Reliability Analysis On Rupture Life Of Heat-Resistant Steel Based On Z Parameter

On the basis of probabilistic analysis on creep rupture data of HK-40 austenite steel and5Cr0.5Mo pearlitic steel, it is investigated the reliability analysis method for heat resistantproperty assessment and life prediction at elevated temperature. From the viewpoint of datascatter, parameter Z was introduced which represented the degree of deviation from themaster curve for each data plot. The concept considered parameter Z as a random variable andrevealed the randomness in rupture property of materials. This parametric method can also beapplied in other life prediction procedures based on Time-Temperature parameter.The integral distribution character of parameter Z was studied. The distributionparameters were estimated by maximum likelihood method. Three considerations were givento check the distribution, which were general fit, consistency of the fit with creep rupturephysics, and safety for design in the tail region. By the K-S methodology confirming therationality of the hypothesis distribution, normal distribution was suggested to be aappropriate statistical model for parameter Z. This method overcomes the disadvantage ofignoring data scatter when using the single master curve for rupture property assessment, andprovides a new clue for the further research and application on the life prediction.Forthermore, the confidence interval range for integral distribution of parameter Z wasestimated. The master curve was proposed to represent the average rupture property undereach temperature and stress level. The distribution regularities of parameter Z in different testconditions were studied, whose result revealed the evolution of the rupture propertydispersion extent.Based on the consideration of rupture property dispersion, the influences of temperatureand stress fluctuations generated from the change of operating terms or other factors duringthe service process was taken into account in additional. A so-called "Service condition-Rupture property"(SRI) reliability analysis model depending on interference theory wasestablished. The reliability for rupture life under certain condition could be predicted byutilizing Monte-Carlo stochastic simulation on computer. It was observed that the reliabilityof service time would decrease as temperature or stress increased, or the fluctuations of boththese two terms amplified. Because the life prediction approaches applied in engineeringdidn't consider both the rupture property dispersion and service condition alterations at thesame time, the SRI model would bring a new subject for rupture life assessment.