation characteristics of bearing vibration signals. To overcome this limitation, a new methodology for the detection and recovery of fault impulses is presented in this paper. First, an improved harmonic product spectrum (IHPS) is proposed to detect and identify the multiple modulation sources buried in a vibration signal. With this method, the fault-related impulsive features could be recognized, while the influence caused by non-fault modulation is eliminated. On this basis, a harmonic significance index is further established to quantify the diagnostic information contained in a narrow band signal. By utilizing this index, the optimal resonance band where the fault impulses are most significant could be accurately determined. Finally, IHPS and sideband product spectrum are integrated to reduce the in-band noise and further recover the fault impulses. The performance of this method is evaluated by both simulated data and real vibration data measured from a train wheel bearing with a naturally developed defect. Compared with Kurtogram and Protrugram, the proposed method can detect the resonance band more precisely even in the presence of heavy noise and other impulsive vibration sources. Moreover, with the impulses recovery scheme, the double impact phenomenon caused by a distributed defect is extracted successfully. Benefiting from this, the defect size of a bearing can be estimated from its vibration signal without dismantling, which makes it a promising tool for the bearing diagnosis and prognosis in industrial applications.

ニューラルネットワークを利用した超音波探傷試験の欠陥性状同定に関する研究

In recent year, the extremely thick plate becomes used in the deck plate of the large ships. To prevent the brittle fracture it is important to detect and measure the inner defects. Ultrasonic testing is used to find defects in structures. However, the precise estimation of defect shape and size is difficult through ultrasound testing of welds; currently, such high precision estimation depends on the experience and skill of the inspecting technician. This study aims to develop a method to quantitatively identify the shape of defects easily, quickly, and with high accuracy from the wave reflected from a defect during an ultrasonic test; in addition, such a method should not require the knowledge and experience of an expert. For this purpose, a neural network that uses numerical simulation results as training data was developed, and a defect identification system based on ultrasonic wave information was investigated. To confirm the usability and estimation accuracy of this system, measurement data from actual ultrasonic tests were used as the input in the defect identification system.

The path of rolling elements in defective bearings: Observations, analysis and methods to estimate spall size

This paper describes the experimental investigation of the vibration signature generated by rolling elements entering and exiting a notch defect in the outer raceway of a bearing. The vibration responses of the bearing housing and the displacement between the raceways were measured and analyzed. These key features can be used to estimate the size of the defect and is demonstrated in this paper for a range of shaft speeds and bearing loads. It is shown that existing defect size estimation methods include assumptions that describe the path of the rolling elements in the defect zone leading to poor estimates of defect size. A new defect size estimation method is proposed and is shown to be accurate for estimating a range of notch defect geometries over a range of shaft speeds and applied loads.

Defect detection and size estimation in billet from profile of time-of-flight using ultrasonic transmission method with linear scanning

In this study, defect detection and size estimation in billet by transmission method with linear scanning were carried out and the validity of the method was evaluated by numerical simulation. In addition, the suitable signal frequency and aperture of transducers were clarified. As a result, the following were found: a defect can be detected, signals with frequencies lower than those generally used in conventional ultrasonic testing (i.e., 0.5–1.5 MHz) are desirable, and the time-of-flight (TOF) deviation Δτ becomes largest when the wavelength at center frequency and the aperture of transducers are comparable. Defect size can be estimated when a single defect exists alone and the defect is not near the surface of a billet. Although defect size estimation becomes difficult when the defect is near the surface of a billet, the defect can be detected by our proposed method.

Comparison of 3-Dimensional Computed Tomography–Based Measurement of Glenoid Bone Loss With Arthroscopic Defect Size Estimation in Patients With Anterior Shoulder Instability

The purpose of this study was to compare four 3-dimensional (3D) computed tomography (CT) methods of measuring glenoid bone loss with the arthroscopic estimation of glenoid bone loss. Twenty patients with recurrent anterior shoulder instability underwent bilateral shoulder CT scans and were found to have glenoid bone loss. Arthroscopic estimation of glenoid bone loss was performed in all patients. Three-dimensional CT reconstruction was performed on the CT scans of each patient. The glenoid bone loss of each patient was measured using the surface area, Pico, ratio, and anteroposterior distance-from-bare area methods. The mean percent loss calculated with each method was compared with arthroscopy to determine the reliability of arthroscopy in the measurement of glenoid bone loss. The mean percent bone loss calculated with arthroscopic estimation, surface area, Pico, ratio, and anteroposterior distance-from-bare area methods was 18.13% ± 11.81%, 12.15% ± 8.50% (P= .005), 12.77% ± 8.17% (P= .002), 9.50% ± 8.74% (P < .001), and 12.44% ± 10.68% (P= .001), respectively. Repeated-measures analysis of variance showed that the 3D CT methods and arthroscopy were significantly different (F4,76= 13.168, P= .02). The estimate using arthroscopy is 55% greater than the average of the 3D CT methods. Our findings suggest that arthroscopy significantly overestimates glenoid bone loss compared with CT and call into question its validity as a method of measurement. A more internally consistent and accurate method for the measurement of glenoid bone loss is necessary to appropriately diagnose and treat shoulder instability. Level IV, case series.

Classification and Size Estimation of Wafer Defects by Using Scattered Light Distribution

SUMMARYUsually, wafer inspection is a two‐step process. The first step, which is very important, consists of detection of defects on the wafer. The second step classification of the detected defects, which is done after defect detection as necessary. Recently, it has become necessary to perform simultaneous detection and classification of wafer defects in order to reduce the wafer inspection time while maintaining high resolution of inspection. Optical wafer inspection is the most effective method for detecting and classifying defects and estimating their size in a short time. In this method, the scattered light distribution is compared with a database of previous scattered light distributions and pattern matching is used to classify defects and estimate their size. Therefore, to achieve optical wafer inspection based on scattered light distribution, it is necessary to gather scattered light distribution data for many samples of various kinds. The first aim of this study is to develop a defect classification and size estimation method utilizing the scattered light distribution. We propose a method of classifying defects on wafers and estimating their size. The proposed method is based on pattern matching utilizing a parametric eigenspace. The second aim of the paper is to validate the proposed method by applying it to scattered light distribution samples. The light scattering distribution samples were generated by discrete dipole approximation (DDA) simulation. A database of the light scattering distribution is easily created using the simulation.

Magnetic SANS study of a sintered Nd–Fe–B magnet: Estimation of defect size

The magnetic microstructure of an isotropic sintered Nd–Fe–B permanent magnet was studied at room temperature by means of unpolarized magnetic-field-dependent small-angle neutron scattering (SANS). Although the microstructure of the material consists of micron-sized Nd2Fe14B crystallites that are separated by a (low volume fraction) nanometer-sized Nd-rich phase, the SANS signal contains a significant contribution due to spin misalignment, which is evidenced by an enhanced scattering signal along the direction of the applied field. The origin of this spin misalignment is presumably related to the perturbing effect of the Nd-rich interphases. Analysis of the SANS data in terms of the autocorrelation function of the spin misalignment provides an almost field-independent length scale lC≅40nm, which, for soft magnetic materials, is usually taken as a measure for the size of inhomogeneously magnetized regions around lattice imperfections. Comparison of our results with a phenomenological expression for lC based on micromagnetic theory suggests that the large anisotropy field of the Nd2Fe14B phase (which is of the order of 8T) reduces the size of gradients in the spin microstructure, so that, for hard magnetic materials, lC represents the (field-independent) size of the characteristic defect in the microstructure.

Varying stiffness and load distributions in defective ball bearings: Analytical formulation and application to defect size estimation

This paper presents an analytical formulation of the load distribution and varying effective stiffness of a ball bearing assembly with a raceway defect of varying size, subjected to static loading in the radial, axial and rotational degrees of freedom. The analytical formulation is used to study the effect of the size of the defect on the load distribution and varying stiffness of the bearing assembly. The study considers a square-shaped outer raceway defect centered in the load zone and the bearing is loaded in the radial and axial directions while the moment loads are zero. Analysis of the load distributions shows that as the defect size increases, defect-free raceway sections are subjected to increased static loading when one or more balls completely or partly destress when positioned in the defect zone. The stiffness variations that occur when balls pass through the defect zone are significantly larger and change more rapidly at the defect entrance and exit than the stiffness variations that occur for the defect-free bearing case. These larger, more rapid stiffness variations generate parametric excitations which produce the low frequency defect entrance and exit events typically observed in the vibration response of a bearing with a square-shaped raceway defect. Analysis of the stiffness variations further shows that as the defect size increases, the mean radial stiffness decreases in the loaded radial and axial directions and increases in the unloaded radial direction. The effects of such stiffness changes on the low frequency entrance and exit events in the vibration response are simulated with a multi-body nonlinear dynamic model. Previous work used the time difference between the low frequency entrance event and the high frequency exit event to estimate the size of the defect. However, these previous defect size estimation techniques cannot distinguish between defects that differ in size by an integer number of the ball angular spacing, and a third feature of the vibration response is therefore required to distinguish between such defects. It is hypothesized and validated through simulations that when the static load remains constant, the third distinguishing feature is the characteristic frequencies of the low frequency event.

Comparison of three commercially available softwares for measuring left ventricular perfusion and function by gated SPECT myocardial perfusion imaging

BackgroundThe three softwares, Quantitative Perfusion SPECT (QPS), Emory Cardiac Toolbox, and 4 Dimension-Myocardial SPECT (4DM) are widely used with myocardial perfusion imaging (MPI) to determine perfusion defect size (PDS) and left ventricular (LV) function. There are limited data on the degree of agreement between these methods in quantifying the LV perfusion pattern and function. Methods and ResultsIn 120 consecutive patients who had abnormal regadenoson SPECT MPI with a visually derived summed stress score ≥4, the correlation between the softwares for measurements of PDS, reversible, and fixed defects was poor to fair (Spearman’s ρ = 0.18-0.72). Overall, estimation of defect size was smaller by QPS and larger by 4DM. There was discordance among the softwares in 62% of the cases in defining PDS as small/moderate/large. The correlation between the softwares was better for measuring LVEF, volumes and mass (ρ = 0.84-0.97), and discrepant results for defining normal/mild-moderate/severe LV systolic dysfunction were prevalent in 28% of the patients. ConclusionThere are significant differences between the softwares in measuring PDS as well as LV function, and more importantly in defining small, moderate, or large ischemic burden. These results suggest the necessity of using the same software when assessing interval changes by serial imaging.

Thermal fly-height controlled glide for disk defect detection and defect size estimation for hard disk drives

Current hard disk drives use thermal fly-height control (TFC) during read/write operations. In this study, we use TFC technology during the disk glide process to determine sub-5 nm height defects. We also utilize TFC to measure the height of the defect during glide operation. Additionally, we magnetically mark the disk locations where defects are detected for further post-processing of the defects using optical surface analysis, atomic force microscopy (AFM), and scanning electron microscopy (SEM). The defect height estimation during the glide was confirmed to be accurate by AFM and SEM analysis. Finally, we will present the TFC glide sensitivity showing capability of detecting smaller defects than conventional non-TFC glide technologies.

Modeling 3D heat flow interaction with defects in composite materials for infrared thermography

Estimation of defect depth and size is an important problem in the domain of Nondestructive Testing in order to perform corrective repairs. The problem of defect depth estimation using Infrared Thermography has been previously studied using 1D heat conduction models. Unfortunately, 1D heat conduction based models are generally inadequate in predicting heat flow around finite-sized defects, especially in composite structures. In this study, a novel approach based on 3D heat conduction is proposed to model the heat flow interaction with the defect in quasi-isotropic composite materials. The modeling process involved coordinate transformations to reduce the anisotropic heat conduction problem to the isotropic domain, followed by separation of variables to solve the partial differential equation of the heat conduction. The validity of the model is established using Pulsed Thermography experiments performed on a composite panel containing rectangular flat-bottom holes of different sizes, present at different depths.