Repeat data on a Ti-6246 alloy at 773 K and 580 MPa provided a unique oppor
tunity to identify stochastic creep properties and to use this information
to build a probabilistic creep damage assessment for this alloy. The stocha
stic nature of creep properties both under a single test condition (using a
generalised gamma distribution) and under various test conditions (by comb
ining this distribution with the Monkman-Grant relationship) were identifie
d. Under the single test condition log failure times had a distribution ske
wed to the right and log minimum creep rates had a distribution skewed to t
he left. Over various test conditions a failure time distribution similar t
o the Weibull distribution was most supported by the data. The observed log
time to failure at 480 MPa was well within the extrapolated 95% confidence
limits obtained from the Monkman-Grant model. Uncertainty about the exact
distributional form increased this 95% confidence limit.