Creep life prediction of new and service exposed 0.5Cr-0.5Mo-0.25V steel pipework

Abstract

There is a growing need to increase the reliability and safety of conventional coal-fired power plants as the UK prepares to reduce its coal-fired capacity in fulfilment of the emissions requirements set in recent years. Over the next decade, the coal-fired power plants that continue to operate will require efficient design and maintenance, such that service shutdowns and inspections are minimised. It is, therefore, fundamental that the remaining life of hazard prone areas be known and informed. A large percentage of the critical high temperature components have exceeded their “design life” of 30 to 40 years, while others are aging considerably and approaching the end of their design life. To realise the continued use of these components, methods and procedures are required to understand the current plant conditions so that the remaining life of the components can be projected. Existing creep models are largely focused on uniaxial test data. In this work, a comprehensive uniaxial and multiaxial creep test programme has been carried out to validate and incorporate existing and novel models for the life assessment of components. The material examined is 0.5Cr-0.5Mo-0.25V steel, removed from a straight pipe section after 250,000 hours of operation at 16 MPa internal steam pressure and a temperature of 843 K. Circumferentially notched bar specimens were used to induce multiaxial stress states using three different notch acuities. The skeletal stress approach was used to characterise the overall behaviour of the specimen. It was shown that creep failure is controlled by the equivalent stress. The majority of the constant load uniaxial and notched bar creep tests were carried out under accelerated conditions in air, with a number of tests carried out in an inert environment to investigate the effects of oxidation. Acceleration was achieved by increasing the stress or temperature or both. It was shown that oxidation is severe in small diameter specimens, especially at temperatures above 843 K. The effects of oxidation were seen to be detrimental on the rupture lives of laboratory specimens. A discrepancy in the axial deformation of the notched bar tests carried out in an inert environment was seen with the finite element data, which was based on a material model derived from tests carried out on air. A compendium of existing as-cast data for multiple batches of 0.5Cr-0.5Mo- 0.25V pipe and tube steel was used to fit the Wilshire model for extrapolation to service conditions. In this work, a methodology is proposed in which the Wilshire model can be used in conjunction with the life fraction rule to predict the remain- ing life of the component. The suitability of the method was assessed by comparing predictions against the results from the test programme. It was shown that the mean diameter hoop stress can be used to characterise the damage incurred in service, thus allowing creep life predictions to be made independent of the specimen location across the pipe thickness. The Wilshire model was implemented into a finite element analysis as a time-based damage law and under the assumption that creep failure is controlled by the equivalent stress. The results were compared to the ductility exhaustion approach based on the Cocks and Ashby method. Conservative rupture life predictions were made for both cases. The potential for the Wilshire model to be incorporated into the formation of a time dependent failure assessment diagram (TDFAD) was also shown. The damage in the pipe prior to testing has been characterised as being high. An admixture of grain boundary separation and cavitation was seen, with cavitation being more severe nearer to the pipe bore. Micrographs of the un-failed notches of the notched bar tests showed that damage was concentrated just ahead of the notch root, which was in agreement with what was seen in the damage contour plots of the finite element results obtained from the Cocks and Ashby damage formulations.