Estimation of remaining creep life of an aluminium alloy from creep crack density measurements

Abstract

Long transverse test pieces of fully aged RR58 plate were stressed in tension at 278 and 308 MPa at 120° C for various fractions of their creep lives. The test pieces were subsequently sectioned, mechanically and electrolytically polished and the numbers of cracks per square millimetre were measured by optical microscopy. The crack density, n, increased linearly with creep strain ε at both stress levels. No accurate assessment of the variation of n with time was possible. Good agreement between the crack densities measured on duplicate microsections was achieved when the crack density was greater than 10 cracks mm⁻². The crack densities in the uniformly strained portions of 11 test pieces from the same plate, fractured at 150° C at stresses within the range 200 to 290 MPa were also measured. The crack density decreased from 45 cracks mm⁻² at 200 MPa to 4 cracks mm⁻² at 290 MPa. A regression equation n/ge=164 − 0.57σ (where σ is the applied stress) was derived assuming linear n versus ε relationships at 150° C. The 90% confidence limits were derived for the determination of an unknown stress level from a single measurement of n/ε. Of the creep life prediction methods discussed, only the correlation of creep crack density and creep strain is of sufficient accuracy and this only when the creep stress and creep temperature are low, i.e. only for those conditions which would develop a high crack density at small fractions of the creep life.