Turbulence measurements in the inlet plane of a centrifugal compressor vaneless diffuser

https://doi.org/10.1016/j.ijheatfluidflow.2008.12.001Get rights and content

Abstract

Detailed flow measurements at the inlet of a centrifugal compressor vaneless diffuser are presented. The mean 3-d velocities and six Reynolds stress components tensor are used to determine the turbulence production terms which lead to total pressure loss. High levels of turbulence kinetic energy were observed in both the blade and passage wakes, but these were only associated with high Reynolds stresses in the blade wakes. For this reason the blade wakes mixed out rapidly, whereas the passage wake maintained its size, but was redistributed across the full length of the shroud wall. Peak levels of Reynolds stress occurred in regions of high velocity shear and streamline curvature which would tend to destabilize the shear gradient. Four regions in the flow are identified as potential sources of loss - the blade wake, the shear layers between passage wake and jet, the thickened hub boundary layer and the interaction region between the secondary flow within the blade wake and the passage vortex. The blade wakes generate most turbulence, with smaller contributions from the hub boundary layer and secondary flows, but no significant contribution is apparent from the passage wake shear layers.

Introduction

The overall efficiency of a centrifugal compressor is equally dependent on the good design of both impeller and diffuser. At the impeller inlet, the flow is uniform but at the impeller outlet the flow is highly distorted and three-dimensional due to strong secondary flows and mixing within the impeller. A major objective of the turbomachinery designer is to minimize the aerodynamic losses which result both within the turbomachine blade passages and downstream as non-uniformities in the flow are mixed out. Such impeller exit flows have been measured by Dean and Senoo, 1960, Krain, 1988, Maksoud and Johnson, 1989, Farge and Johnson, 1992.

The mixing out of this highly distorted and 3-d flow in the diffuser has been observed by previous researchers (Inoue and Cumpsty, 1984, Senoo and Ishida, 1975, Pinarbasi and Johnson, 1994, Hagelstein et al., 2000, Hillewaert and Van den Braembussche, 1999, Shum et al., 2000). The blade wakes were observed to mix out rapidly, while the passage wake, which is located on the shroud side of the passage, mixes out slowly. At the diffuser exit, a Hele–Shaw flow had developed between the diffuser walls.

Researchers showed that, (Johnson and Moore, 1980, Moore et al., 1987) for a turbine cascade, a major contributor to loss production within the turbulent boundary layers and passage vortex is the conversion of mean kinetic energy to turbulent kinetic energy and hence to total pressure loss due to turbulent dissipation. Such a conversion process is only possible where shear is present in the mean flow, e.g. within the boundary layers or within the secondary flow regions associated with a passage vortex. Most turbomachines discharge a complex 3-d flow exhibiting several flow features which will ultimately result in turbulent dissipation and loss. The designers’ problem is in identifying features which are the major contributors to the loss such that he can attempt to modify his design to reduce their effect.

The centrifugal compressor vaneless diffuser is a good example for the study of losses due to turbulent dissipation. The geometry is simple, but because the flow entering the diffuser is highly non-uniform with strong secondary flows, turbulent dissipation leads to relatively high losses. The purpose of this paper is to study the flow mechanisms which lead to the losses within a low speed centrifugal compressor vaneless diffuser.

Section snippets

Experimental procedure

A schematic of the low speed centrifugal compressor rig used in the study is shown in Fig. 1. The impeller was a De Havilland Ghost impeller, the geometry of which is given by Johnson and Moore (1983). In the current study, the original radial outlet section was replaced to provide a 30° backswept outlet angle as detailed in Fig. 2. The vaneless diffuser has straight walls and a constant cross sectional area. The geometry, operating conditions and measurement stations are summarized in Table 1

Experimental results

Results will only be presented at station 1, although mean velocity results from stations 2 and 3 have been used to evaluate the radial gradients of mean velocity using finite difference techniques.

Conclusions

  • (1)

    Four regions of high shear rate are identified within the flow – within the blade wake, between the passage wake and jet, within the thickened hub boundary layer and between the blade wake secondary flow and passage vortex. Each of these regions is associated with high turbulent kinetic energy and with high levels in at least one component of vorticity.

  • (2)

    Significant Reynolds stresses are generated in all these regions except between the passage wake and jet. The author believes that these

References (22)

  • T.Z. Farge et al.

    Effect of flow rate on loss mechanisms in a backswept centrifugal impeller

    J. Heat Fluid Flow

    (1992)
  • R.C. Dean et al.

    Rotating wakes in vaneless diffusers

    ASME J. Basic Eng.

    (1960)
  • Farge, T.Z., Johnson, M.W., Maksoud, T.M.A., 1988. Tip leakage in a centrifugal impeller. ASME Paper...
  • D. Hagelstein et al.

    Experimental and numerical investigation of the flow in a centrifugal compressor volute

    J. Turbomach.

    (2000)
  • Hathaway, M.D., Chriss, R.M., Wood, J.R., Strazisar, A.J., 1992. Experimental and computation of the NASA low speed...
  • K. Hillewaert et al.

    Numerical simulation of impeller–volute interaction in centrifugal compressors

    J. Turbomach.

    (1999)
  • M. Inoue et al.

    Experimental study of the centrifugal impeller discharge flow in vaneless diffusers

    ASME J. Eng. Gas Turbines Power

    (1984)
  • M.W. Johnson et al.

    The development of wake flow in a centrifugal compressor

    ASME J. Eng. Power

    (1980)
  • M.W. Johnson et al.

    Influence of flow rate on the wake in a centrifugal impeller

    ASME J. Eng. Power

    (1983)
  • Jorgensen, F.E., 1971. Directional sensitivity of wire and fibre film probes. DISA Information No. 11, p....
  • H. Krain

    Swirling impeller flow

    ASME J. Turbomach.

    (1988)
  • Cited by (0)

    View full text