Data-driven analysis of ultrasonic inspection data of pressure tubes

Research output: Contribution to journalArticle

Abstract

Pressure tubes are critical components of CANDU reactors and other pressurized heavy water–type reactors because they contain the nuclear fuel and the coolant. Manufacturing flaws as well as defects developed during in-service operation can lead to coolant leakage and can potentially damage the reactor. The current inspection process of these flaws is based on manually analyzing ultrasonic data received from multiple probes during planned, statutory outages. Recent advances in ultrasonic inspection tools enable the provision of high-resolution data of significantly large volumes. This highlights the need for an efficient autonomous signal analysis process. Typically, automation of ultrasonic inspection data analysis is approached by knowledge-based or supervised data-driven methods. This work proposes an unsupervised data-driven framework that requires no explicit rules or individually labeled signals. The framework follows a two-stage clustering procedure that utilizes the Density-Based Spatial Clustering of Applications with Noise density-based clustering algorithm and aims to provide decision support for the assessment of potential defects in a robust and consistent way. Nevertheless, verified defect dimensions are essential in order to assess the results and train the framework for unseen defects. Initial results of the implementation are presented and discussed, with the method showing promise as a means of assessing ultrasonic inspection data.
LanguageEnglish
Pages153-160
Number of pages8
JournalNuclear Technology
Volume202
Early online date1 Mar 2018
DOIs
Publication statusPublished - 30 Jun 2018

Fingerprint

inspection
ultrasonics
Inspection
Ultrasonics
tubes
Defects
defects
reactors
coolants
Coolants
Heavy water
heavy water
signal analysis
nuclear fuels
Signal analysis
Nuclear fuels
automation
Outages
Clustering algorithms
leakage

Keywords

  • CANDU
  • pressure tubes
  • ultrasonic inspection
  • machine learning
  • clustering

Cite this

@article{348a13bcd9714d87af32b67c22d39803,
title = "Data-driven analysis of ultrasonic inspection data of pressure tubes",
abstract = "Pressure tubes are critical components of CANDU reactors and other pressurized heavy water–type reactors because they contain the nuclear fuel and the coolant. Manufacturing flaws as well as defects developed during in-service operation can lead to coolant leakage and can potentially damage the reactor. The current inspection process of these flaws is based on manually analyzing ultrasonic data received from multiple probes during planned, statutory outages. Recent advances in ultrasonic inspection tools enable the provision of high-resolution data of significantly large volumes. This highlights the need for an efficient autonomous signal analysis process. Typically, automation of ultrasonic inspection data analysis is approached by knowledge-based or supervised data-driven methods. This work proposes an unsupervised data-driven framework that requires no explicit rules or individually labeled signals. The framework follows a two-stage clustering procedure that utilizes the Density-Based Spatial Clustering of Applications with Noise density-based clustering algorithm and aims to provide decision support for the assessment of potential defects in a robust and consistent way. Nevertheless, verified defect dimensions are essential in order to assess the results and train the framework for unseen defects. Initial results of the implementation are presented and discussed, with the method showing promise as a means of assessing ultrasonic inspection data.",
keywords = "CANDU, pressure tubes, ultrasonic inspection, machine learning, clustering",
author = "Panagiotis Zacharis and Graeme West and Gordon Dobie and Timothy Lardner and Anthony Gachagan",
note = "This is an Accepted Manuscript of an article published by Taylor & Francis in Nuclear Technology on 01 March 2018, available online: http://www.tandfonline.com/10.1080/00295450.2017.1421803.",
year = "2018",
month = "6",
day = "30",
doi = "10.1080/00295450.2017.1421803",
language = "English",
volume = "202",
pages = "153--160",
journal = "Nuclear Technology",
issn = "0029-5450",

}

Data-driven analysis of ultrasonic inspection data of pressure tubes. / Zacharis, Panagiotis; West, Graeme; Dobie, Gordon; Lardner, Timothy; Gachagan, Anthony.

In: Nuclear Technology, Vol. 202, 30.06.2018, p. 153-160.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Data-driven analysis of ultrasonic inspection data of pressure tubes

AU - Zacharis, Panagiotis

AU - West, Graeme

AU - Dobie, Gordon

AU - Lardner, Timothy

AU - Gachagan, Anthony

N1 - This is an Accepted Manuscript of an article published by Taylor & Francis in Nuclear Technology on 01 March 2018, available online: http://www.tandfonline.com/10.1080/00295450.2017.1421803.

PY - 2018/6/30

Y1 - 2018/6/30

N2 - Pressure tubes are critical components of CANDU reactors and other pressurized heavy water–type reactors because they contain the nuclear fuel and the coolant. Manufacturing flaws as well as defects developed during in-service operation can lead to coolant leakage and can potentially damage the reactor. The current inspection process of these flaws is based on manually analyzing ultrasonic data received from multiple probes during planned, statutory outages. Recent advances in ultrasonic inspection tools enable the provision of high-resolution data of significantly large volumes. This highlights the need for an efficient autonomous signal analysis process. Typically, automation of ultrasonic inspection data analysis is approached by knowledge-based or supervised data-driven methods. This work proposes an unsupervised data-driven framework that requires no explicit rules or individually labeled signals. The framework follows a two-stage clustering procedure that utilizes the Density-Based Spatial Clustering of Applications with Noise density-based clustering algorithm and aims to provide decision support for the assessment of potential defects in a robust and consistent way. Nevertheless, verified defect dimensions are essential in order to assess the results and train the framework for unseen defects. Initial results of the implementation are presented and discussed, with the method showing promise as a means of assessing ultrasonic inspection data.

AB - Pressure tubes are critical components of CANDU reactors and other pressurized heavy water–type reactors because they contain the nuclear fuel and the coolant. Manufacturing flaws as well as defects developed during in-service operation can lead to coolant leakage and can potentially damage the reactor. The current inspection process of these flaws is based on manually analyzing ultrasonic data received from multiple probes during planned, statutory outages. Recent advances in ultrasonic inspection tools enable the provision of high-resolution data of significantly large volumes. This highlights the need for an efficient autonomous signal analysis process. Typically, automation of ultrasonic inspection data analysis is approached by knowledge-based or supervised data-driven methods. This work proposes an unsupervised data-driven framework that requires no explicit rules or individually labeled signals. The framework follows a two-stage clustering procedure that utilizes the Density-Based Spatial Clustering of Applications with Noise density-based clustering algorithm and aims to provide decision support for the assessment of potential defects in a robust and consistent way. Nevertheless, verified defect dimensions are essential in order to assess the results and train the framework for unseen defects. Initial results of the implementation are presented and discussed, with the method showing promise as a means of assessing ultrasonic inspection data.

KW - CANDU

KW - pressure tubes

KW - ultrasonic inspection

KW - machine learning

KW - clustering

UR - https://www.tandfonline.com/toc/unct20/current

U2 - 10.1080/00295450.2017.1421803

DO - 10.1080/00295450.2017.1421803

M3 - Article

VL - 202

SP - 153

EP - 160

JO - Nuclear Technology

T2 - Nuclear Technology

JF - Nuclear Technology

SN - 0029-5450

ER -