Fatigue is a common failure mode in mechanical components that are subject to repeated load cycles. In nuclear plants the cyclic loading is usually associated with variations of pressure and temperature. The American Society of Mechanical Engineers (ASME) Boiler and Pressure Vessel Code provide fatigue evaluation procedures in ASME III Section NB-3220. It is required to calculate a Cumulative Usage Factor (CUF). The Code criterion for acceptance is that the calculated CUF must be less than one. The ASME method is deterministic and it is assumed that uncertainties are accounted for by largely un-quantified inherent safety margins in the assessment.
The conservatism in the ASME III Code fatigue evaluation methods may come from the assessment procedure used and/or the fatigue design curve. In this paper the inputs to an ASME III fatigue assessment that is based upon a finite element determination of stresses are outlined and assessed. Based upon these inputs possible areas of conservatism within the analysis are then discussed and conclusion drawn on the inherent margins.
In addition a probabilistic model that uses Monte Carlo (MC) methods is developed. By assuming the scatter in the fatigue data, used to generate the design fatigue curve, to be the most significant random variable, the MC model was used to calculate the conditional probability of initiation associated with CUFs from a number of plant representative examples.