Abstract :
Wilshire–Scharning have recently developed a
new methodology that has been demonstrated to deliver
accurate longer term creep life predictions, and so offers
the prospect of cost-effective acquisition of long-term
creep design data. This methodology differs from existing
approaches to creep life prediction by normalising the
applied stress through the appropriate tensile strength. This
article develops a generalisation of this Wilshire–Scharning
model that has the potential to increase the predictive
accuracy of this methodology—which will be so essential
if it is to be adopted as a way of economising on the
acquisition of creep design data. When applied to 1Cr–
1Mo–0.25 V steel, it was found that this generalisation
reduced the average error in prediction from 2.5 under the
Wilshire–Scharning specification to 1.8 years when
extrapolating from 5,000 out to over 100,000 h. Further,
over this time scale the generalised model produces a mean
absolute percentage error of 28%. This compares to 47%
obtained using the traditional 4H projection technique and
26% using a modification of this methodology as recently
proposed by Evans.
