Pages: 13 Published: Jan 1983
We have often discussed creep as though it could be measured and modeled exactly. In practice neither is possible. There are uncertainties associated with both the measurements and models, and these uncertainties must be estimated and accounted for in both component design and licensing. In fact, the desired estimate of many phenomena is not the best estimate of what will be observed but the probability that the next observation(s) will be outside some limiting value. It may be important to build a model that makes estimates of the following type: “The probability that the creep rate will be between 4.5 × 10−6 and 4.9 × 10−6/h is 0.95.” An estimate that the creep rate will be 4.7 × 10−6/h, on the other hand, may be useless.
Paper ID: STP34417S