scholarly journals The Snooker Algorithm for Ultrasonic Imaging of Fatigue Cracks in order to use Parameter-Spaces to Aid Machine Learning

2021 ◽  
Author(s):  
Channa Nageswaran
Keyword(s):  
Machine Learning ◽  
Fatigue Cracks ◽  
Ultrasonic Inspection ◽  
Ultrasonic Imaging ◽  
Parameter Spaces ◽  
New Approach ◽  
Accurate Knowledge ◽  
Ultrasonic Techniques ◽  
Human Operators ◽  
Inspection Techniques

Fatigue cracks in a wide array of industrial components and structures pose a significant threat to their integrity. Detecting fatigue cracks using ultrasonic inspection techniques is a widespread activity for economic reasons but there are limitations to the techniques due to the morphology of fatigue cracks. In addition to detection there is a need to measure the size of the cracks which are often within the volume of the material. Ultrasonic techniques are well-suited to look inside the volume of the material but achieving sufficient sensitivity to the tip of the cracks in particular is practically difficult. Without an accurate knowledge of where the tip of the crack lies there can be significant uncertainty in sizing measurements. Machine Learning (ML) techniques are being developed to aid in the inspection and monitoring tasks but presenting the ultrasonic data in a suitable way for machine learning is very important. This paper presents a new approach to condition the ultrasonic data for machine learning settings so that they can be used effectively and confidently to detect and size fatigue cracks. The new approach, using images termed parameter-spaces, will also aid in conventional inspections as they are able to give information to human operators as to the existence or not of these very dangerous cracks.

scholarly journals OutlierNets: Highly Compact Deep Autoencoder Network Architectures for On-Device Acoustic Anomaly Detection

Sensors ◽  
2021 ◽  
Vol 21 (14) ◽  
pp. 4805
Author(s):  
Saad Abbasi ◽  
Mahmoud Famouri ◽  
Mohammad Javad Shafiee ◽  
Alexander Wong
Keyword(s):  
Machine Learning ◽  
Anomaly Detection ◽  
Detection Methods ◽  
Detection Accuracy ◽  
Network Architectures ◽  
Design Exploration ◽  
Convolutional Autoencoder ◽  
Acoustic Anomaly ◽  
Human Operators ◽  
Computational Resources

Human operators often diagnose industrial machinery via anomalous sounds. Given the new advances in the field of machine learning, automated acoustic anomaly detection can lead to reliable maintenance of machinery. However, deep learning-driven anomaly detection methods often require an extensive amount of computational resources prohibiting their deployment in factories. Here we explore a machine-driven design exploration strategy to create OutlierNets, a family of highly compact deep convolutional autoencoder network architectures featuring as few as 686 parameters, model sizes as small as 2.7 KB, and as low as 2.8 million FLOPs, with a detection accuracy matching or exceeding published architectures with as many as 4 million parameters. The architectures are deployed on an Intel Core i5 as well as a ARM Cortex A72 to assess performance on hardware that is likely to be used in industry. Experimental results on the model’s latency show that the OutlierNet architectures can achieve as much as 30x lower latency than published networks.

Extending FMC Based Ultrasonic Imaging Practices to Smaller Wall Thickness

2021 ◽  
Author(s):  
Niels Pörtzgen ◽  
Ola Bachke Solem
Keyword(s):  
Carbon Steel ◽  
Wall Thickness ◽  
Oil And Gas ◽  
Ultrasonic Inspection ◽  
Ultrasonic Imaging ◽  
Advanced Technology ◽  
Front Wall ◽  
Thin Wall ◽  
Destructive Testing ◽  
Girth Welds

Abstract During the construction of pipelines for the transportation of oil and gas, the inspection of girth welds is a critical step to ensure the integrity and thereby the safety and durability of the pipeline. In this paper we present an advanced technology ‘IWEX’ for the non-destructive testing of welds based on 2D and 3D ultrasonic imaging. This technology allows for safe, fast, and accurate inspection whereby the results are presented comprehensively. This will be illustrated with results from a recent project. The IWEX technology is based on an ultrasonic inspection concept, whereby ‘fingerprints’ of ultrasonic signals are recorded, also referred to as ‘full matrix capture’ (FMC) data. Then, an image area is defined, consisting out of pixels over an area large enough to cover the inspection volume. With the FMC data, image amplitudes are calculated for each pixel so that the shape of geometry (back wall, front wall, cap, and root) and possible indications are revealed. As opposed to traditional ultrasonic testing strategies, the detection and sizing of indications is therefore less dependent on its orientation. The project concerned the inspection of J and V welds from a 5.56″ diameter carbon steel pipe with an 8.4mm wall thickness. The wall thickness is relatively thin compared to common inspection scopes. Therefore, the inspection set-up was adapted, and procedural changes were proposed. Consequently, additional validation efforts were required to demonstrate compliance with the required inspection standard; DNVGL-ST-F101: 2017. As part of this, welds were scanned with seeded indications and the reported locations were marked for macro slicing under witnessing of an independent representative from DNVGL. The resulting images from the indications in the welds showed great detail with respect to the position, orientation and height of the indications. A quantitative comparison with the results from the macro slices was performed, including a statistical analysis of the height sizing and depth positioning accuracies. From the analysis, it could be observed that the expected improvements with respect to the resolution and sizing accuracy were indeed achieved. Thereby, the procedure has proven to be adequate for the inspection of carbon steel girth welds within the thin wall thickness range (~6mm to ~15mm). The IWEX technology is a member of the upcoming inspection strategy based on imaging of ultrasonic FMC data. This strategy can be considered as the next step in the evolution of inspection strategies after phased array inspection. The IWEX technology has been witnessed and qualified by independent 3rd parties like DNVGL, this makes the IWEX technology unique in its kind and it opens opportunities for further acceptance in the industry and other inspection applications.

Research of actual information on well casing using machine learning and neural networks

2021 ◽  
pp. 108-119
Author(s):  
D. V. Shalyapin ◽  
D. L. Bakirov ◽  
M. M. Fattahov ◽  
A. D. Shalyapina ◽  
V. G. Kuznetsov
Keyword(s):  
Machine Learning ◽  
Probabilistic Models ◽  
Western Siberia ◽  
Network Modeling ◽  
Horizontal Wells ◽  
Categorical Variables ◽  
Neural Network Modeling ◽  
New Approach ◽  
Attachment Process

In domestic and world practice, despite the measures applied and developed to improve the quality of well casing, there is a problem of leaky structures in almost 50 % of completed wells. The study of actual data using classical methods of statistical analysis (regression and variance analyses) doesn't allow us to model the process with sufficient accuracy that requires the development of a new approach to the study of the attachment process. It is proposed to use the methods of machine learning and neural network modeling to identify the most important parameters and their synergistic impact on the target variables that affect the quality of well casing. The formulas necessary for translating the numerical values of the results of acoustic and gamma-gamma cementometry into categorical variables to improve the quality of probabilistic models are determined. A database consisting of 93 parameters for 934 wells of fields located in Western Siberia has been formed. The analysis of fastening of production columns of horizontal wells of four stratigraphic arches is carried out, the most weighty variables and regularities of their influence on target indicators are established. Recommendations are formulated to improve the quality of well casing by correcting the effects of acoustic and gamma-gamma logging on the results.

scholarly journals Process Intensification 4.0: A New Approach for Attaining New, Sustainable and Circular Processes Enabled by Machine Learning

2021 ◽  
pp. 108671
Author(s):  
Enrique A. López-Guajardo ◽  
Fernando Delgado-Licona ◽  
Alejandro J. Álvarez ◽  
Krishna D.P. Nigam ◽  
Alejandro Montesinos-Castellanos ◽  
...  
Keyword(s):  
Machine Learning ◽  
Process Intensification ◽  
New Approach

scholarly journals Introducing the Viewpoint in the Resource Description using Machine Learning

2021 ◽  
Author(s):  
Ouahiba Djama
Keyword(s):  
Machine Learning ◽  
Experimental Study ◽  
Search Engines ◽  
Machine Learning Technique ◽  
New Approach ◽  
Learning Technique ◽  
Resource Description

Search engines allow providing the user with data and information according to their interests and specialty. Thus, it is necessary to exploit descriptions of the resources, which take into consideration viewpoints. Generally, the resource descriptions are available in RDF (e.g., DBPedia of Wikipedia content). However, these descriptions do not take into consideration viewpoints. In this paper, we propose a new approach, which allows converting a classic RDF resource description to a resource description that takes into consideration viewpoints. To detect viewpoints in the document, a machine learning technique will be exploited on an instanced ontology. This latter allows representing the viewpoint in a given domain. An experimental study shows that the conversion of the classic RDF resource description to a resource description that takes into consideration viewpoints, allows giving very relevant responses to the user’s requests.

scholarly journals Training a deep learning model for single-cell segmentation without manual annotation

2021 ◽  
Vol 11 (1) ◽  
Author(s):  
Nizam Ud Din ◽  
Ji Yu
Keyword(s):  
Machine Learning ◽  
Training Data ◽  
Machine Learning Techniques ◽  
Cell Segmentation ◽  
Learning Techniques ◽  
Signal Characteristics ◽  
Human Operators ◽  
Deep Learning Model ◽  
Microscopy Images ◽  
Segmentation Models

AbstractAdvances in the artificial neural network have made machine learning techniques increasingly more important in image analysis tasks. Recently, convolutional neural networks (CNN) have been applied to the problem of cell segmentation from microscopy images. However, previous methods used a supervised training paradigm in order to create an accurate segmentation model. This strategy requires a large amount of manually labeled cellular images, in which accurate segmentations at pixel level were produced by human operators. Generating training data is expensive and a major hindrance in the wider adoption of machine learning based methods for cell segmentation. Here we present an alternative strategy that trains CNNs without any human-labeled data. We show that our method is able to produce accurate segmentation models, and is applicable to both fluorescence and bright-field images, and requires little to no prior knowledge of the signal characteristics.

Sign in / Sign up

Export Citation Format

Share Document