Flow-induced vibration testing of replacement thermowell designs

Abstract

Listen

Translate article

Inconel 600 Primary Water Stress Corrosion Cracking (PWSCC) in Nuclear Pressurized Water Reactors (PWRs) has necessitated the repair/replacement of various small bore nozzles. These repairs/replacements must be designed to avoid unwanted vibrations. So, to this end, new RTD-Thermowell-Nozzle replacement designs were developed and subjected to flow testing over a velocity range from 9.14 to 33.53 m/s (30–110 ft/s), and temperatures ranging from 121°C to 316°C (250–600°F). The replacement nozzles are welded on the pipe OD, rather than on the pipe ID. A split, tapered ferrule is used to support the nozzle tip inside the pipe bore. This maintains high thermowell tip-resonance frequencies with the objective of avoiding self-excitation from vortex shedding that is believed to have caused failures in an earlier design during initial, precritical plant startup testing. The flow testing was complicated by the small size of the thermowell tips (5.08 mm or 0.2 in ID), which necessitated use of a complement of low temperature and high temperature instrumentation. Since the high temperature device had an internal resonance (750 Hz) within the frequency range of interest (0–2500 Hz), adequate sensor correlations had to be derived from low temperature tests. The current nozzle/thermowell design was tested concurrently with two slight variations of the replacement design. The acceleration signals were acquired during incremental and continuous flow sweeps, nominally at 5 kHz sampling rates and for time domain processing as high as 25 kHz. Whereas vortex-shedding frequencies were predicted to prevail between 400 and 1500 Hz, no such response was observed at these frequencies. Rather, the thermowell tips responded due to turbulent buffeting with a peak response that was related directly to flow velocity. Lift direction response was always larger than drag direction response. The thermowell tips also responded at their natural tip frequencies in a narrow band random fashion. At the higher flow rates, one replacement design experienced an instability mode leading to high tip stresses. Although this instability did not repeat, this particular design was eliminated from consideration. The second replacement design performed almost identically to the current in-plant design. The experimental data were used to extract forcing functions and thermowell responses that were used as input into the design calculations.