Severed Tubes in Operating Nuclear Steam Generators: Case Analysis — I

During a routine inspection in October 2001, it was discovered that a plugged tube in one of the steam generators at a nuclear station was completely severed at the upper tubesheet. The severed surface appeared to be fresh, clean, 360-deg. and with no evidence of necking down, ductile failure or visible flaws. Eddy current inspection revealed that the four neighboring tubes downstream of the severed tube each had wedge-shaped wear marks starting from the tubesheet end. However, there was no visible mid-span impact wear marks. Furthermore, the tube was visibly swollen due to over-pressurization. This tube was plugged in 1986 with no indications of flaw or wear marks. It was replugged later. The tube was not stabilized because there was no apparent potential for a circumferential sever to occur anywhere in the tube. Flow-induced vibration based on classic linear theory showed that this tube should have wide margin against fluid-elastic instability, turbulence or vortex-induced fatigue failure. However, it also was found that if for any reason the tube was laterally restrained (clamped) at the support plates, the margin against fluid-elastic instability would be reduced significantly. Still, this tube should have been stable and turbulence or vortex-induced vibration alone, as predicted by the classic linear theory without fluid-structure interaction, should not have caused fatigue failure. This study points out a deficiency in the classic linear turbulence-induced vibration analysis without taking into account the effect of fluid-structure coupling. When the cross-flow gap velocity is near the critical velocity, fluid-structure coupling would greatly increase the vibration amplitude over what is predicted by the classic acceptance integral approach. It is this increase in turbulence-induced vibration that caused this particular tube to fail by fatigue.