Fuzzy logic methods in fatigue and creep

Abstract

Fatigue and creep laboratory data for metals normally exhibit scatter, which implies an element of uncertainty or vagueness in the results. Such data are usually treated by empirical correlations or by mathematical models with some theoretical basis. Confidence limits are sometimes given based upon an assumed probability distribution. On a fine scale, fracture mechanics studies consider the mechanism of crack growth, assuming a uniformly smooth continuum. The treatments aim to provide design working stresses and also the reliability basis for the formulation of inspection and maintenance schedules. Further uncertainty in the interpretation of laboratory test data in terms of in-service material characteristics arises from a few other sources, even if the material types are nominally the same. Fuzzy logic is a natural means of expressing vague categories of information by means of fuzzy sets and also provides the means of performing logic operations on them. In this work, consideration is given to the application to some aspects of fatigue and creep. Some examples, including fuzzy boundaries between safe and unsafe states, are given to illustrate the methodology, the conclusions are also initially in the form of fuzzy sets. Compared with other methodologies, richer meaning is found in the results.