Search

We aimed to develop a computer-aided characterization system that could support the diagnosis of dysplasia in Barrett's esophagus (BE) on magnification endoscopy. Videos were collected in high-definition magnification white-light and virtual chromoendoscopy with i-scan (Pentax Hoya, Japan) imaging in patients with dysplastic and nondysplastic BE (NDBE) from 4 centers. We trained a neural network with a Resnet101 architecture to classify frames as dysplastic or nondysplastic. The network was tested on 3 different scenarios: high-quality still images, all available video frames, and a selected sequence within each video. Fifty-seven patients, each with videos of magnification areas of BE (34 dysplasia, 23 NDBE), were included. Performance was evaluated by a leave-1-patient-out cross-validation method. In all, 60,174 (39,347 dysplasia, 20,827 NDBE) magnification video frames were used to train the network. The testing set included 49,726 i-scan-3/optical enhancement magnification frames. On 350 high-quality still images, the network achieved a sensitivity of 94%, specificity of 86%, and area under the receiver operator curve (AUROC) of 96%. On all 49,726 available video frames, the network achieved a sensitivity of 92%, specificity of 82%, and AUROC of 95%. On a selected sequence of frames per case (total of 11,471 frames), we used an exponentially weighted moving average of classifications on consecutive frames to characterize dysplasia. The network achieved a sensitivity of 92%, specificity of 84%, and AUROC of 96%. The mean assessment speed per frame was 0.0135 seconds (SD ± 0.006). Our network can characterize BE dysplasia with high accuracy and speed on high-quality magnification images and sequence of video frames, moving it toward real-time automated diagnosis. [Display omitted] [ABSTRACT FROM AUTHOR]

The marriage of welding and Non-Destructive Testing (NDT) processes at the point of manufacture has enabled the detection and correction of defects during the welding process. This has demonstrated clear financial and production benefits by reducing weld rework and ensuring schedule certainty, however this is yet to be demonstrated for use with narrow-groove welding practises. Narrow-groove welds are notoriously difficult to inspect using traditional Phased Array Ultrasonic Testing (PAUT) techniques due to large thicknesses and the vertical nature of Lack-of-Sidewall Fusion (LOSWF) defects. This is further complicated by the presence of partially-filled weld geometries during in-process inspection, which cause geometric reflections which can mask or falsely indicate the presence of a defect. A solution to this is proposed in this work, by adapting a dual-tandem phased array imaging system for the imaging of LOSWF defects in a partial weld geometry. This considers a two array system utilising a phased array probe on each weld side, coupled with an advanced dual-aperture Full Matrix Capture (FMC) acquisition technique. Advanced multi-mode image processing algorithms such as the Total Focusing Method (TFM) and Phase Coherence Imaging (PCI), with adaptive delay law calculation, have shown high sensitivity to LOSWF defects in a mock partial weld geometry. Additionally, an adaptive Probe Centre Spacing (PCS) technique is defined for in-process inspection based on amplitude and phase coherence sensitivity in partial weld geometries, with the effects of partial weld reflections analysed and discussed. These results have demonstrated the effectiveness of a dual-tandem phased array approach to imagine LOSWF defects during the in-process inspection of narrow-gap welds. [ABSTRACT FROM AUTHOR]

In recent years, there have been increasing economic and industrial drivers for the development of real-time non-destructive evaluation directly at the point of manufacture. Real-time inspection and monitoring of welding processes can help to reduce fabrication costs by detecting defects as they occur, enabling more efficient and cost-effective builds. This paper shows, for the first time, the use of phased array ultrasonics to monitor and analyse the molten weld pool during deposition of multi-pass gas tungsten arc welds. The received ultrasonic signals are shown to contain information related to key physical transitions occurring within the welding process, namely the melting and solidification of the weldment. Furthermore, the technique used here is shown to be effective for determining weld quality in real-time with significant signal changes occurring when defects such as Lack of Root Penetration are present. The accurate focusing and steering capabilities offered by phased arrays are used to successfully isolate the molten weld pool from the surrounding solidified weldment during deposition of multiple layers of a multi-pass weld. [ABSTRACT FROM AUTHOR]

This study tested whether a gradual reduction in sow contact during lactation, achieved through housing the sow in a ‘sow only’ area, would influence piglet stress responses to weaning. Gradual reduction in sow contact was achieved by separating the sow from her piglets (SP, n = 30) for 5, 7, and 9 h per day on days 10–15, 16–20 and 20 to weaning, respectively. Litters from 20 sows were followed as controls (CON), remaining in full contact with one another until weaning. Weaning occurred on day 28 ± 1.3 of lactation. Piglet body weight, injury scores and evidence of creep ingestion were measured throughout lactation and after weaning. Continuous video footage was collected for 6 h on the days following weaning for behavioural analyses. After weaning, SP piglets were lighter than CON piglets (6.8 ± 0.22 versus 7.6 ± 0.16 kg); however, by day 7 post-weaning piglet weights were similar (P > 0.05) for the CON (8.6 ± 0.22 kg) and SP (8.4 ± 0.15 kg) treatments, possibly reflecting a reduced growth check in SP piglets. There was a significant effect of treatment (CON versus SP) on the duration of aggressive (6.5 ± 1.1 versus 4.2 ± 0.8 s) and belly nosing (6.3 ± 2.0 versus 2.4 ± 1.3 s) events post-weaning. Injury scores were higher for CON piglets on almost all days examined (P < 0.05). Plasma circulating cortisol concentrations following weaning were increased in CON piglets (18.7 ± 13.3 nmol/L), and decreased in SP piglets (−12.3 ± 14.1 nmol/L; F 1,127 = 4.425, P < 0.05). These findings imply sow separation during lactation provides welfare benefits for piglets around the highly stressful weaning period. [ABSTRACT FROM AUTHOR]

The growing adoption of Carbon Fibre Reinforced Plastics (CFRPs) in the aerospace industry has resulted in a significant reliance on Non-Destructive Evaluation (NDE) to ensure the quality and integrity of these materials. The interpretation of large amounts of data acquired from automated robotic ultrasonic scanning by expert operators is often time consuming, tedious, and prone to human error creating a bottleneck in the manufacturing process. However, with ever growing trend of computing power and digitally stored NDE data, intelligent Machine Learning (ML) algorithms have been gaining more traction than before for NDE data analysis. In this study, the performance of ML object detection models, statistical methods for defect detection, and traditional amplitude thresholding approaches for defect detection in CFRPs were compared. A novel augmentation technique was used to enhance synthetically generated datasets used for ML model training. All approaches were tested on real data obtained from an experimental setup mimicking industrial conditions, with ML models showing improvement over amplitude thresholding and statistical thresholding techniques. The advantages and limitations of all methods are reported and discussed. [ABSTRACT FROM AUTHOR]

On modern manufacturing production lines, Non-Destructive Testing (NDT) is frequently a bottleneck which could greatly be alleviated by integrating the inspection of components as they are manufactured. By moving inspection to the point of manufacture, greater economic and productivity benefits are realised in terms of reduced rework and schedule slippage, however, new technical challenges emerge. For welded components, high temperatures and the resulting thermal gradients, present challenges when performing ultrasonic inspection at the point of manufacture. The thermal gradients introduce positional misalignment due to "beam bending" effects arising from refraction as the material properties change with temperature. This paper presents for the first time, through simulation and practical experiments, a novel thermal compensation strategy to mitigate for thermal effects when performing ultrasonic inspection of welded components at the point of manufacture. To understand the thermal gradients experienced during standard Tungsten Inert Gas (TIG) welding, 3-dimensional thermal simulations were developed and experimentally-validated with an average error of 1.80% (mean error: 4.1oC, max error: 19.2oC). The output from the thermal simulations in combination with material properties that vary over temperature, allowed for generalised time of flight maps to be created via the Multi-Stencils Fast Marching Method (MSFMM) and the ultrasonic data to be imaged by the Total Focusing Method (TFM). The thermal compensation strategy was initially proved on synthetically generated finite element Full Matrix Capture (FMC) datasets, and it was shown that reflector positional accuracy could be increased by ∼ 3 mm. Experimental results also showed marked improvements with reflector positional accuracy also being increased by ∼3 mm. Over both simulated and experimental datasets, the SNR was shown to be negligibly altered between uncompensated and compensated images. The results show how high-quality ultrasonic images can be generated in-process and help bring inspection closer to the point of manufacture. [ABSTRACT FROM AUTHOR]

Near-vertical flaws in thick-section narrow-gap welds are notoriously difficult to detect and size using traditional Phased Array Ultrasonic Testing (PAUT) techniques. Self-tandem shear inspection has been considered for detection but lacks appropriate full weld sensitivity. The dual-tandem phased array inspection method is proposed to enable multi-mode Total Focusing Method (TFM) pulse-echo imaging with longitudinal through-transmission inspection, by introducing a second, opposite-facing probe on the far weld side. Pulse-echo and through-transmission datasets are obtained in a single Full Matrix Capture (FMC) acquisition, for each weld side. Analysis of the optimum Probe Centre Spacing (PCS) based upon image sensitivity suggests that a spacing corresponding to a longitudinal beam crossover at 2/3 sample thickness provides greatest weld coverage. Testing on near-vertical EDM notches in a 120 mm thick mock narrow-gap carbon steel sample has shown high sensitivity TFM imaging, at notch depths of 27.5 mm, 60 mm and 92.5 mm. Pitch-catch through-weld transmission imaging has exhibited tip-diffraction images with a Signal-to-Noise Ratio (SNR) of up to 17.3 dB. Pulse-echo imaging using multi-mode TFM has shown tip-diffraction and reflection with up to 19.2 dB and 34.8 dB SNR respectively. The possibility of image mixing by considering image defect response type has provided full 2D notch reconstruction. These results demonstrate the strength of the dual-tandem method in enhancing the detection reliability and detection of near-vertical defects in thick-section narrow-gap welds. [ABSTRACT FROM AUTHOR]

[Display omitted] • Multi-layer adaptive ultrasound imaging using VSA for dry-coupled roller-probe inspection of AM components. • Inspection of as-built dry-coupled Titanium WAAM components with artificial calibration reflectors. • Increased data acquisition rate, increased speed of inspection and reduced data size while maintained sufficient SNR. Wire + Arc Additive Manufacture (WAAM) is an economical manufacturing technique to build components. To maintain the commercial and technical benefits, inspection is desired at each layer of the build. New conformable phased array roller-probes offer the potential to inspect the as-built sample in-process or on completion. A challenge with such inspections is the refraction of the ultrasonic waves at multiple interfaces. The Synthetic Aperture Focusing Technique (SAFT) and Total Focusing Method (TFM), combined with Full Matrix Capture (FMC) acquisition enabled imaging through the as-built WAAM surface. However, single-element firings, associated with FMC are a limiting factor due to the lower energy at transmission, while the high number of firings and subsequent larger data size negatively affect inspection speed. This work introduces the concept of Virtual Source Aperture (VSA) for ultrasonic roller-probe inspection, through modelling of VSA on custom-designed calibration WAAM block. The concept is then demonstrated on a mock as-built WAAM sample inspection with reference defects. The results demonstrate that a markedly lower number of transmissions are required to match the performance of the FMC counterpart while increasing the energy levels. Moreover, an almost 50% reduction in the data size enabled a doubling of the inspection speed. [ABSTRACT FROM AUTHOR]

This investigation introduces two new techniques to quantitatively address the challenging problem of understanding Hydrogen Induced Cracking (HIC) in welding processes. The first technique is a novel procedure to create a known and controlled HIC in a welded sample. The second is an in-process monitoring technique to measure the initial formation and subsequent growth of the HIC in a multi-pass weld whilst being compatible with the high temperatures associated with the welding process. The HIC was initiated using a localised quenching method of the weld and its character was verified using both macrograph and microscopic investigations. During HIC initiation and growth, the sample was monitored every 1–30 min for a total of 96 h using a custom non-destructive testing (NDT) system, mounted on a robot which ensured repeatable inspection positioning. Combining these techniques has therefore allowed for the first time, a detailed understanding of the evolution of HIC in a multi-pass welded sample. Our findings reveal that the HIC was initiated 43 min after the weld ended and that it then grew rapidly for about 15 min and continued growing at a slower rate for around 24 h. No significant growth was observed for the remaining 72 h of the experimental measurement. Image 1 • A novel procedure to create a known and controlled hydrogen crack is introduced. • An in-process high-temperature monitoring technique is also used. • Above combination allowed HIC measurement without cooling requirements of the weld. • HIC was initiated 43 min after the weld ended. • HIC grew rapidly for about 15 min and continued growing for 24 h. [ABSTRACT FROM AUTHOR]

Automated in-process Non-Destructive Testing (NDT) systems are rapidly gaining traction within the manufacturing industry as they reduce manufacturing time and costs. When considering calibration and verification of such systems, creating defects of known geometry and nature during the deposition of a weld can: (I) help examine the capability of the automated system to detect and characterise defects, (II) be used to form a database of signals associated with different defect types to train intelligent defect classification algorithms, and (III) act as a basis for in-process gain calibration during weld inspection at high temperatures, where the ultrasound beam can be skewed as a result of velocity gradients. In view of this, this paper investigates two unique methodologies for introducing: (a) lack of fusion weld defects by embedding tungsten in the weld and (b) creating artificial weld cracks by quenching to imitate the real cracking scenarios. According to the results of Phased Array Ultrasound Testing (PAUT) inspections, the methodologies used for embedding the artificial defects were successful. The validity of inspections was also verified by extracting micrographs from the defective sections of the welds, and model-based simulations were carried out to gain a better understanding of the wave propagation path and interaction with the generated defects. Unlabelled Image • Tungsten tubes could be successfully used for reference amplitude and gain calibration purposes. • CIVA as a simulation tool was found to provide reliable predictions for the weld inspections. • Local inter-pass water quenching of the weld can introduce cracks in the weld. • Water quenching-induced cracks and tungsten inclusions can be successfully detected by phased array ultrasound. [ABSTRACT FROM AUTHOR]

In multi-pass welding, there is increasing motivation to move towards in-process defect detection to enable real-time repair; thus avoiding deposition of more layers over a defective weld pass. All defect detection techniques require a consistent and repeatable approach to calibration to ensure that measured defect sizing is accurate. Conventional approaches to calibration employ fixed test blocks with known defect sizes, however, this methodology can lead to incorrect sizing when considering complex geometries, materials with challenging microstructure, and the significant thermal gradients present in materials during the inter-pass inspection period. To circumvent these challenges, the authors present a novel approach to calibration and introduce the concept of in-process calibration applied to ultrasonic Non-Destructive Testing (NDT). The new concept is centred around the manufacturing of a second duplication sample, containing intentionally-embedded tungsten inclusions, with identical process parameters as the main sample. Both samples are then inspected using a high-temperature robotic NDT process to allow direct comparative measurements to be established between the real part and the calibration sample. It is demonstrated that in-process weld defect detection using the in-process calibration technique can more reliably identify defects in samples which would otherwise pass the acceptance test using a traditional calibration. Unlabelled Image • It is proposed that in-process inspection must include in-process calibration. • This is centred around the manufacture of a second duplicate sample. • The calibration sample contains intentionally-imbedded tungsten defects. • In-process inspection and calibration together could reliably identify defects. [ABSTRACT FROM AUTHOR]

Automated weld deposition coupled with the real-time robotic Non-Destructive Evaluation (NDE) is used in this paper. For performance verification of the in-process inspection system, an intentionally embedded defect, a tungsten rod, is introduced into the multi-pass weld. A partially-filled groove (staircase) sample is also manufactured and ultrasonically tested to calibrate the real-time inspection implemented on all seven layers of the weld which are deposited progressively. The tungsten rod is successfully detected in the real-time NDE of the deposited position. The same robotic inspection system was then used to continuously monitor an intentionally-manufactured crack for 20 h. The crack was initiated 22 min after the weld ended and it grew quickly within the next 1.5 h. The crack growth stopped approximately after 2 h and no considerable change in the reflection signal was detected for the next 18 h of monitoring. Unlabelled Image • Automated weld deposition coupled with the real-time robotic NDT is discussed. • An intentionally embedded defect, a tungsten rod, is introduced for verification. • A partially-filled groove sample is also manufactured and ultrasonically tested. • The system is then used for continuous monitoring of a crack for 20 h. • The crack was initiated 22 min after the weld and grew for 1.5 h. [ABSTRACT FROM AUTHOR]