Thermoelastic Stress Analysis (TSA) is a well established technique for studying stress fields around a crack tip. Recent work utilised the observation that individual isopachics (contours of constant stress) around a crack tip appeared to take the form of a cardioid. Various computational approaches were successfully used to estimate stress intensity factors using the first order Westergaard equations. Whilst this worked well for mode I cracks, there was less success when applying the equation to mixed mode situations; it appeared as though a pure cardioid form was not an appropriate description of the isopachics for mixed mode cracks. The work carried out in this thesis concentrates on (i) developing an algorithm that is able to accurately estimate crack tip parameters from thermoelastic data using the Williams expansion and (ii) evaluating the effectiveness of the approach using simulated fields. In the first part of the study it is shown that the higher order terms in the Williams expansion are responsible for creating the rotation of the isopachics noted in previous work. One of the higher order terms in the expansion is the T-stress which is thought to be responsible for controlling the stability of crack growth. A differential evolution is used to estimate the parameters of the Williams expansion and fit the parameters to thermoelastic data. This would allow the calculation of the stress intensity factors and T-stress for the crack, as well as estimate the crack tip position. Both mode I and mixed mode plates are used in an effort to accurately calculate the crack tip parameters. It is shown from both experimental results and simulations that using DE to fit the Williams expansion directly to thermoelastic data is technique that requires further work and investigations. Although the crack tip can be accurately located, the stress intensity factors and T-stress estimations are consistently inaccurate. It is also shown that the results of the parameters are dependent upon the size of the data array used. The T-stress is never accurately estimated and the results for the stress intensity factors are inconsistent. However, the shortcomings of the DE algorithm are highlighted, and possibilities to remedy the problems are suggested.
| Identifer | oai:union.ndltd.org:bl.uk/oai:ethos.bl.uk:560794 |
| Date | January 2012 |
| Creators | Hebb, Richard Ian |
| Publisher | University of Southampton |
| Source Sets | Ethos UK |
| Detected Language | English |
| Type | Electronic Thesis or Dissertation |
Page generated in 0.0018 seconds