Super resolution array imaging of embedded defects within safety-critical components
File(s)
Author(s)
Elliott, Joshua Benjamin
Type
Thesis
Abstract
There is a constant drive within the nuclear power industry to improve upon the characterization capabilities of ultrasonic Non-Destructive Evaluation (NDE) inspection techniques in order to improve safety and reduce costs, with particular emphasis placed on the ability to characterize small defects. The usage of ultrasonic phased array technologies have led to significant advancements in NDT performance relative to conventional monolithic transducers and they have also led to the development of several advanced imaging algorithms. A group of these called Super Resolution (SR) algorithms have been shown to demonstrate a capability to resolve scatterers separated by less than the diffraction limit when deployed in representative NDE inspections.
In this thesis, the Factorisation Method (FM) and the Time Reversal Multiple Signal Classification (TR-MUSIC) algorithms were investigated in the imaging of embedded defects. The performance of these SR techniques in accurately characterising smooth embedded planar defects of varying size and orientations was investigated via two-dimensional (2D) Finite Element (FE) simulations and the results were experimentally validated. These studies were extended to consider more realistic three-dimensional (3D) smooth embedded planar defects in experimental trials and rough embedded planar defects, the latter being explored using 2D FE Monte Carlo simulations.
The SR algorithms were also benchmarked against the conventional array Total Focusing Method, which is recognised to be a high performing and robust imaging technique. The SR algorithms were also applied to the imaging of 2D and 3D volumetric defects in order to determine if direct, image-based sizing could be achieved with these ultrasonic methods. The final and most challenging inspection case considered within this thesis was the inspection of embedded defects within austenitic stainless steel welds. These materials exhibit spatially-varying anisotropic coarse grained microstructures which can lead to significant ultrasonic signal attenuation and beam bending effects that make their NDE inspection difficult.
In this thesis, the Factorisation Method (FM) and the Time Reversal Multiple Signal Classification (TR-MUSIC) algorithms were investigated in the imaging of embedded defects. The performance of these SR techniques in accurately characterising smooth embedded planar defects of varying size and orientations was investigated via two-dimensional (2D) Finite Element (FE) simulations and the results were experimentally validated. These studies were extended to consider more realistic three-dimensional (3D) smooth embedded planar defects in experimental trials and rough embedded planar defects, the latter being explored using 2D FE Monte Carlo simulations.
The SR algorithms were also benchmarked against the conventional array Total Focusing Method, which is recognised to be a high performing and robust imaging technique. The SR algorithms were also applied to the imaging of 2D and 3D volumetric defects in order to determine if direct, image-based sizing could be achieved with these ultrasonic methods. The final and most challenging inspection case considered within this thesis was the inspection of embedded defects within austenitic stainless steel welds. These materials exhibit spatially-varying anisotropic coarse grained microstructures which can lead to significant ultrasonic signal attenuation and beam bending effects that make their NDE inspection difficult.
Version
Open Access
Date Issued
2020-08
Date Awarded
2021-01
Copyright Statement
Creative Commons Attribution NonCommercial NoDerivatives Licence
Advisor
Lowe, Michael
Huthwaite, Peter
Sponsor
Engineering and Physical Sciences Research Council (EPSRC)
Great Britain. Royal Commission for the Exhibition of 1851
Great Britain. Ministry of Defence
Grant Number
EP/I017704/1
Publisher Department
Mechanical Engineering
Publisher Institution
Imperial College London
Qualification Level
Doctoral
Qualification Name
Engineering Doctorate (EngD)