Fatigue crack initiation and small crack propagation in additive manufactured Ti-6Al-4V

  • Muhammad Shamir

    Student thesis: Doctoral ThesisDoctor of Philosophy


    This research aims to quantify the effect of as-deposited surface on crack initiation and early-stage propagation in a wire + arc additive manufactured Ti-6Al-4V (WAAM Ti64). The main objectives were: 1) to conduct fatigue testing and microstructural examination of specimens in the as-deposited conditions; 2) to develop a model to predict a fatigue life; 3) to study the effect of crystallographic orientations of grains surrounding a defect
    on crack initiation and small crack propagation. To achieve these objectives, fatigue testing and numerical modelling were designed and conducted. Key findings are summarised below.

    Firstly, specimens with as-deposited surface were tested under bending fatigue load. The surface waviness feature was characterised using a metrology tool and represented as
    "groove" or "notch" as in the classic fatigue method. Therefore, fatigue life prediction was based on the stress concentration factor of the centre notch geometrical profile using the notch-stress approach. A range of values representing the shallowest and deepest notches in the WAAM Ti64 build was investigated. The model has worked well for test samples under higher applied stresses while it over-estimated the fatigue life at lower applied stresses.

    An alternative approach was subsequently proposed based on the fracture mechanics method assuming the surface notch as an initial crack length based on the equivalent initial flaw size (EIFS) concept. Given the notch size being sub-mm range, small crack
    growth testing was performed using a recommended geometry by ASTM E647 standard. It was found that small cracks grow faster than long cracks under the same applied stress intensity factor. Hence small crack data must be used to estimate the fatigue life of AM parts containing defects.

    Furthermore, the small crack test data was used to verify the Hartman-Schijve variant of NASGRO equation that only requires testing a long crack specimen. Based on this work, a recommendation was made to use long crack data to predict small crack growth behaviour. Moreover, a unified crack growth rate curve was computed using the Hartman-Schijve equation, and as a demonstrator, it was further used for the durability prediction of another geometry in as-deposited condition. Predicted life was found to be slightly conservative compared with the notch stress approach.

    Finally, microstructure characterisation of internal gas pores was conducted showing one of the reasons for fatigue test scatter in WAAM Ti64 being the variation in the α lath crystallographic orientation and its width in the vicinity of the crack initiating pore.
    Pyramidal slip systems with high Schmid factor and smaller α laths resulted in longer fatigue life.

    In summary, a unified crack growth curve was proposed and verified by small crack experimental data. The model capability was demonstrated by durability analysis of as-deposited WAAM Ti64 and found to be in good agreement with the test result. The
    fracture mechanics based approach is more suitable for durability analysis of AM parts containing defects as the crack length is included in the model. Nevertheless, the classical notch stress method has also worked well for predicting the crack initiation life from sub-mm notches. The microstructural analysis has shown that the durability of AM parts with defects could be improved by reducing the α lath width and/or increasing the number of
    active pyramidal slip systems.
    Date of Award2022
    Original languageEnglish
    Awarding Institution
    • Coventry University
    SponsorsBAE Systems
    SupervisorAbdul Syed (Supervisor) & Xiang Zhang (Supervisor)

    Cite this