Froude Modeling Of The Flixborough "By-Pass" Pipe

The Flixborough works in the United Kingdom was destroyed by a powerful explosion twenty-five years ago (June 1, 1974). The cause of the explosion was determined to be the failure of a temporary pipe bridge (the 'dog-leg' pipe), installed between two reactors. The failure mechanism of the dog-leg pipe has been the subject of extensive research and controversy. To date, no failure mechanism, that did not require excessive over pressure, has been accepted. The flow in a one-sixth scale model of the 'dog-leg' 'by-pass' pipe assembly of the Flixborough chemical plant is examined using Strouhal and Froude number analyses, as well as laser Doppler velocimeter (LDV), video and proximeter displacement transducer measurements. The work investigates the one possibility that appears to have received little consideration - that of a partial circumferential fatigue crack in one of the bellows supporting the ends of the pipe caused by flow induced vibration. In this analysis, the approximate axial, lateral and angular (rocking) natural frequencies of the pipe and bellows assembly were determined. These, scaled to the one-sixth flow model used, were compared to flow frequencies measured by video analysis. LDV, and proximeter displacement measurements as well as those determined from Strouhal and Froude number analyses of the possible flows past the bellows.The results indicate that both axial and lateral near resonance conditions of the bellows and the structure could have occurred. Such conditions may have been responsible for a fatigue initiated two-step failure process that is better supported by the physical facts and eyewitness accounts and does not require the over pressure necessary for the single step simultaneous failure of both bellows accepted, albeit reluctantly, by the Court of Inquiry.