Noise on Gullfaks C Takes a Turn for the Better with DRAG®

When Statoil’s Gullfaks C platform recently uprated its compressors, they revamped their piping and replaced their compressor-recycle valves. During initial testing on Train A in early October 2001, they observed excessive noise levels – as high as 126 dBA – within the revamped piping surrounding a new recycle valve and the valve itself, Figure 1.

The piping revamp included five new elbows and two flange joints within 15-ft (4.5-m) of the valve inlet. It also included an 18.5-in. (455-mm) spool piece to reduce the pressure class just downstream of the valve. This was accommodated by a spool piece that included one 1500 ANSI flange joint, one 600 ANSI flange joint and a slight expansion in the spool itself to accommodate the decrease in pipe wall thickness, Figure 2. The flanged joints were all the ring-type joint (RTJ) design.

Because the high frequency noise was judged to be in excess of 85 dBA, and the noise seemed to have a “trigger” point associated with a specific valve lift, CCI was contacted by Statoil and Aker Engineering to assist in determining the source of the noise. An audit confirmed that the valve should not be producing the noise, and that there was no noticeable vibration of the valve stem or actuator, something frequently related to noise.

Statoil and Aker Engineering contracted a consulting firm to learn about the specifics of the observed noise. They carried out detailed and accurate noise measurements under varying compressor and valve operating conditions. The dominant noise frequency was 2000 Hz and the peak noise level was 126 dBA. This peak noise level occurred at the ANSI 1500/600 transitional spool piece. The noise level in the upstream piping ranged from 112 to 115 dBA. Farther downstream, where the piping was insulated, the noise level dropped off from 123 dBA to 113 dBA.

It was decided that a spare valve made at the same time as the one at Train A would be disassembled and inspected while the installed valve continued to be used for compressor testing. It was also decided that at the next shutdown, a flexible perfluoro-elastomer filler material would be installed into the RTJ flange gaps at the inlet and outlet end of the spool piece to assess its impact on reducing the noise at that location. At that time, the installed valve could also be visually inspected through the valve outlet to see if any constructional anomaly could be observed. After this modification was completed, and upon restarts of the Train A compressor, a significant noise reduction was observed (in excess of 10 dBA) such that the dominant noise now seemed to be coming from the upstream side of the valve. The dominant frequency remained at about 2000 Hz.

Based on this experience, it was concluded that the source of the noise was probably not the valve, but rather the piping, and it was further decided to install soft fill-in gaskets in the RTJ flanges upstream of the recycle valve as well. This was carried out a few days later during a scheduled compressor stop. New noise measurements were again taken that showed the previous high-frequency noise in the piping system adjacent to the anti-surge valve had been eliminated.

The final conclusion was that the noise arose solely from significantly increased gas velocity in the piping following the uprated capacity of the export compressor. When the gas velocity in the piping, as determined by the valve opening position, became sufficient, it triggered a resonant frequency in the flange gaps (similar to blowing across the top of an empty bottle).

Checking conventional design rules for gas service pipe sizing also showed that the compressor uprating had produced a significantly increased gas velocity over normal recommendations for the pipe sizes involved. It was further concluded that the anti-surge recycle valves did not represent any significant noise problem in Satoil’s current operations.

Published in SOLUTIONS Summer 2002