On the Prediction of Fretting Wear of Heat Exchanger/Steam Generator Tubing

Abstract

The phenomenon of tube fretting wear in Nuclear Steam Generator Tubing is of significant importance to Canadian nuclear industry, as well as to other users of heat exchangers. It occurs due to the Flow Induced Vibration of the tube bundle due to flow on the shell side of heat exchanger or steam generator. Fretting wear of tubes results in the reduction in the thickness of the tube wall, leading ultimately to failure of the tubes. This type of failure is observed in the regions of U-bend supports and tube supports. It requires the plugging of tubes which in turn requires significant down time and cost to the operators. In the present study, a predictive model for simulating the wear that occurs in the tubing is developed. The results of the model are compared against experimental values obtained for identical contact configuration and from other experimental studies where a similar experimental setup was used. The predictive models incorporate the concepts of surface statistics, contact mechanics and fracture mechanics to calculate wear particle geometry and the number of cycles to failure through specific failure modes. The cyclic strains developed on the wearing surface is used for estimate wear by considering two distinct wear mechanisms, namely cyclic fatigue and ratcheting. The resulting wear volume and the number of cycles to failure are calculated by the history of strain cycling and tangential work equivalence for low cycle fatigue mechanism and by analyzing the history of strain cycling for ratcheting failure. Subsequently, it is shown that under relatively low normal loading levels over a range of friction coefficients, either no plastic deformation results or the plastic zone is contained within the subsurface, far below the wearing surface. But experiments conducted for identical situations show crack growth and wear particle formation. Thus a predictive model is developed based on the principles of fracture mechanics simulating crack growth and particle detachment. Finite element techniques are utilized for the simulation. The values of wear volume and aspect ratio of wear particles from the numerical studies are compared against wear volume and crack growth measured experimentally. It is seen that the experimental values agree well with the numerically computed values from the predictive wear models. This model is proposed to be expanded to study the fretting wear of tubes by incorporating the actual geometry and material combinations involved.