Peridynamic analysis of fatigue crack growth in fillet welded joints

Student thesis: Master's Thesis

Abstract

Fatigue assessment is one of the significant factors to be considered for a design life of structure and estimation of structural reliability during operation. Especially, for welded structures, various welding effects, such as stress concentration, residual stresses, weld geometry and weld quality, make the structure more vulnerable to fatigue failures. This requires more effective approaches for estimating fatigue performances of welded structures.Existing classical methods to predict the crack propagation under cyclic loadings have some difficulties in treating complicated patterns of crack growth. A peridynamic theory, however, has powerful advantage on discontinuities. A peridynamic fatigue model, which is a bond damage model of remaining life, is used to demonstrate two phases of fatigue failure, crack nucleation and crack growth. Two types of numerical tests are conducted to validate the peridynamic fatigue model. One is tensile test for the phase of crack nucleation and the other is compact tension test for the phase of crack growth. All results from numerical tests are compared with experimental test data to validate the peridynamic fatigue model.After validation of peridynamic fatigue model, numerical tests with peridynamic fatigue model are performed to investigate a weld effect of the length of unwelded zone on the fatigue performance of load-carrying fillet welded joint. Numerical results of fatigue performance and path of fatigue crack growth are compared with existing experimental data.In this thesis, the peridynamic fatigue model is validated by two different fatigue tests which are uniaxial tension-compression tests for the crack nucleation and ASTM E647 standard compact tests for the crack growth. After validation, the fatigue performance of the fillet welded joint is estimated with respect to the length of unwelded zone by simulating the fatigue crack growth under cyclic load conditions with the peridynamic fatigue model.
Date of Award1 Sep 2018
LanguageEnglish
Awarding Institution
  • University Of Strathclyde
SupervisorSelda Oterkus (Supervisor) & Erkan Oterkus (Supervisor)

Cite this