Measured Performance of Two-Stage Centrifugal Compressor Under Wet Gas Conditions

Abstract

The oil and gas industry is moving forward to access the most remote gas reserves and enhance the exploitation of the existing installation or postponing their tail-end. To achieve these accomplishments several technology challenges are being unveiled. In topside upstream application both offshore and onshore, one important technology issue is the capability to compress gas with a significant amount of liquids and it assumes a special interest in case of the facilities revamping. Nevertheless is in the subsea environment where this technology issue becomes really challenging. In order to properly design and size a compressor/motor system for subsea wet gas compression, one must be able to adequately predict the compressor performance with mixed phase flow. This paper presents the results from an experimental test program which investigated the performance of a centrifugal compressor at various wet gas conditions with elevated suction pressure. Performance tests are completed on a two stage centrifugal compressor with a mixture of air and water at suction pressures of 20 bar (300 psi). The compressor is subjected to flow with liquid volume fractions ranging from 0 to 5% along three speedlines. The performance measurements are made in accordance with ASME PTC-10 specifications with an additional torque measurement on the shaft between the compressor and gearbox. At each test condition, once the liquid is injected in the air flow, an increase in pressure ratio occurs. This testifies the compressor is still able to work in presence of water. However, increasing the amount of liquid injected a decreased polytropic head together with an increased absorbed actual power by the compressor cause a deterioration of its efficiency. Moreover when liquid is introduced into the flow, the discharge temperature of the compressor reduces significantly. The performance results and trends mentioned above are reviewed in the detail in this paper.