I. INTRODUCTION

The purpose of this paper is to describe pulsation and transient flow phenomena in centrifugal compressor installations and to describe techniques for predicting or simulating these phenomena. Of particular emphasis will be an electronic analog simulator capable of duplicating surge and other fluid transient phenomena in which the compressor interacts with the transient response of its attached piping. Theoretical aspects of this compressor/piping interaction will be discussed briefly to explain the analog action, and performance of the analog will be compared to field data. The discussions will generally be limited to low frequencies in the range of zero to several hundred Hertz. Centrifugal compressors also are noted for high intensity pulsations, noise and turbulence problems well above this frequency range due to blade passing phenomena and high shear flows, but fluid interaction effects between the compressor and its piping are most pronounced at frequencies below the plane wave cutoff frequency of the pipe. Host of the potentially damaging piping span resonances and pulsation resonances occur in this portion of the frequency spectrum as well.

II. BACKGROUND

While compressor surge effects have long been recognized (if not totally understood), other low frequency pulsation phenomena have become increasingly apparent as compressor horsepower and pipe sizes have increased in the past few decades. SwRI has been involved in a variety of such problems over the past 10 years ranging from reciprocating compressor pulsations which apparently throw a centrifugal into surge, to cases where high level pulsations destroy compressor internals when no reciprocating compressor is within 50 miles. A synopsis of salient observations from both field and laboratory studies is given below.

  • 1)

    It has been shown that a compressor can either amplify or attenuate external pulsations (as from a reciprocating compressor, vortex flow past tees, valves, etc.).

  • 2)

    Even with no positive source of pulsations (e.g., a reciprocating compressor) in the piping system, low frequency pulsations can be experienced at levels sufficiently high to fatigue compressor internals or severely shake the piping. Such effects were observed at flows substantially above the normal machine Burge point.

  • 3)

    These pulsation problems can often be mitigated by changing the pulsation response of the compressor piping (lengths, diameters, etc.).

  • 4)

    Frequencies are not harmonically related to, and do not vary with, compressor speed. The severe pulsation frequencies normally correspond to one of the major length resonances of the piping system. Measurements along the piping show a strong standing wave pattern, often existing across the compressor.

  • 5)

    Pulsation levels are most severe when the compressor is situated at or near a velocity maximum in the pulsation standing wave field.

  • 6)

    The onset, frequency, and severity of machine surge can also vary as the piping system is changed. Surge point and frequency as measured in the test stand often bears little relationship to those characteristics observed in the field.

  • 7)

    External pulsations can induce surge in a centrifugal (e.g., paralleling a centrifugal with a reciprocating compressor).

  • 8)

    Some machines exhibit little if any evidence of surge.

This content is only available via PDF.
You can access this article if you purchase or spend a download.