Further analysis of the Monkman–Grant relationship for 2·25Cr–1Mo steel using creep data from the National Research Institute for Metals

Abstract

The National Research Institute for Metals in Japan regularly publishes long term creep data which is not as readily available in the UK. Parametric extrapolation techniques have proved unsuccessful on such data sets and so interest has recently turned to the Monkman–Grant relationship. A number of shortcomings in past estimates of this relationship make them unsuitable for extrapolation. These include failure to make use of censored data and to identify the distribution for rupture lives, and neglect of the known dependency of the Monkman–Grant relationship on test conditions. This paper addresses these issues to derive an empirical relationship capable of extrapolation. Incorporating unfailed specimens into the analysis substantially altered the predictions obtained from the Monkman–Grant relationship. Uncertainty over the shape of the failure time distribution also increased the confidence limits associated with these predictions. The intercept of the Monkman–Grant relationship was shown to vary with absolute temperature and this was modelled to allow the relationship to be extrapolated to operating conditions for boiler tubes. A weak linear dependency of variability in rupture life on the minimum creep rate was also identified.