1979 年 28 巻 308 号 p. 365-371
Logarithmic plastic strain rate (log εp) versus logarithmic residual stress (logσr) curves of stress relaxation data show an inflection of a convex nature in a certain stress range. Although the low stress range of the curve has been called the secondary relaxation and considered to correspond to the steady state creep, no experimental verification has been given.
In this paper, the correspondence of the secondary relaxation to the steady state creep is examined using long-time stress relaxation data up to 10000hr on 1Cr-0.5Mo-0.25V, 12Cr-1Mo-1W-0.25V and 19Cr-9Ni-Mo-W-Nb high temperature bolting steels.
Good correlation could not be obtained between the residual stress at the point of inflection on log σr vs. log εp curve and the stress estimated graphically from a trace of plastic strain of stress relaxation on the transient point of creep curves using the strain-hardening theory. The comparison of the time at the point of inflection showed no good correlation between stress relaxation and creep. There was no good agreement between the dependence of the plastic strain rate on residual stress at secondary relaxation and that of the steady state creep on creep stress. The apparent activation energy of relaxation was different between at higher stresses (primary relaxation) and at lower stresses (secondary relaxation). A good linear relation between the time at the point of inflection on log σr vs. log εp curves and that at the starting-point of structural changes was observed.
It was concluded that the secondary relaxation does not correspond to the steady state creep, but corresponds to the metallurgical structural changes during the course of relaxation.