| Summary: | Thesis (M.Eng.)--Memorial University of Newfoundland, 1996. Engineering and Applied Science Bibliography: leaves 91-101. Although many studies on the fatigue behaviour of tubular T-joints have been carried out thus far, the present study is done with the view to evaluating the total fatigue life of a joint in the presence of an asymmetric crack, using the principles of fracture mechanics. The studies made so far have only analyzed symmetric cracking, which does not generally occur during fatigue cracking, since it is assumed that the cracks at both the hot spot regions grow at the same rate. Also very few studies address the issue of crack initiation life in tubular T-joints. D.I. Nwosu (1993) used the stress-strain approach using the Manson-Coffin equation, to evaluate the initiation life. The elastic life component was used to determine the coefficients. In this study the author has used both lives (plastic strain life, as well as elastic strain life) to determine the coefficients used in the equation. The experimental results reported by Iida [1987] on tubular joints have been used to determine them. The range of obtained coefficients check with the prescribed limits given in earlier studies. The purpose of the study was to use and compare the results of the numerical model in verifying and correlating the experimental investigation of the tubular T-joints, which is being carried out in the Strength Laboratory of Faculty of Engineering, Memorial University. -- The line spring element was used to model the crack; the reason being that the other known method, viz., that of using singular three dimensional elements would render the problem impractical in the light of computer time and memory space available in the Faculty of Engineering. Using the stress intensity factors obtained from the line spring model, the through-thickness crack propagation life was obtained using Paris' law. While the crack initiation compared very well with the experimental value, the crack propagation life compared only favourably with the experimental value. The procedure required to improve the computed crack propagation life is given in the discussion of results.
|
|---|