Fatigue is a pervasive failure mode that affects many industrial sectors including the high value aerospace, nuclear and automotive sectors. It remains a source of in-service failure on the one hand and inefficient over-engineered conservative design on the other and so generates considerable risks and cost (capital and operating) to industry. In the lower stress regimes where high or very high cycles to failure occur there are several factors that complicate fundamental understanding of fatigue failure and how to manage it effectively in engineering practice: (i) testing methodologies generally only test to ~1 million cycles while safety critical components may see much longer service periods of a hundred to a thousand times longer in the aerospace and nuclear sector forcing extrapolation into unknown and untested regimes (ii) there is considerably more scatter in fatigue lives in (very) high cycle fatigue compared to low cycle fatigue which is linked to a greater influence of microstructure (iii) the crack initiation process takes up a much larger fraction of the total fatigue life but as no crack is present it is difficult to know where in the material microstructure to make observations that will capture local processes that will eventually lead to crack nucleation (iv) residual stresses from processing and machining make a more significant contribution to the total stress state when the external loading is smaller. This research programme will deliver a step change in high cycle fatigue testing by combining ultrasonic technology with small scale miniature test-piece designs. The tests will be conducted at 20 kHz at which a million cycles takes just less than a minute and a billion cycles takes only 1 day. The sample dimension will be in two regimes. Firstly, a micro-regime with Focused Ion Beam (FIB) cut sample widths only a fraction to a few micrometres across allowing testing of individual selected features of a microstructure (grain, grain boundary, inclusion...). Secondly, a meso-regime with samples a few tens to a few hundreds of micrometres wide cut using laser micro-machining and allowing small patches of microstructure to be tested. The meso-samples are sufficiently small that frequent intermittent microscopic characterisation methods can be used to the local evolution of local deformation, stress, and dislocation content in regions where crack initiation is guaranteed to occur eventually. Greater understanding of processes leading to crack initiation and how local variation in microstructure control fatigue crack initiation lifetimes are the key scientific and technological outcomes sought. This step change advance will be exploited in the first instance to characterize effects of process conditions on the fatigue crack initiation response of (i) linear friction welds & (ii) peened surfaces in Ti-6Al-4V.

The concept of grain size playing an important role in the engineering application of polycrystalline metals is well established. During casting and subsequent wrought processing, tried and tested methods are used to refine grain size in order to enhance ductility and increase tensile, yield and fatigue strengths. The advent of electron microscopy based experimental techniques such as electron back scatter diffraction (EBSD) and focussed ion beam (FIB) plus nano-indentation have provided novel, intriguing insights into the deeper aspects of both structural evolution and structure / property relationships. This has included preliminary identification of the critical role of effective structural unit size (rather than grain size) in determining mechanical behaviour. However, understanding of the the relationship between processing and effective structural unit size remains in its infancy for most systems. Consequently, significant progress can now be made in understanding the evolution of structures including recrystallisation processes and variant selection during phase transformation. This offers the potential of refining the structure of a wide range of engineering materials for which phase transformation plays an important role during processing such as steel, titanium, zirconium etc. The fatigue process is very complex but can be simplified conceptually into initiation and crack growth. For high cycle fatigue (HCF) regimes where the number of applied stress cycles can easily exceed 10,000,000 material evaluation relies on specimen or component testing. The majority of the HCF life is spent initiating a defect that then grows rapidly to failure. For materials subject to such HCF regimes, the design principle is to stay below an empirically defined endurance stress so that initiation is prevented. For low cycle fatigue (LCF) the situation is different in that initiation life and growth life can both be used to predict a safe component life. Typically, initiation is again determined empirically by mechanical testing. The current inability to predict fatigue initiation from basic principles stems from the fact that crack initiation is dominated by interactions from grain to grain which are inherently difficult to quantify and to model. Thus, for significant end user applications, the engineer has minimal knowledge defining what aspects of a material, or its processing, influence its performance other than by mechanical testing, which is very time consuming and expensive.Considerable scientific exploration of fatigue has until recently largely failed to assist the material producer and end user in other important ways. In the specific case of the titanium-based alloys, the definition of grain boundaries and subsequent measurement of grain size are notoriously difficult through optical inspection alone. The existence of large colonies of similarly orientated crystallographic units can encourage extensive planar slip structures to develop. In turn, through a process of stress redistribution between relatively weak and strong units , this can have a potentially disastrous effect on component performance. Key issues which determine mechanical properties of interest to the end user include:a) How boundaries behave and what constitutes a boundary for a given load regime.b) Factors in processing and heat treatment that dictate effective structural unit size.c) Modelling capability to provide quantitative predictions of mechanical behaviour including HCF initiation and short crack growth rates.All of these issues form the basis of the current proposal for research..

The concept of grain size playing an important role in the engineering application of polycrystalline metals is well established. During casting and subsequent wrought processing, tried and tested methods are used to refine grain size in order to enhance ductility and increase tensile, yield and fatigue strengths. The advent of electron microscopy based experimental techniques such as electron back scatter diffraction (EBSD) and focussed ion beam (FIB) plus nano-indentation have provided novel, intriguing insights into the deeper aspects of both structural evolution and structure / property relationships. This has included preliminary identification of the critical role of effective structural unit size (rather than grain size) in determining mechanical behaviour. However, understanding of the the relationship between processing and effective structural unit size remains in its infancy for most systems. Consequently, significant progress can now be made in understanding the evolution of structures including recrystallisation processes and variant selection during phase transformation. This offers the potential of refining the structure of a wide range of engineering materials for which phase transformation plays an important role during processing such as steel, titanium, zirconium etc. The fatigue process is very complex but can be simplified conceptually into initiation and crack growth. For high cycle fatigue (HCF) regimes where the number of applied stress cycles can easily exceed 10,000,000 material evaluation relies on specimen or component testing. The majority of the HCF life is spent initiating a defect that then grows rapidly to failure. For materials subject to such HCF regimes, the design principle is to stay below an empirically defined endurance stress so that initiation is prevented. For low cycle fatigue (LCF) the situation is different in that initiation life and growth life can both be used to predict a safe component life. Typically, initiation is again determined empirically by mechanical testing. The current inability to predict fatigue initiation from basic principles stems from the fact that crack initiation is dominated by interactions from grain to grain which are inherently difficult to quantify and to model. Thus, for significant end user applications, the engineer has minimal knowledge defining what aspects of a material, or its processing, influence its performance other than by mechanical testing, which is very time consuming and expensive.Considerable scientific exploration of fatigue has until recently largely failed to assist the material producer and end user in other important ways. In the specific case of the titanium-based alloys, the definition of grain boundaries and subsequent measurement of grain size are notoriously difficult through optical inspection alone. The existence of large colonies of similarly orientated crystallographic units can encourage extensive planar slip structures to develop. In turn, through a process of stress redistribution between relatively weak and strong units , this can have a potentially disastrous effect on component performance. Key issues which determine mechanical properties of interest to the end user include:a) How boundaries behave and what constitutes a boundary for a given load regime.b) Factors in processing and heat treatment that dictate effective structural unit size.c) Modelling capability to provide quantitative predictions of mechanical behaviour including HCF initiation and short crack growth rates.All of these issues form the basis of the current proposal for research.