Experience With Empirical Criteria For Rotating Stall In Radial Vaneless Diffusers.

A seven stage, centrifugal compressor, for gas lift service on a North Sea oil production platform, exhibited subsynchronous shaft vibrations during full load, high density, factory tests. Pressure transducers in each stage characterized two stages as stalled and traced the cause of rotor vibration to rotating stall in the diffuser of the last stage. Modifications to that diffuser of the 97 last stage corrected the subsynchronous shaft vibration. The other stage that exhibited characteristics of diffuser rotating stall was left unchanged as it did not yield harmful subsynchro­ nous shaft vibration. The original compressor configuration was designed to meet Senoo and Kinoshita's criterion [ 1, 2] for inception of rotating stall in a vaneless diffuser. However, its diffuser widths, b3, were less than the impeller tip widths, b2, while b3 and b2 widths were equal in Senoo 's experiment. Kobayashi, Nishida, et a!. [3], and Nishida, et a\. [ 4], extended Senoo 's criteria by accounting for a b:/b2 ratio less than one. The subject compressor was analyzed using the Kobayashi, et al. [3], criterion. Data reduction of the testing condition producing diffuser rotating stall demonstrated significant margin to the criterion. The generous shape of the inlet diffuser contour was suggested as being the source of the disparity between the criterion and the data reduction values, EQUIPMENT AND SERVICE The compressor of this study was the high pressure inboard compressor of a two compressor train. The rated conditions (Table 1) included a specific gravity of 0.92 and rated discharge pressure of 174 Bara (2526 psia). The compressor operated between the first and second lateral critical frequencies of the rotor support system. The design resulted in a first critical (approximately 4000 cpm) 23.6 percent below the compressor mir.imum continuous speed, while the second critical (approxi­ mately 10,000 cpm) had a margin of 27.3 percent above maxi­ mum continuous speed. Additional features of this compressor (Figure 1) included dry gas seals and individual stage casing drains. Figure 1. Compressor Schematic. 98 PROCEEDINGS OF THE TWENTY-FOURTH TURBOMACHINERY SYMPOSIUM Table 1. Compressor Rated Conditions. Table 2. Compressor Full Load, High Density, Test Conditions. Qi ACMS (ACFM) 0.61 ( 1301) Qi ACMS (ACFM) 0.66 (1405) pi Bara (psia) 40.3 (585) pi Sara (psia) 34.5 (500) Ti oc ( oF) 35.7 (96.3) Ti oc ( OF) 37.8 ( 100) 3 ( lb /ft 3 ) 52.8 3 (lb /ft 3 ) pi kg/m ( 3. 3) p kg/m 54.3 (3.4) m i m pd Sara (psia) 174.0 (2526) pd Sara (psia) 169.3 (2456) T d oc (F) 150.7 (303.3) Td oc ( OF) 236.2 (457.2) pd kg/m 3 (lb /ft3) 153.3 (9.6) pd kg/m 3 (lb /ft 3) 169.3 (10.58) m m Gas Type (MW) Nat. Gas (26.6) Gas Type (MW) Inert (37.6) N RPM 7381 N RPM 7381 BASELINE TESTING Configuration 1 Initial testing of the high pressure compressor occurred at low pressure on the manufacturer's compressor test stand. Mechan­ ical testing followed the guidelines of API Standard 617 Centrif­ ugal Compressors for General Refinery Service [5], and included a four hour test at maximum continuous speed. Aerodynamic performance was measured under Class III conditions following the guidelines of ASME PTC 10-1965 for Compressors and Exhausters [6]. The compressor scope of supply included eddy current type shaft proximity probes that were used in both tests. External instrumentation consisted of diaphragm type, pressure transmitters, and thermocouples for flange pressure and temper­ ature measurement, respectively. The performance test incorpo­ rated test loop flow measurement by means of an orifice plate. Neither of these tests produced a substantial subsynchronous rotor vibration component as seen by the proximity probes.