Thesis

Non-destructive evaluation of residual stress using phased array ultrasonics

Downloadable Content

Download PDF
Creator
Rights statement
Awarding institution
  • University of Strathclyde
Date of award
  • 2026
Thesis identifier
  • T18200
Person Identifier (Local)
  • 202258458
Qualification Level
Qualification Name
Department, School or Faculty
Abstract
  • Residual stress (RS) affects the structural integrity and fatigue performance of welded and additively manufactured metallic components and can contribute to distortion, cracking, and premature failure. Established RS measurement techniques, including incremental hole drilling, neutron diffraction, and the contour method, provide valuable stress information but can be limited by localised sampling, lack of in-situ capability, or destructive material removal. Ultrasonic methods offer a non-destructive alternative, particularly through longitudinal critically refracted (LCR) waves, whose velocity is affected by material stress through the acoustoelastic effect. However, ultrasonic RS measurement relies on nanosecond-scale time-of-flight (ToF) changes, making it sensitive to external sources of ToF variation. This thesis develops a phased-array, acoustoelasticity-based framework progressing from controlled stress measurement in bolts, to phased array ultrasonic residual stress measurement (PAURS), and finally to microstructure-sensitive interpretation (PAURS+). Full Matrix Capture (FMC) enables multiple ultrasonic paths to be assessed using ToF, amplitude, attenuation-related response, signal quality, and path-wise variance, providing a basis for determining measurement quality. Bolted connections are used as the controlled-stress benchmark, with ultrasonic measurements compared against mechanical loading data and finite-element-derived average axial stress. Phased-array ultrasonics resolved preload-sensitive ToF changes and path-wise variation associated with asymmetric loading or disrupted acoustic paths. PAURS was subsequently developed using a pitch-catch phased-array configuration to generate and receive LCR waves in available welded S275 steel samples, with further S275 steel acquired for complementary investigations. Simulation and experiments were used to verify LCR generation, assess penetration depth, calibrate the acoustoelastic response, and quantify couplant gel-thickness effects. Uncontrolled coupling variation produced practically significant apparent stress errors, while adjacent-receiver subtraction and multi-path processing reduced the compensated mean gel-induced ToF error to 0.27 ns, corresponding to an equivalent apparent stress error of 1.09 MPa using the approximation of 1 ns ≈ 4 MPa present in steel. This value represents the mean across three steel plates. This is treated as a contributing source of error, and not an overall uncertainty. PAURS was applied to a robotically welded S275 steel sample, selected as a representative weldable structural material for development of the residual-stress measurement methodology, and assessed against incremental hole-drilling measurements performed externally on the same specimen, and neutron-diffraction measurements obtained from a nominally identical welded specimen. Accounting for differences in measurement principle, gauge volume, and sampling depth, PAURS reproduced the principal residual stress trends observed by the independent techniques where suitable wedge–specimen contact was achieved. Agreement was reduced near the weld centre, where weld geometry and poor wedge–specimen coupling affected the LCR response; measurements in these regions were therefore interpreted as semi-quantitative or qualitative. FMC reduced path-wise variability relative to matched linear paths and provided an internal measure of measurement consistency; this improvement is interpreted as increased precision rather than evidence of reduced absolute bias. The final stage extends PAURS to microstructure-sensitive characterisation through PAURS+. Electron Backscatter Diffraction (EBSD), k-Wave simulation, frequency-dependent analysis, and robotic angular LCR measurements were used to investigate Al7075 produced by additive friction stir deposition. Amplitude and attenuation-related angular responses were used as the principal microstructure-sensitive observables. Angular attenuation signatures distinguished between equiaxed and columnar grain structures, indicating near-surface microstructural anisotropy and demonstrating that microstructure should be considered when interpreting ultrasonic RS measurements. Overall, the thesis establishes an FMC-based phased-array methodology combining calibrated ToF stress calculation, multi-path processing, couplant-error control, signal-quality assessment, and comparison with independent RS measurements. The framework identifies when ultrasonic outputs support numerical stress calculation, trend-based interpretation, or rejection from quantitative assessment, and extends the same phased-array platform towards microstructure-sensitive indication through PAURS+.
Advisor / supervisor
  • Javadi, Yashar
  • MacLeod, Charles Norman
  • Pierce, Gareth
Resource Type
DOI

Relations

Items

Thumbnail Title Date Uploaded Visibility Actions
file details PDF of thesis T18200 2026-10-08 Public Download