PSI - Issue 7

2

M. Hughes / Structural Integrity Procedia 00 (2017) 000–000

Martin Hughes / Procedia Structural Integrity 7 (2017) 33–35

34

grid. Over the past 70 years, the gas turbine has of course benefitted from significant developments in materials technology which is one of the factors affecting the significant improvements in performance and efficiency; the early nickel base alloy developments, columnar grained alloys, single crystal alloys, thermal barrier coatings, ceramics, ceramic matrix composites and additive manufacturing. Typical simple cycle engine efficiencies have improved significantly, benefitting from the increasing temperature capability of the high strength alloys. In parallel with the development of the gas turbine, there have of course been enormous developments in other technologies which have helped feed the improvements in the gas turbine. The first of these is computing power. Initially a combination of analogue and digital technologies contributed to understanding and solving the analytical problems relating to GT design. The digital world is now key to design, operation and maintenance. The related areas of numerical methods have experienced huge strides forward. Early finite element codes appearing in the 1960s, driven very much by the aerospace and nuclear sectors, started to make an initial impact, although initially requiring extensive computing facilities to generate useful data. By the 1980s, commercial finite element codes were being regularly used for specialised non-linear problems. Today, analyses with millions of degrees of freedom are regularly solved on large CPU/GPU clusters. This clearly gives some indication of the high resolution that can be achieved in today’s gas turbine design environment. The requirements for reliable and long life components under increasingly demanding operating conditions has been a significant factor in the drive towards the use of these high fidelity models. Conversly, the ability to create these types of models has also had an impact on how we go about addressing mechanical integrity and component life prediction. When used in combination with advanced material specimen testing techniques, our understanding of the mechanisms of failure and our ability to characterise the process has improved significantly. But what is the nature of the loading and material damage processes in gas turbine components? Why does it require all this intense analytical and computational attention? Clearly, for a gas turbine, the operating temperature is key. Combine this with high rotational speeds, and you have a collection of high energy parts operating close to their physical limits. But that’s not all. We want to start it up and shut it down, possibly quite frequently. At the temperatures of interest, time dependent material deformation behaviour can also make a significant contribution. Localised inelastic strain is well known to be a fundamental driver for fatigue crack nucleation under cyclic conditions. The later stages of component damage development may be related to how cracks develop over time. Fatigue crack growth under high stress levels may lead to local non-linear material behaviour. Low load levels may cause crack arrest. Time dependent material deformation processes will also influence the crack growth behaviour, as can other material degradation mechanisms such as oxidation and corrosion. Inherent material defects in a component may significantly reduce or even eliminate any initiation phase, therefore characterisation of the shape, orientation and distribution of such defects within a component becomes key to any integrity assessment. This is inherently linked to the manufacturing process. Consider first a major component like the rotor. There are different methods of gas turbine rotor construction. For example, the welded rotor configuration and the bolted assembly of individual discs/shafts. This will clearly affect the type of assessments required to ensure that the machine operates as specified. However, the common themes in the rotor include: the rotational dynamics behaviour, the centrifugal loading and also the stress fields arising from the high thermal loads. It is this loading, particularly the transient thermal behaviour during start-up, shut-down, and other load fluctuations which will have a strong impact on how fast you can start the machine, and how many times you can do it. Accurate modelling of hot gas flows and secondary air flows to characterise the thermal response is therefore a key input to the mechanical assessment process. Reliable mechanical assessments will rely very much on knowledge of the thermo-mechanical behaviour and significantly, how well the material is characterised and modelled. From simple linear elastic models and local strain approximations to highly complex material constitutive models for time independent and time dependent behaviour. The prediction of the local stress-strain behaviour needs to be carried out in a manner appropriate for the required accuracy. For lower strength rotor and disc materials where creep is active at typical operating conditions, then the cyclic damage mechanism will likely involve a creep and fatigue interaction with creep active at one end of the cycle and plasticity at the other. Consequently where prediction accuracy and component criticality are of importance then an advanced material model may be a key requirement. Similar mechanisms apply of course in many of the high thermally loaded components of the combustor and