Estimation of stress intensity factors in ship structural connections

Rights statement
Awarding institution
  • University of Strathclyde
Date of award
  • 2021
Thesis identifier
  • T16052
Person Identifier (Local)
  • 201173741
Qualification Level
Qualification Name
Department, School or Faculty
  • Stiffened plated structures such as ships and box girder bridges, result in connection details that contain sharp internal corners. Many failures in ship structure have been found to be associated with fatigue crack propagation at the side shell connections between longitudinal and transverse structure. According to elastic stress analysis, these sharp corners are geometric singularities that have an infinite stress in the corner. A further complication of stiffened structures is that a crack may grow through intersections, e.g. of plates and stiffeners and changes of plate thickness before it causes a catastrophic structural failure. In this thesis, a new approach is developed to simplify the analysis of these issues. The singular stress contribution is, as usual, characterised by Y, the non-dimensional the stress intensity factor but within this method simplified analysis is used to calculate the Y values. The method combines a ratio of non-singular linearized ligament stresses to estimate the effect of large changes in crack length and changes in plate thickness with an empirical methods to estimate the local effect as the crack grows through a change of thickness. The method does not require an analysis of the actual singularity, so saving analysis time and, importantly, giving the engineer some feeling for the result and the possibility of a “back of the envelope” calculation for the SIF or Y. This work is based on running finite element analyses, to determine the Stress Intensity Factor and Y and using the results to test the empirical or analytical methods.The derived methods are useful both for assessment of existing structures and for design application. Comparing the results from the application of this new methodology with the FE method and existing fatigue analysis guidance, the new method is very much quicker and easier to apply. It is though less accurate than FE analysis and so is most appropriate for, (1) preliminary assessment, (2) reliability assessment where many structural and defect variations are to be considered and (3) for checking whether a more detailed analysis is producing sensible results. For design calculations often a stress concentration factor or SCF is needed that can be used with an S-N curve. The actual predicted peak stress and hence SCF will, for finite element analysis, depend on the element size and will normally increase as the element size decreases. The existing guidance on determining an appropriate stress value for fatigue analysis of a sharp corner is commonly in terms of linearly extrapolating finite element calculated surface stresses from a number of plate thicknesses t away from the singularity to the corner. A simpler approach, developed for planar plates with sharp corners, assesses the stress on the basis of the dimensions of the corner. This thesis includes checks on the applicability, to more complicated 3-d geometry, of these previous recommendations for the assessment of corner singularities.
Advisor / supervisor
  • Barltrop, N. D. P. (Nigel D. P.)
Resource Type