Extending fatigue life of industrial low-pressure FV566 turbine blades: efficacy of a lifetime extension strategy to extend service life

Abstract

The fir-tree-root-fillets of turbine blades are prone to fatigue cracking on the notch surfaces and pose a potential risk of fatigue failure before expected service life due to increasing exposure to start-stop fatigue loading conditions. Turbine systems are additionally subjected to 10 % over-speed testing annually. Turbine blades are being replaced at high cost due to the serious safety risk in the case of catastrophic failure. The efficacy of a proposed lifetime-extension strategy involving grinding out existing cracks, identifiable via non-destructive testing (thus changing the notch geometry), followed by T0 shot peening and the effect of additional overload cycles was investigated. FV566 martensitic stainless-steel material extracted from ex-service turbine blades was tested and the characteristic baseline material microstructure, mechanical properties and fatigue behaviour examined. Single edge notch bend and U-notch specimens (with industry representative geometries) were subjected to 3-point bending fatigue loading in the low cycle fatigue regime. A finite element model of the test specimens with elastic-plastic material modelling was developed to calculate required loads and stress and strain distributions in the notch field with various U-notch geometries. The effects of industry representative ‘as-received’ and polished notch surface conditions on the crack initiation behaviour was investigated. The effect of relatively small industry representative overloads up to 150 % of the maximum baseload every 150 baseload cycles were compared with constant amplitude loading and the effect on initiation behaviour and crack growth rates examined. Various U-notch geometries expected after grinding out existing cracks were fatigue tested at different strain ranges to establish the influence of notch geometry on fatigue behaviour. T0 industrial representative shot peening was applied to U-notched specimens and the residual stress profiles obtained via XRD. The effect shot peening on fatigue behaviour was investigated for various loading conditions (including overload) and notch geometries. A lifetime prediction method was developed which was capable of accounting for the application of the proposed lifetime extension strategy and was compared with experimental testing. Early crack initiation was observed in the as-received notch surface condition due to the presence of corrosion pits. A polished surface condition increased the number of cycles to crack initiation. The presence of an overload 110 % or less of the maximum baseload every 150 baseload cycles did not affect fatigue life. An overload of 150 % retarded both long and short crack growth rates, attributed to compressive residual stress ahead of the crack tip, resulting in an overall increase in fatigue life. Additionally, a compressive residual stress induced in the notch field increased the number of cycles to crack initiation, improving fatigue life overall. Shot peening damaged the notch surface causing pre-existing cracks and early crack initiation. However, the compressive residual stress field from shot peening significantly retarded short crack growth and increased overall fatigue life. After 1.2 % total strain range the lifetime extension benefit from shot peening was diminished. Notch geometry was not found to have a notable difference on fatigue life when tested at identical strain ranges. The lifetime extension strategy was found to increase the overall life of Unotch samples assuming the strain range remained constant throughout the test. The lifetime prediction model was able to predict the number of cycles to failure for a sample subjected to the lifetime extension strategy and offer a lifetime prediction that accounts for a likely increase in strain range after changing the notch geometry. The service life of turbine blades may be improved by adopting the lifetime extension strategy. Finite Element modelling of actual turbine blade geometry with industry relevant loading conditions is required to establish the increase in strain range expected from grinding out existing cracks in-situ. Further experimentation is then required to apply the method developed in this thesis to actual turbine blade geometry prior to incorporating the lifetime extension strategy as a maintenance procedure.

More information

Identifiers

ORCID for Benjamin Cunningham: orcid.org/0000-0002-2604-4242

ORCID for Philippa Reed: orcid.org/0000-0002-2258-0347

Catalogue record

Export record

Altmetrics