| Popis: |
This thesis is concerned with the influence of two surface hardening procedures, mechanical shot peening (MSP) and deep cold rolling (DCR), on the microstructures of single crystal nickel-based superalloys. These alloys have excellent high temperature properties, and consist of a randomly substituted face centred cubic matrix phase known as γ, together with a large volume fraction of an ordered precipitate phase, most often the primitive cubic γ' phase. Superalloys cast as single crystals are the premier materials for the manufacture of turbine blades in aerospace jet engines, where the material experiences high temperatures and fatigue stresses for prolonged periods of time in service. Turbine blade superalloys have traditionally been optimised for creep resistance. However, a continual industry drive to improve turbine jet engine efficiency has led to greater low cycle fatigue damage being accumulated by the single crystal turbine blades, as metal stresses and temperatures rise. A case in point is the poor fatigue performance of intermediate pressure turbine blades in the Rolls Royce Trent 1000 engine, which was a major industrial issue at the time of writing. The application of MSP and DCR was identified as a prospective way to improve turbine blade fatigue resistance. This called for a systematic investigation of the effects of these procedures on single crystal superalloy materials, which had previously been largely unexplored. Surface hardening procedures such as MSP and DCR function by plastically deforming the surface of a workpiece, generating a surface compressive residual stress and a layer of cold worked material. Both the residual stress and cold work are thought to impede the fatigue propagation of surface cracks, but previous studies have established that at high cycling stresses and high temperatures, the cold work confers the main benefit to fatigue life. In MSP, the surface of the material is bombarded by a stream of small hard shot, while in DCR, the surface is deformed by a roller under hydrostatic pressure. Each procedure is defined by several key process parameters: intensity, coverage and shot size for MSP, and pressure and roller ball diameter for DCR, and setting the correct combinations of these parameters is important for optimising a particular procedure for a given application. The ultimate goal of the Work described in this thesis was to enable the informed application of surface hardening to single crystal turbine blades, so as to achieve the greatest improvement in fatigue behaviour. The specific brief of this Work was to examine the effects of MSP and DCR on the microstructures of single crystal superalloys, and to characterise how different procedure parameters affected the resultant work hardened layer of material, particularly the distribution of microscopic cold work. Changes in the as hardened microstructure following high temperature heat treatments were also to be assessed, since thermal stability is a key requirement for the turbine blade. Local misorientation (LM) analysis of electron backscattered diffraction (EBSD) data was the principal tool for characterising cold work structures produced by MSP and DCR. The development of the LM analysis methodology for single crystal superalloys and the creation of a Python script to implement this methodology were major outcomes of this Work. Cold work depths and magnitudes, the latter being the amount of cold work per depth, were defined on the basis of LM data and were used to quantify the average extents and densities of the cold worked layers. Test samples made from the superalloy CMSX-4, treated by MSP or DCR with varied combinations of parameters, were examined using LM analysis and backscattered electron imaging. It was shown that in MSP, greater intensity increased the depth of cold work, while peening coverage chiefly influenced the amount of cold work per depth. Greater shot size tended to produce deeper cold work but with less magnitude. Likewise, in DCR, greater roller diameter and greater rolling pressure both led to greater cold work depth, with the rolling pressure having a further significant influence over the amount of cold work per depth. Both the depth and the magnitude of cold work were also seen to depend somewhat on the secondary crystallographic orientation of the crystal matrix in a given sample. Overall, DCR resulted in much deeper cold worked layers than MSP, with much lower magnitudes. The lateral cold work distributions produced by MSP and DCR were also observed to differ substantially. In MSP, a thick continuous surface layer of cold work, ~35 µm in depth, was seen, while in DCR, cold work was concentrated into narrow surface streaks. Sets of slip bands, denser but more shallow in MSP than DCR, penetrated in depth from the hardened surface, often past the depth of the bulk cold work. To assess the response of the peened and rolled sample microstructures to highly elevated temperatures, samples were heat treated at 900 and 1100 °C under vacuum with no externally applied stress. Significant amounts of recrystallisation (RX) and topologically close packed phase (TCP) precipitation were observed and quantified in samples treated at 900 °C, both phenomena being promoted by the slip bands induced in the material. After 500 h at 900 °C, samples with sufficiently low cold work magnitudes displayed RX depths fairly in line with a control sample, though almost all MSP samples displayed RX grains of ~30 µm or more after 1000 h. At 1100 °C RX occurred more far more rapidly in all samples, whereas TCP formation was hindered by the recovery of slip bands and the lower thermodynamic stability of TCPs at this temperature. Samples treated by DCR generally displayed smaller RX depths and lower rates of TCP formation than those treated by MSP. Directional coarsening, or rafting, of the γ/γ' microstructure was noted at 1100 °C. The extent of rafting was independent of the hardening conditions, and increased with depth before reaching a transition to the bulk microstructure. It was surmised that rafting was driven by residual stresses generated within the material by the surface hardening, and that cold work impeded rafting. The latter conclusion was supported by the invariable occurrence of peak rafting after the measured cold work depth, highlighting the difference between the related, but distinct phenomena of cold work and residual stress. Finally, work was carried out on turbine blades manufactured from the superalloy RR3010 and treated by MSP, to validate the results obtained in this Work. The cold work structures of the blades were seen to correspond well with those in test samples peened under similar conditions. Preliminary investigation was also carried out on peened blades with Pt Cr diffusion coatings, which are proposed to be used in conjunction with MSP to increase the corrosion and oxidation resistance of intermediate pressure turbine blades. The coatings did not themselves change the underlying cold work structure significantly. However, they were found to occupy the area of greatest cold work magnitude in the cold work profile typical of MSP, so that the amount of cold work transmitted to the alloy beneath was generally low. The present Work has resulted in development of greater understanding of the effects of MSP and DCR on single crystal nickel based superalloys, including the variation of cold work depth and magnitude with procedure parameter, the differences between the cold work structures produced by the two procedures, and the microstructural consequences of exposing the work hardened material to high temperatures. This will aid industry in rationalising and refining hardening procedures when applied to turbine blades, and will lead to blades having greater fatigue resistance and longevity, ultimately improving jet engine efficiency. This thesis consists of seven chapters and an appendix. A brief summary of the contents is given at the beginning of each chapter. Chapter 1 is the introduction, where the industrial context, relevant background information and a review of the known effects of surface hardening procedures on superalloys are given. Chapter 2 contains the experimental details of the materials and methods employed throughout the Work. Chapters 3 and 4 deal with the as hardened states of the MSP and DCR samples, respectively. The effects of static heat treatments on these samples are described and discussed in Chapter 5. Chapter 6 is concerned with the studies performed on both coated and uncoated turbine blade specimens. The conclusions drawn from the Work are summarised in Chapter 7, together with suggestions for possible future work in this field. Appendix A contains the Python script used to conduct the LM analysis and a discussion of its functionality. |
