Papers by Author: Janice M. Dulieu-Barton

Abstract: This paper describes the development of a stress / strain based in-situ damage inspection strategy focused around, but not exclusively, using thermoelastic stress analysis (TSA). The underlying philosophy is that defects and damage in a component or structure only constitute a cause for concern if these influence the stress field, i.e. the defect or damage acts as a stress raiser that reduces the service load limit. To assess this, it is necessary for the inspection method to map the distribution of stresses in the component, rather than the location and extent of an irregularity in the material. Imaging based techniques, such as TSA, digital image correlation (DIC) or digital speckle pattern interferometry (DSPI) provide non-contact maps of the surface stresses, deformations and/or strains. The full field data enables the engineer to evaluate if stress concentrations are present within the structure and, if data from a previous inspection is available, to assess if the distribution of stresses within the structure has changed from a previous 'undamaged' state. One of the key issues addressed in the current work has been the transition from a standard test setup, as typically used in laboratory work, to a more flexible (portable) setup relevant to industry requirements, e.g. site inspections. An approach that enables similar resolution (by comparison to current laboratory standard setups) stress and strain data to be captured using natural frequency excitation of a structure has been demonstrated on various full scale components.

Abstract: Polymer closed cell foam beam specimens manufactured from H100 Divinycell (Diab) are tested in four point bend at three loading speeds using a specially designed rig and an Instron VHS test machine. Synchronised high speed images are captured using white light and infra-red thermography (IRT) to obtain the mid-point full-field deflection and strains using digital image correlation (DIC) along with the temperature evolutions. There is a marked increase in the maximum load to failure with loading rate and the optical techniques provide an opportunity to analyse the strain and temperature evolution within the specimens.

Abstract: A study has been carried out to characterize the effect of variation of processing parameters on the phase contrast data between defective and defect-free areas obtained through the use of pulse phase thermography (PPT). Processing parameters used to implement the fast Fourier transform (FFT) have been varied. Phase contrast was maximized when the datum used as the start point for the FFT was taken as the frame just after the pulse. Optimum recording duration was found when the surface temperature had returned to its initial temperature. A truncation shorter than this resulted in a reduced phase contrast. Sampling interval and range is required to be balanced against the quantity of data produced and the computational expense. A sampling frequency of 0.06 Hz was suggested for the sample studied as this allowed peak phase contrast to be captured without unnecessarily increasing data size. Repeatability of tests was also investigated. It was found that PPT phase contrast results have been found to be repeatable with a maximum standard deviation of 6°.

Abstract: The work described in the paper investigates the stresses in the vicinity of a debond between the face sheet and core in a foam core / composite sandwich beam. Experiments were conducted in which the damaged side of the beam was loaded in compression in a four point bend test. Thermoelastic stress analysis (TSA) was used to obtain a measure of the stress field in the central part of the beam to evaluate its potential for damage detection and assessment. It is shown that TSA is capable of both identifying the damage and assessing its severity in terms of the remnant load carrying capacity of the beam. Fatigue tests were also performed, which showed that the cyclic load applied as a necessity for TSA did not cause the damage to grow, thereby demonstrating the potential of TSA as a non-destructive method for inspecting sandwich structures.

Abstract: In some metals it has been shown that the introduction of plastic deformation or strain modifies the thermoelastic constant, K. If it was possible to define the magnitude of the change in thermoelastic constant over a range of plastic strain, then the plastic strain that a material has experienced could be established based on a measured change in the thermoelastic constant. This variation of the thermoelastic constant and the ability to estimate the plastic strain that has been experienced, has potential to form the basis of a novel non-destructive, non-contact, full-field technique for residual stress assessment using thermoelastic stress analysis (TSA). Recent research has suggested that the change in thermoelastic constant is related to the material dislocation that occurs during strain hardening, and thus the change in K for a material that does not strain harden would be significantly less than for a material that does. In the work described in this paper, the change in thermoelastic constant for three materials (316L stainless steel, AA2024 and AA7085) with different strain hardening characteristics is investigated. As the change in thermoelastic response due to plastic strain is small, and metallic specimens require a paint coating for TSA, the effects of the paint coating and other test factors on the thermoelastic response have been considered.

Abstract: The feasibility of using pulse phase thermography (PPT) to identify defects in adhesively bonded joints is assessed. Artificial defects created in solid materials are successfully identified using the phase images produced by PPT. Contaminants typical of those found in the manufacturing environment have been used to recreate kissing defects in joints in polymer composite materials constructed using a single shot process and by using secondary bonding. It is shown that PPT has the potential to identify kissing defects.

Abstract: Digital image correlation (DIC) is an optical technique for full field deformation measurement. The spatial resolution and precision of the measurements are limited by the number of pixels within the image. The use of magnifying optics provides greater spatial resolution images, enabling smaller displacements to be observed with greater accuracy. Increasing the magnification of an image significantly changes the appearance of the non-periodic, stochastic speckle pattern which provides the grey scale contrast necessary for the image correlation method. In the paper a methodology is developed to evaluate the properties of different speckle pattern types under a range of resolutions up to 705 pixel / mm. Numerical deformation of the patterns is also undertaken to evaluate how the changes in the pattern properties affect the accuracy of the DIC measurements.

Abstract: Polymer composites are increasingly being used in high-end and military applications, mainly due to their excellent tailorability to specific loading scenarios and strength/stiffness to weight ratios. The overall purpose of the research project is to develop an enhanced understanding of the behaviour of fibre reinforced polymer composites when subjected to high velocity loading. This is particularly important in military applications, where composite structures are at a high risk of receiving high strain rate loading, such as those resulting from collisions or blasts. The work described here considers an approach that allows the collection of full-field temperature and strain data to investigate the complex viscoelastic behaviour of composite material at high strain rates. To develop such a data-rich approach digital image correlation (DIC) is used to collect the displacement data and infra-red thermography (IRT) is used to collect temperature data. The use of optical techniques at the sampling rates necessary to capture the behaviour of composites subjected to high loading rates is novel and requires using imaging systems at the far extent of their design specification. One of the major advantages of optical techniques is that they are non-contact; however this also forms one of the challenges to their application to high speed testing. The separate camera systems and the test machine/loading system must be synchronised to ensure that the correct strain/temperature measurement is correlated with the correct temporal value of the loading regime. The loading rate exacerbates the situation where even at high sampling rates the data is discrete and therefore it is difficult to match values. The work described in the paper concentrates on investigating the possibility of the high speed DIC and synchronisation. The limitations of bringing together the techniques are discussed in detail, and a discussion of the relative merits of each synchronisation approach is included, which takes into consideration ease of use, accuracy, repeatability etc.

Abstract: Polymer composites are increasingly being used in high-end and military applications, mainly due to their excellent tailorability to specific loading scenarios and strength/stiffness to weight ratios. The overall purpose of the ongoing research at the University of Southampton is to develop an enhanced understanding of the behaviour of fibre reinforced polymer composites when subjected to a range of loading scenarios. The measurements lecture reviews progress in using the thermoelastic stress analysis technique as a tool that enables a better understanding of the behaviour of polymer composite single skin and sandwich structures.

Abstract: Thermoelastic Stress Analysis (TSA) is a well-established full-field technique for experimental stress analysis that has proved to be extremely effective for studying stress fields in the vicinity of cracks. Recently, work has focused on the observation that the stress-sum contours (isopachics) obtained from TSA take the form of a cardioid. Genetic Algorithms (GAs) and Differential Evolution (DE) have proved successful for accurate parameter estimation of the cardioids, thus allowing the SIFs to be calculated. Originally, some curve-fits indicated that a pure cardioid form is inappropriate for the base model, especially for mixed-mode cracks. The deviation from the cardioid form has been shown to be due to higher-order terms within the stress function. The objective of the current paper is to use a modified version of the original methodology (that fitted parameters to a single isopachic) to find the higher-order parameters from the entire data field obtained from the TSA.