A probabilistic assessment technique for shell-and-tube heat exchanger inspection

Plow-induced vibration is a common phenomena in shell-and-tube heat exchangers. The resulting vibrations can lead to component failure by fretting wear or fatigue. The intent of thus research project is to develop a methodology that can be used at the design stage to give an overall estimate of service life and during component operation to monitor known flaws and ensure that they will not fail during operation.;The vibration response of a four-span tube subjected to turbulence and low-level Hurid-elastic forces was calculated for varions support conditions. It was determined that the clearance at the supports has a large effect out the critical flow velocity. As the clearance is increased, the critical flow velocity will increase. This is associated with an increase in the overall system damping. As the tube impacts the supports, energy is transferred into higher modes of vibration which dampen out more quickly. It is also shown that friction at the supports has no effect on the vibration response below the critical flow velocity.;An initial through-wall crack is assumed directly beside the tube sheet where the bending stress is highest. The crack propagates due to cyclical stress caused by vibration and is allowed to grow up to fracture. A series of Monte Carlo simulations were run where the clearance at the supports and the support friction were varied according to probability density functions. It is shown that the time-to-fracture for the crack is highly dependent the clearance of the loose support directly beside the tube sheet. As the clearance is increased the overall displacement and vibration amplitudes increase resulting in higher cyclical stresses. It is shown that a Weibull distribution can be fit reasonably well to the time-to-fracture data for the crack. Leak rate calculations are made on the crack as it grows toward fracture. This data shows how the leak rate changes with time depending an the input conditions and can he used to determine the maximum leakage before fracture. A Weibull distribution is also fit to leak rate data based on the time for the leak rate to reach 5 kg/day. This statistical data can then be used in an analysis to see if LBB criteria will apply.;Leak-before-break (LBB) is an analysis based on the assumption that there is negligible chance that a flaw will fail before a leak can be detected. As part of this project. leak rate calculation are made on growing flaws. This information can then be used when assessing whether LBB criteria will apply.