抄録
The 105hr rupture strength, which may be obtained by the creep rupture testing over the period of more than twelve years, will undoubtedly provide us a crucial keynote for the design of power generating facilities as well as chemical plants. In this respect, much effort has been made to predict the strength by use of several extrapolation methods. Unfortunately, however, few seem to have been successful.
On the other hand, it is a standard and necessary practice for engineers particularly in electric power companies to evaluate, in short time, the creep properties of defective materials in case of troubles. Hence, they are in an urgent need to know how they can estimate and what is the best method for the prediction of creep rupture strength.
To meet the aforementioned requirement, an attempt has been made in the present study to clarify to what extent each extrapolation method is successful for the prediction. Five extrapolation methods have been selected and applied to the creep data originally published by the National Research Institute for Metals. In fact, these data are outcomes of intensive and consecutive work of creep testing of more than several years.
3×104hr and 105hr strength of twentyeight heats of three different types of high temperature steels have been predicted by using these methods on the basis of a vertually limited size of creep data (i. e. partial aggregation of creep data). Secondly, the predicted strength was compared with the corresponding actual data with their coincidence being evaluated statistically.
It was found that 3×104hr rupture strength can be predicted within fifteen percent error by using Larson-Miller equation with 3rd order regression on the basis of the creep rupture data up to 3×103hr.
