Probabilistic prediction of fretting-wear damage of nuclear fuel rods

The development of a probabilistic method to predict the fretting-wear performance of fuel rods is presented. In this approach, the dynamic response of the nuclear fuel rod and its supports is calculated by a non-linear vibration model (called VITRAN for VIbration TRansient Analysis—Nonlinear). This model predicts the motion of a fuel rod exposed to a turbulent flow and takes into account the presence of the rod-to-grid gaps in the assembly grids. Numerical values for the grid work and wear rates are calculated from the predicted forces during the rod impacts against the supports. Due to the large variability of the fuel assembly parameters affecting the fretting-wear performance, a Monte Carlo method was chosen to perform the transient simulations. As a result, average values, standard deviations and other statistical parameters of the probability density functions of the wear rates and the wear scars population can be estimated. A description of VITRAN and an example of application of the fretting-wear methodology to evaluate the performance of a fuel design are presented.