UDC 539.434 Known parametric methods and a modified base diagram method were used to analyze extrapolation results for creep rupture diagrams of heat-resistant steels. Two model alternatives are examined: viz when the hypothesis of the "unified" curve is valid and when its application leads to significant errors in prediction of creep strain processes. The efficiency of proposed procedures and their advantages over other approaches are demonstrated.Keywords: extrapolation, parametric methods, creep rupture diagrams, modified base diagram method.
Introduction.Nowadays the life extension of critical structural elements, in particular in heat power engineering of Ukraine, is a currently central problem. Analysis of the condition of constructions and feasibility of their further service is based on creep rupture diagrams of the materials. The latter are plotted by the test results of the specimens in uniaxial and multiaxial stress states under the action of constant stresses and temperatures [1][2][3][4][5]. Creep rupture tests can last from several days to tens of years. Therefore, the construction of mathematical models for creep rupture prediction has been and remains the urgent goal of engineering practice.Despite a great number of creep rupture prediction methods of the materials, the Larson-Miller, OrrSherby-Dorn, Manson-Haferd, Trunin parametric relationships have received the widest practical acceptance [6][7][8][9].The relationships, which reduce the description of the set of creep rupture curves to that of the "unified" curve covering all test temperatures, are termed parametric ones. Then the function that approximates experimental results takes the form s h = F 1 ( ) or t h = F 2 ( ),