Material damage evaluation during long time service in turbine rotors

Structural integrity of gas turbine rotors components naturally degrades over time when subjected to the harsh environment of gas turbine operating conditions; at the end of their design life they should not be necessarily replaced, as they may have a potential to safety operate for additional time. In particular, the possibility of extending the service duration is particularly important for rotor components. In order to come up with the best possible judgment of the rotor condition, these components are subjected to specific investigation procedures that involve removal, disassembly and a full inspection with traditional and high tech nondestructive testing (NDT) tools and numerical methods. This approach can be used to ensure the safety to keep in service or to recommend the necessary restoration activities. The term “life extension” has often been misunderstood, hence it is important to remark that the purpose of life extension activities is not to continue the operation of a component beyond its useful life, but merely to ensure full utilization up to its useful life. Structural integrity and residual life assessment require a multidisciplinary and integrated approach to perform a pragmatic engineering analysis. As a matter of fact, inspection via NDT methods, suitable material properties of aged materials, detailed turbine engine operation data and maintenance history, a deep structural analysis and life prediction models for the operating failure mechanisms are essential to ensure reliable and safe remaining assessment of Gas Turbine rotors and discs. Since rotors contain significant potential energy, an unexpected failure of rotating components represents serious safety risks to personnel and it can cause collateral damage to other parts of the machine until to a total breakdown of the system with a significant economic loss, or worse, the loss of human lives. Depending on the type of service, a wide spectrum of damage mechanisms and various forms of material aging and metallurgical degradation may sometimes interact, accelerating the rate of damage accumulation faster than design requirements. Therefore, in-depth understanding of expected damage evolution together with indication about how the degree of damage can be measured for each failure mode, are essential in order to correctly evaluate the actual status of components and the risks in going ahead with further operations. In this context, the research carried out in the present work contributes to the knowledge of the damage accumulation for rotors after long-term service exposure. The thesis is focused on the study of numerical-experimental techniques to face the complex challenge of material damage estimation for particular types of damage phenomena, specific for both heavy duty and aero-derivative gas turbines. The final objective is to identify useful parameters that can indicate a time-dependent material degradation in order to give additional information on material state. More in detail, three types of practical approaches are devised to improve the rotor condition assessment and the life management processes. Initially the thesis reports an overview of the aspects which introduce uncertainties into life assessments and provides the basis for this research topic. Then the thesis deals with the modellization of a particular embrittlement phenomenon which occurs in many kinds of low alloy rotor steels, as consequence of the exposure temperature. Such a phenomenon can be monitored by measuring the increase in the ductile to brittle fracture appearance transition temperature (FATT) and it leads to a severe fracture toughness degradation with an increased risk of brittle fracture. An artificial neural network-based algorithm able to capture the complexity of the embrittlement susceptibility has been developed to enable the prediction of a typical indicator of embrittlement: the above mentioned FATT. Results of the model are in agreement with microstructural theories and observations reported in the literature. The neural model provides the evolution of the FATT value with time for a specific alloy composition. It can be used to study the influence on the material embrittlement of every element of the alloy, and also to find out the optimal steel composition. The model has been applied to NiCrMoV low alloy steels; yet with the appropriate database, it can be easily extended to other steels commonly used for turbine rotors, as CrMoV and 9-12%Cr. Metallurgical assessments, performed with various techniques, as for example with replicas technique, commonly integrate every residual life assessment program. However, many degradation mechanisms are difficult to be observed until items life has expired. For this reason, inspections may not be able to observe and to assess the evolution of damage. Advanced microstructural investigation, carried out with transmission electron microscopy or alternatively with other advanced techniques as atom force microscopy and atom probe field ion microscopy, can provide a complete investigation of precipitates and structures. However these techniques require a complex sample preparation procedure, being a critical phase conditioning the reliability of the observations results. Moreover, only a small volume of material can be examined for each analysis. Due to some limitations introduced by economic considerations, a limited number of examination is commonly performed; the limited areas of investigation, together with the restricted statistical basis, introduce a problem of accuracy and repeatability of measurements. In contrast to these methods, a suitable microstructural strategy of investigation which combines optical microscopy, scanning electron microscopy and X-ray powder diffraction has been proposed. The quantitative determination of the volume fraction of the nanosized carbides and carbonitrides precipitates is challenging due to their fine sizes, wide particle size distributions, and low volume fractions. The proposed procedure, developed with reference to high resistant 9-12%Cr stainless steels, has been carried out to identify microstructural measurable parameters which may be useful to assess relationships between the variation of mechanical properties and the microstructural evolution. The typical NDT inspections, performed during a rotor overhaul, are useful to identify defects and they are necessary for the estimation of minimum or average life in terms of crack initiation and propagation. From a material point of view, NDT do not give any ageing degradation information. With the aim at overcoming some limitations of conventional NDT an application of the micro-indentation technique has been introduced. Instrumented indentation testing (IIT), also known as depth-sensing indentation, is a relatively new type of mechanical testing: it provides a continuous record of the variation of indentation load as a function of the depth of penetration into the indented specimen. This technique is a promising and attractive tool nowadays available in materials engineering and science, particularly for the fact that indentation tests are non-destructive and also inexpensive from an economical point of view. In the technical literature, several numerical methods were devised to evaluate monotonic properties of the material starting from the load-indentation curves. Differently from these approaches, in this thesis an investigation to link IIT results with damage has been carried out. In particular, two parameters coming from the indented curve are proposed, and they can indicate the damage level resulting from temperature field and stress acting on high pressure rotor components during service. In a few words, this thesis aims to suggest some smart techniques in the field of the complex challenge of material damage quantification, which is a necessary and fundamental phase for the evaluation of the life extension of turbine rotors.