Creep under Varying Load

Abstract

The creep behaviour of nickel under cyclic loading has been studied, the aim being to relate the creep strain to the metallurgical recovery that occurs during each load reduction. Square-wave loading between stresses σ₁ (40 or 80 N/mm²) and σ₂ (0–σ₁) for times t ₁ (0.01–0.1 h) and t ₂ (0.1–20 h) was applied at 650 or 750° C, and a point of principle was to apply the full stress σ₁ initially until the metal was fully work-hardened so that all subsequent changes could be ascribed to recovery. The strain per cycle consists of three components, which could be expressed in terms of experimental conditions: (1) ∆ε₁, the normal strain in time t ₁ at a stress σ₁; (2) ∆ε₂, the strain in time t ₂ at stress σ₂; ε₂/ε₁ = (σ₂/σ₁)^m, where m = 10 at 650° C and 5.2 at 750° C; (3) ∆ε_r, the extra strain on reloading from σ₂ to σ₁; ∆ε_r = bt ₂ ^⅔ + c, where b and c have been determined. The factor c can be negative, probably because strain-ageing occurs. A consequence is that three functions are needed to describe accurately the state of nickel during creep.

At high temperature and low stress, recovery did not accumulate from cycle to cycle. By fitting the acceleration factor (∆ε₁ + ∆ε₂ + ∆ε_r)/∆ε₁ into a standard creep-strain/time equation, the envelope strain/time relation for cycling is described with satisfactory accuracy. At low temperature and high stress the creep rate steadily accelerated, presumably because recovery accumulated from cycle to cycle. Quite good predictions were made from the first few cycles. In both cases the predictive accuracy is more reassuring than it is with the conventional strain-hardening or time-hardening methods. The physical metallurgical approach enables this final result to be obtained with relatively little experimentation, though on this first occasion many extra experiments were performed to check various points.