Use this URL to cite or link to this record in EThOS:
Title: The development of a strain-based defect assessment technique for composite aerospace structures
Author: Christian, William J. R.
ISNI:       0000 0004 6496 2663
Awarding Body: University of Liverpool
Current Institution: University of Liverpool
Date of Award: 2017
Availability of Full Text:
Access from EThOS:
Access from Institution:
This thesis details the work conducted over three years on the development of strain-based defect assessment techniques for carbon-fibre reinforced composites. This material, whilst exhibiting a high specific strength, is sensitive to defects and thus there is an industrial need for assessment techniques that are capable of characterising defects and obtaining predictions of residual strength or life. The most commonly applied techniques are currently ultrasonic and thermographic non-destructive evaluation. A strain-based defect assessment could lead to more accurate predictions of residual strength, resulting in a reduction of the costs associated with operating composite aerospace structures. The aim of this project is to increase the quality and confidence in residual strength information gained from the non-destructive evaluation of composite defects using strain-based assessments, in addition to currently applied ultrasonic practices for composite structures. A literature review on composite defects and existing techniques for assessing defects was conducted. Knowledge gaps were then identified that if filled, could improve residual strength predictions. Initially, a statistical framework was developed that used Bayesian regression to predict the residual strength of impacted composites, based on ultrasonic non-destructive measurements, that is robust to data outliers. As part of this framework a performance metric for quantifying the accuracy of residual strength predictions was introduced, allowing currently applied assessment techniques to be compared with the novel strain-based assessment. Then, a novel technique for performing strain-based defect assessments was developed that utilised image decomposition and the statistical framework to make residual strength predictions. Digital image correlation was used to measure strain fields which were then dimensionally reduced to feature vectors using image decomposition. The difference between feature vectors representing virgin and defective laminates were quantified, resulting in a strain-based defect severity measure. Bayesian regression was used to fit an empirical model capable of predicting the residual strength of an impacted laminate based on the strain-based defect severity. The accuracy of the strain-based predictions were compared to the accuracy of ultrasound-based predictions and found to outperform the currently applied ultrasonic technique. Strain-based assessment of in-plane fibre-waviness was also explored, as minimal research had been conducted studying waviness defects with full-field techniques. This required the development of a procedure for creating controlled levels of local waviness in laminates. The same strain-based assessment used for assessing impact damage was applied to the fibre-waviness specimens, but for this defect the accuracy of predictions were found to be comparable to the ultrasound-based predictions. However, residual strain measurements were found to be effective for predicting the strength of laminates, indicating that knowledge of the residual strains around a waviness defect may be important when predicting a laminates residual strength.
Supervisor: Patterson, Eann ; Diaz De la O, Alex Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral