Flow-induced vibration (FIV) of nuclear fuel rods remains a persistent cause of fuel failure in pressurized water reactors (PWRs). About 65% to 75% of the IAEA-reviewed fuel-leak events are attributed to rod fretting, and most research has addressed grid-to-rod fretting wear (GTRF) of the cladding. A recent outage at a domestic nuclear power plant (NPP) exposed a different failure mode: the anti-debris filter plate of the bottom nozzle had worn the lower end plug into a nine-cell pattern, and the wear depth reached the welding zone with the cladding, releasing fission gas into the primary circuit. It is the first fuel failure of its kind reported in China, and single-rod analysis cannot explain it, because the load path is set by the full-core flow field, the multi-span support of the rod and the fretting wear of lower end plug countered by anti-debris filter plate. This paper reviews the three threads on which an explanation has to be built: full-core coolant-flow simulation, FIV of fuel rods, and end plug fretting wear modelling. The first two are mature in isolation, but full-core models resolve the flow split and not the pressure fluctuation that drives a single rod, beam models idealise the end plug boundary, and wear models need a contact load that no full-core study provides. A three-scale framework is proposed: it carries the excitation from the full-core flow through the assembly dynamics to a local contact submodel, feeds the worn geometry back into the dynamics, and quantifies the uncertainty of the controlling parameters by interval and probabilistic methods.

flow-induced vibration; PWR fuel assembly; fuel rod end plug fretting wear; anti-debris filter plate; full-core CFD; multiscale coupling; interval uncertainty