A Comprehensive Approach to Predicting Vortex-Shedding-Induced
Pulsation Amplitudes in Piping Systems, 18-R8325
Principal Investigators
Eugene L. Broerman
Sarah Simons
Rebecca Owston
Aaron McClung
Nathan Poerner
Inclusive Dates: 07/01/12 – Current
Background — One of the major design criteria for centrifugal compressor piping systems is preventing piping system vibration and valve failures resulting from vortex-shedding-induced (VSI) pulsations. In recent years, computational fluid dynamics (CFD) has been used to theoretically predict the dynamic shaking forces that cause vibration when the subject phenomenon occurs in piping systems. However, a practical validated approach to applying these theoretical methods has yet to be developed. SwRI's current approach to address this concern is a screening type vortex-shedding analysis ("Strouhal analysis") for determining whether there is a coincidence associated with the frequency of the vortex-shedding of the gas flow and the acoustic natural frequency of piping stubs. If a coincidence is predicted based on the screening analysis, piping changes are recommended to avoid the coincidence. These piping changes are typically expensive, time consuming, and possibly unnecessary in some circumstances. There is currently no way to determine the severity of the resulting pulsation amplitudes and, therefore, the severity of possible piping vibration. The lack of an accurate prediction method requires added conservatism that has forced many operators to reduce operating flexibility or make costly piping changes that could otherwise be avoided. VSI pulsation amplitude predictions would allow SwRI to offer a significant cost-saving service to its clients.
Approach — This project is a combined experimental and analytical effort. An experimental Strouhal test program is being conducted in conjunction with a supporting effort that utilizes a commercial CFD package to provide supplemental pulsation amplitude predictions in order to populate pulsation amplitude response surfaces. The response surfaces will be used as the basis for developing a new boundary condition to be implemented in SwRI's proprietary Transient Analysis Pulsation Solver (TAPS). The new boundary condition will incorporate a source term representing the acoustic excitation due to Strouhal-related shedding at a piping stub. This new simulation capability will augment the current SwRI Strouhal screening methods by providing a means of predicting pulsation amplitudes when acoustic resonance cannot be avoided. The primary objectives of this project are to (1) develop validated simulation capabilities (i.e., using TAPS and CFD) to accurately predict the amplitudes of vortex-shedding-induced piping pulsation and (2) gain knowledge during the execution of this investigation that will enable the definition of design guidelines for using clamps to restrain shaking forces resulting from acoustic coincidences in small-bore piping.
Accomplishments — An experimental test matrix has been developed that includes mainline piping diameters ranging from 3 inches to 6 inches, flow conditions associated with Reynolds numbers that range from 2(10)5 to 4(10)6, and branch piping diameters that range from 1.5 inches to 3 inches. Test sections have been fabricated such that appropriate measurements (dynamic pressure, temperature, static pressure, flow, etc.) can be obtained during each data sampling. Laboratory testing is currently under way, and preliminary results show good agreement between the calculated/predicted acoustic natural frequencies and the measured natural frequencies. Vortex-shedding frequencies have also been observed to be at approximately the predicted frequencies. Some of the measured pulsation amplitudes have been lower than anticipated; therefore, additional testing is planned to provide more insight into this unexpected result. In particular, future testing will be performed in a system in which much higher Reynolds numbers can be achieved (approximately 1.7(10)7). Appropriate CFD software packages and turbulence models have been assessed and model verification has begun. Investigations are under way to determine an acceptable numerical approach that will minimize required computational resources.
