By Steve Blankinship, Associate Editor
Con-Ed’s East River Station - the world’s largest steam distribution plant - was upgraded and brought on line in the spring of 2005. The East River Repowering Project (ERRP) was undertaken to provide steam from four different sources: two existing high pressure boilers and two new heat recovery steam generators utilizing combined cycle turbine exhaust. The send-out steam from these boilers feeds to a ring header that permits operators to direct the steam to a large district heating system through several control valve stations. One control valve station features five parallel control valves that pass steam to a header and out to the street.
But piping vibrations occurred during the initial shake-down operation following commissioning. They disrupted flow meters and other equipment in the send-out piping. Operation of the No. 1 valve (far left in Figure 1) resulted in the high vibration level. The No. 1 valve is in a 14-inch line that enters the 28-inch outlet header to the left of the header outlet tee. The other four valves enter the header to the right.
 |
A project to investigate the vibrations’ cause was undertaken. The project included pipe testing using dynamic pressure transducers, accelerometers and modeling the acoustics and piping structure. The testing was performed over two days and included multiple valve positions and flow rates to cover a range of operating conditions. It was initially thought that a flow-induced vibration issue related to the operation of the No. 1 valve was the primary cause of the piping vibration. Typically, flow-induced vibration arising from a location in the piping causes the pressure at one or more frequencies to be excited and resonate throughout the piping network.
Large increases in dynamic pressures (by as much as five times) and pipe acceleration (by as much as a four times) were recorded when the No. 1 valve was opened, both alone and in combination with other valves. The proportional increase of these two confirmed flow-induced vibration.
The next issue involved locating the source of the pressure fluctuation. An acoustic model of the system was used to predict pressure amplification over a range of frequencies for various potential sources. The model results were compared to the test data. The acoustic model showed that the No. 1 valve was not the source of the pressure oscillation. Instead, the likely source was either the header inlet tee downstream of the No. 1 valve or the header outlet tee leading to the outlet pipe. Test data indicated that the header outlet tee was a more likely source.
From an engineering perspective, these two tees were closely located, approximately five feet apart or about 2.2 times the internal diameter of the 28-inch header. Local flow separation issues at the first tee may not be resolved when the flow reaches the outlet tee, since this can require three to five pipe diameters.
Pipe stress models were used to quantify pipe stresses caused by the measured vibration and to investigate potential support changes that could eliminate the problem. Unfortunately, the system’s thermal growth is on the order of several inches, while the vibration, although undesirably high, is on the order of tenths of an inch. Therefore, support options were determined to be an unrealistic fix to the vibration problem.
 |
A piping configuration change was seemed the best solution (Figure 2). The modification featured extending the No. 1 valve outlet pipe and shortening the outlet header to bypass the outlet header tee. This would eliminate the vibration’s source if it were caused by flow separation and instability between the two tees.
After the piping modification, tests showed the reconfiguration successfully mitigated the vibration. Vibration and pressure levels were reduced, effectively eliminating the vibration concern. In fact, the pressure and vibration are now lower than during operation of the remaining four control valves.