Accounting for Batch to Batch Variation when Predicting the Safe Life of Materials Operating at High Temperatures: An Ap
- PDF / 678,269 Bytes
- 14 Pages / 593.972 x 792 pts Page_size
- 74 Downloads / 206 Views
INTRODUCTION
THE structural integrity of many AGR components needs to be assessed at long lifetimes. In order to perform such assessments, it is necessary to have confidence in the relevant material property predictions which may involve significant extrapolation beyond the range of the available data on a given material. This can prove particularly problematic when equations are based upon polynomial expressions (such as the methods proposed by LarsonMiller,[1] Orr-Dorn-Shepherd,[2] and Manson-Haferd[3]) since at low stresses, the equations of these models are often subject to turn back and hence it is not possible to extrapolate to long durations. This paper focuses on 1Cr1Mo0.25V steel for turbine rotors and shafts, for which this is a particular problem. Without techniques for the accurate extrapolation of short-term property measurements (obtained from 1 or less years of testing), reliance must instead be placed on very protracted and expensive test programs lasting 12 to 15 years to determine safe operating life. Further, the determination of safe life must also be based on the analysis of multiple heats or batches of the same material. For example, the British Steel Makers’ Creep Committee elevated-temperature data[4] on 1Cr1Mo0.25V steel contain eight different batches that vary slightly in their chemical composition, geometry (bar, tube and plate), heat treatment, and steel making process. Similarly, the National Institute for Materials Science (NIMS) in its MARK EVANS, Senior Lecturer, is with the College of Engineering, Swansea University, Singleton Park, Swansea SA2 8PP, UK. Contact e-mail: [email protected] Manuscript submitted July 10, 2012. Article published online December 11, 2012 METALLURGICAL AND MATERIALS TRANSACTIONS A
creep data sheet No. 9b[5] has long-term creep data out to over 100,000 hours for nine different batches of this steel. It is therefore of little surprise to note that a reduction in this 12- to 15-year materials’ development cycle has been defined as the No. 1 priority in the 2007 UK Energy Materials—Strategic Research Agenda.[6] The theta methodology, first proposed by Evans and Wilshire,[7] offers a framework of analysis that avoids the problem of turn back that occurs when using arbitrary polynomials and so offers a solution to the need to use such long-term test programs. But, while the equations of the theta methodology are easily manipulated to yield the required creep property predictions, it has the disadvantage that the only quantitative information currently available on the values for the unknown constants of this model is derived from an investigation by Evans et al.[8] and Evans[9] on just one batch of this material. Consequently, there is some concern as to how representative the estimated parameters of this model, and thus the resulting creep property predictions, are of a wider range of 1Cr1Mo0.25V heats. Indeed, the batch analyzed by Evans et al.[8] is believed to represent the lower bound strength of this material. The aim of this paper is to apply the theta meth
Data Loading...
