Axial flow characteristics within a screw compressor.

Angle-resolved axial mean flow and turbulence characteristics were measured inside the working chamber of the male rotor of a screw compressor with high spatial and temporal resolution using laser Doppler velocimetry at two rotor speeds, 750 and 1000 rpm. Measurements were performed through a transparent window near the discharge port to allow the application of various laser techniques. The results showed that an angular resolution up to 2[degrees] could fully describe the flow variation inside the chamber. The cyclic flow variation between different working chambers was found to be similar in both the mean and turbulence velocities. The effect of the discharge port opening on the axial mean and root mean square velocities was found to be significant near the leading edge of the rotors, causing a steep increase in mean and root mean square velocities of the order of 4.2 times the pitch velocity, [V.sub.p] . This effect is less pronounced on the flow near the root of the rotor; large fluctuations and instability in the mean flow was caused by rapid flow expansion during the port opening. The obtained data will be used to validate a computational fluid dynamics model of the fluid flow within twin screw compressors, which could allow reliable optimization of various compressor designs.

INTRODUCTION

The use of screw compressors in the HVAC industry is widespread, as they have replaced the traditional reciprocating compressor in a large range of applications, particularly in air compression and refrigerant compression as well as engine supercharging in the automotive industry. Improvements in screw compressors are continually sought in order to increase their performance and reduce energy consumption, noise generation, and manufacturing costs. Screw compressor rotors are contained in a casing where their meshing lobes form a series of working chambers within which the compression process takes place. As the rotors turn, air is admitted through the space between the rotor lobes and the suction port. Further rotation of the rotor leads to cutoff of the suction port, and the trapped volume is pushed forward axially and circumferentially toward the discharge port by the action of the screw rotors, during which the trapped volume in each passage is reduced and its pressure is increased. This process continues until the working volume between the rotors is exposed to the discharge port, allowing gas with high pressure to flow out (Stosic 1998). The performance of such compressors depends on the flow field characteristics of the gas trapped between the chambers of the screw rotors and the housing. It is thus essential to have a good understanding of the gas motion within the compressor by quantifying the velocity field in the suction, discharge, and working chambers, as well as the flow through the clearance gaps, in order to complete the picture about the sequence of processes that occur within the compressor. These velocities are presently estimated today by means of mathematical models of different levels of accuracy and, to the authors' best knowledge, no velocity measurements within a screw compressor have been published and no attempt to do so has been reported.

The material presented in this paper is the initial phase of a long-term research project attempting to measure the mean velocity distribution and the corresponding turbulence fluctuations at various cross sections of the compressor's working volume within the interlobe space at different phase angles. This is expected to reveal how major features of the heat and fluid flow within the compressor are affected by the rotor and lobe geometry. In addition, the major integral properties of the compressor, including the suction and discharge flow velocities, pressures, and temperatures, as well as the torque/power, will be measured with various instruments and compared with predicted values of the same properties derived from an existing computational fluid dynamics (CFD) model (Kovacevic et al. 2000, 2002); the latter is under development and can assist in the design of future screw compressors. Potential benefits to be derived from such a project are the ability to design rotors and other components with correct allowance for distortion and increased compressor efficiencies due to minimization of clearances.

As described above, the flow in screw compressors is complex, three-dimensional, and strongly time dependent, i.e., it is similar to those in the cylinders of gasoline and diesel engines (Arcoumanis and Whitelaw 1987; Kampanis et al. 2001), centrifugal pumps (Liu et al. 1990), or in related and mixed-flow turbines (Zaidi and Elder 1993; Arcoumanis et al. 1997, 1998) and mixing reactors agitated by turbines (Hockey and Nouri 1996; Distelhoff et al. 1997). This implies that the measuring instrumentation must be robust to withstand the unsteady aerodynamic forces, have high spatial and temporal resolution, and most importantly must not disturb the flow. Only point optical diagnostics like laser Doppler velocimetry (LDV) can fulfill these requirements, as was demonstrated by previous research in similar flows. Recent relevant experimental work by Schleer and Abtahi (2006) made use of an LDV system to measure the flow characteristics within a vaneless parallel diffuser downstream of the impeller. The spatially and temporally resolved data showed a complex and highly vortical flow structure within the diffuser that depended on the relative tip clearance.

The first phase of the proposed method of research is the characterization of the axial component of the fluid velocity and turbulence fluctuations at a range of preselected measurement points inside the working chamber from the root of the rotor to its tip using an existing dual beam laser Doppler velocimeter; other components of the flow, in particular the circumferential component, will be measured in the next phase, since it requires a specially designed optical window. These results are the continuation of Nouri et al.'s (2006) work and will be complemented by flow field measurements using high-speed particle image velocimetry (PIV) in the near future. The main advantage of PIV is its ability to characterize two-dimensional (rather than point) instantaneous flow fields, thus revealing two-dimensional angle-phased flow structures. A parallel work by the authors, a comparison between the axial velocity measurements and predictions obtained by an in-house CFD model, is in progress and will be presented in a future publication.

Nouri et al. (2006) established the feasibility of using LDV inside the compressor, but they also identified possible sources of problems. The main objective of the present work was to expand the angle-resolved axial velocity measurements to characterize the axial flow within the rotors using the same dual-beam LDV system at several axial locations near the discharge port and at different radial positions inside the working chamber. A transparent window made of acrylic (Perspex) was installed on the pressure side and was large enough to allow near backscatter light collection. The internal profile of the window was exactly the same as that of the rotor's casing (see Figure 2c) in order to minimize any flow disturbance and preserve the integrity of the flow motion. The only difference is the roughness of the wall, which is smoother in the case of the window. The following sections of the paper, in turn, describe the flow configuration and experimental techniques, discuss the results, and summarize the main findings and their implications.

FLOW CONFIGURATION AND MEASURING INSTRUMENTS

The performance test measurements were performed in the compressor laboratory of City University. The test rig meets Pneurop/Cagi PN2CPTC1 requirements for screw compressor acceptance tests. The compressor was tested according to ISO 1217, Displacement Compressors--Acceptance Tests (ISO 1996), and its delivery flow was measured following ISO 5167, Measurement of Fluid Flow by Means of Pressure Differential Devices (ISO 1991). High-accuracy test equipment was used for the measurement of all relevant parameters. All measurements were taken by transducers and both recorded and processed in a computerized data logger for real-time presentation, as was, for example, the data reported by Stosic et al. (2006). The test rig layout is presented in Figure 1. A 75 kW electrical motor with variable speed controlled by a frequency converter was used as a prime mover via a torque meter.

[FIGURE 1 OMITTED]

For the LDV velocity measurements, a standard screw compressor was used. The test rig was modified to accommodate the new optical compressor and the transmitting and collecting optics and their traverses. The transparent window was machined from Perspex to the exact internal profile as the rotor case (Figure 2a) and is positioned at the pressure side of the compressor near the discharge port, as shown in Figure 2b. After machining the internal and external surfaces of the window, the surfaces were fully polished to allow optical access. Angle-resolved axial velocity measurements were performed along a radial plane inside the working chambers of the male rotor at three locations from the discharge port. Figure 3 shows the adopted coordinate system for the axial velocity measurements within the working chamber geometry of the male and female rotors and the interrogation area; there are five working chambers with the male rotor in one cycle. The radial plane along which the velocity measurements were made was at [[alpha].sub.p] = 27[degrees] to vertical plane at a range of axial distances of [H.sub.dp] = 58-75 mm from the center of the discharge port. The zero radial position coincides with the center of the rotor, but the radial locations within the working chamber vary from [R.sub.P] = 42 mm near the root of the rotor to [R.sub.P] = 62 mm near the tip of the rotor. The direction of the measured axial velocity component is along the axes of the rotor with positive values from suction port to discharge port as shown in Figure 3.

[FIGURE 2 OMITTED]

Under normal operation of the compressor, oil was injected for cooling purposes. The presence of the oil, even at the lowest possible flow rate, caused serious flooding of the oil inside the working chamber, especially on the inner surface of the optical window, forming unstable thick oil film. Under such condition, obviously no LDV or any other optical method is practical, as the laser beams passing through this oil film will be subjected to major distortion, causing uncertainties in the geometry of the measuring volume of unknown magnitude, in particular the crossing of the two beams and the measuring volume location. Thus, in order to be able to measure the velocity field, it was decided to run the compressor with no oil injection at the lower speed. Under this lower speed range, the maximum air temperature was 60[degrees]C [+ or -] 3[degrees]C for a speed of 1000 rpm, and the corresponding value at 750 rpm was 55[degrees]C [+ or -] 3[degrees]C. Despite the imposed working conditions, internal compression ensures a fluid structure that should be comparable with a CFD simulation.

Air was used as the working fluid, and the flow regime was identified by the Reynolds number defined as Re = [V.sub.p][D.sub.p]/[NU], where [V.sub.p] is the male rotor pitch velocity (the blade's velocity) calculated to be 2.957 and 3.94 m/s for rotor speeds, N, of 750 and 1000 rpm, respectively; [D.sub.p] is the male rotor pitch diameter, which is 81.8 mm; and [NU] is the air kinematic viscosity. For the calculation of the blade's pitched velocity, the rotor's lead helix angle was 42.63[degrees]. The calculated Reynolds number was found to be 13,200 and 17,100 for rotor speeds of 750 and 1000 rpm, respectively; therefore, the flows under these speeds can be considered to be turbulent. The volumetric flow rates through the compressor were measured by an orifice plate installed in the exhaust pipe and were 0.842 [m.sup.3]/min and 1.047 [m.sup.3]/min for rotational speeds of 750 and 1000 rpm, respectively. For the LDV application, it is essential to seed the flow so that the seeding droplets can scatter light when passing through the measuring volume from which their velocity can be obtained. The size of these particles/droplets should be small enough to ensure that all of them are following not only the mean flow but also the velocity fluctuations. For that reason, a silicone oil atomiser, purposely made for LDV measurements, was used; it can produce droplet sizes in the range of up to 1 to 2 mm. A low-viscosity silicone oil of 5 cSt was used in this study.

The laser Doppler velocimeter was operated in the dual-beam near full backscatter mode and is composed of a 600 mW argon-Ion laser, a diffraction-grating unit to divide the light beam into two equal intensity beams and to provide frequency shift, and collimating and focusing lenses to form the control volume. A fiber optic was used to direct the laser beam from the laser to the transmitting optics, and a mirror was used to direct the beams from the transmitting optics into the compressor through the transparent window, as shown in Figure 2c. The collecting optics were positioned around 25[degrees] to the full backscatter position and were composed of collimating and focusing lenses, a 100 mm pinhole, and a photomultiplier equipped with an amplifier. The size of the pinhole forms the effective length of the measuring volume, and the measuring volume diameter and fringe spacing were calculated from the geometrical optics to be 79 and 4.33 [micro]m, respectively. The signal from the photomultiplier was processed by a new TSI processor interfaced to a personal computer and led to angled-average values of the mean and root mean square (rms) velocities. In order to synchronize the velocity measurements with respect to the location of the lobs (blades), a shaft encoder that provides a pulse per revolution and 3600 train pulses, giving an angular resolution of 0.1[degrees], was used; it was fixed at the end of a driving shaft. Instantaneous velocity measurements were made over thousands of shaft rotations to provide a sufficient number of samples, which in the present study was set at 50,000 samples.

A Matlab program was written in which the information of the shaft's angular position from the shaft encoder was used to resolve the velocity with respect to the rotor--so-called "gated" measurements. This was done by collecting the sum of all the instantaneous velocities over a given time-window (2[degrees] in this experiment); then the ensemble mean, U, and rms, u, values were calculated, as presented in Figure 4. This method of gated measurements was efficient since the data was collected continuously as the rotor turned and provided ensemble averages for every 2[degrees] over the entire 360[degrees] cycle. The overall measuring time at each point was up to 25 min, which gave more than 600 samples per time-window and provided good statistical uncertainties of less than 1.6% and 5.5% in the ensemble mean and rms velocities based on a 95% confidence level and a velocity fluctuation of 20% of the mean value. A second program was written to calculate the locations of the measuring volume in the compressor, including the effects of light reflection and refraction as it passes through the transparent window.

[FIGURE 3 OMITTED]

TEST RESULTS AND DISCUSSION

The results of axial mean and rms velocities are presented in Figures 5 through 12 to quantify the effects of different resolving angles, chamber-to-chamber flow velocity variations, and axial flow characteristics at different radial and axial locations. Note that in all the figures, the rotation of the rotor is counterclockwise so that the region immediately behind the rotor's blade is referred to as the trailing edge region and in front of the blade is referred to as the leading edge region, as indicated in the figures.

The effect of resolving angle resolution on the mean and rms velocity was examined first; a sample is shown in Figure 5 where the mean and rms velocities resolved over two different angles of 0.5[degrees] and 2[degrees]. It is clearly evident that the mean flow obtained at both resolving angles leads to the same mean flow pattern with the same magnitude; the overall fluctuation in mean velocities is well within the experimental error. Similar behavior can be seen in the turbulence velocity fluctuation variations with almost the same profiles. The only difference is the number of samples, which obviously is much higher with the 2[degrees] resolution. The results suggest that the 2[degrees] resolved angle-widows give the same mean and turbulence flow structures as those of 0.5[degrees], and thus an angle-window of 2[degrees] would be sufficient to describe the flow. In the present paper, time resolutions of 1.5[degrees] and 2[degrees] were chosen to resolve the mean and rms velocities without losing any mean and turbulence flow structure details. The main advantage of selecting the 2[degrees] resolved angle-window over the 0.5[degrees] is that the number of samples in the larger window is much higher, which minimizes the statistical uncertainties in the values of mean and rms velocities.

[FIGURE 4 OMITTED]

Chamber-to-Chamber Flow Variation

The mean flow variation from one working chamber to the next is presented in Figure 6 for a speed of 1000 rpm, a distance [H.sub.dp] of 73 mm from the discharge port, and at two radial locations, [R.sub.p] = 42 mm (near the root of the rotor) and [R.sub.p] = 62 mm (near the tip of the rotor). In general, a similar mean flow variation can be seen in all the working chambers except chamber number 1, where the mean flow is substantially different from the other four chambers, especially in the region close to the leading edge of the rotor. This can be due to some defect in the geometrical boundary of flow chamber number 1 and will be investigated once the compressor is dismantled; this flow defect was observed at all the measuring points and at all operating conditions. The mean flow structures in the other chambers (numbers 2 through 5) are remarkably similar, with a maximum deviation up to 10% close to the leading edges of the rotor. This similarity will allow combining the results of these working chambers to produce a single curve to present the mean flow variation inside the working chamber from the trailing edge to the leading edge and will be referred to as the overall averaged value. It should be noted that the angular flow passage at [R.sub.p] = 42 mm (well inside the working chamber near the root) is around 11[degrees]. This is much smaller than that at [R.sub.p] = 62 mm (near the tip), which is around 55[degrees], as would be expected due to the change of geometry of the rotor's profile (as is evident from Figure 3).

[FIGURE 5 OMITTED]

Similar results to that of Figure 6 are presented in Figure 7, but for turbulence velocity fluctuation; the same conclusion can be drawn from the results except that the difference between the rms velocity profile of working chamber number 1 and the other working chambers (2 through 5) is not as large as those observed in the mean flow variation of Figure 6. Again, the same conclusions were obtained at all radial locations and all operating conditions.

[FIGURE 6 OMITTED]

[FIGURE 7 OMITTED]

Axial Flow Characteristics

Mean axial flow variations inside the working chamber at different radial locations from [R.sub.p] = 42 mm (near the root of the rotor) to [R.sub.p] = 62 mm (near the tip of the rotor) are presented in Figure 8 for a speed of 1000 rpm. In Figure 8a (near the root of the working chamber) the axial mean velocity variation exhibits a near-uniform profile with velocity increasing slowly from the trailing edge to a peak value of 4.1 m/s (or 1.04[V.sub.p]), at around 4[degrees] behind the blade, and then decreasing toward the leading edge. As the radius increases (Figures 8b to 8d), the velocity variation from the trailing to leading edges becomes steeper, with the peak value shifting toward the trailing edge so that at [R.sub.p] = 62 mm, near the tip of the rotor; the maximum velocity is at the trailing edge with a value of 6.5 m/s (or 1.64[V.sub.p]). This jet-like flow at the trailing edge region is due to the jet flow leaking through the clearance between the tip of the rotor and the compressor casing. At this radial location, the velocity values close to the leading edge are almost uniform at around 2 m/s (or 1.64[V.sub.p]). A larger axial mean velocity than the pitched blade velocity, [V.sub.p], in the trailing edge region is expected, because as the air is squeezed by the action of the rotor it tries to escape to the next working chamber by passing through the clearance gap where the flow accelerates substantially and enters into the next chamber with a much higher velocity.

[FIGURE 8 OMITTED]

The turbulence velocity fluctuation at all radial locations shows, in general, similar patterns at all radial locations with the maximum value at the trailing edge and gradual reduction towards the leading edge. However, the turbulence levels at different radial locations are markedly different so that near the bottom of the chamber the rms velocities are higher, and as the radius increases the rms levels decrease, with the minimum values near the tip of the blades at [R.sub.p] = 62 mm. This suggests that the flow mixing is better deep inside the working chamber than at the near-tip region. It should be noted that at this axial location, [H.sub.dp] = 73 mm, no sign of the discharge port opening can be observed from the results of Figure 8.

The differences in the mean and rms velocities at different radial locations can be shown more clearly by plotting their curves on the same graph, as presented in Figure 9. The mean axial velocity profiles (Figure 9a) exhibit a peaked profile at the lowest measured radius, r = 42 mm, near the bottom of the chamber, and as the flow passage expands at the larger radii, the velocity profiles are modified with a shift of the peak value toward the trailing edge of the rotor. The velocity profile near the tip, r = 62 mm, has a maximum value at the trailing edge and reduces rapidly toward the leading edge, where it become almost uniform. This flow transformation with radius suggests that a strong transverse flow (in radial and particularly in circumferential directions) exists in the working chamber and needs to be identified in the next phase of measurements. A comparison between the turbulence velocity fluctuations at different radial positions, given in Figure 8b, clearly shows a reduction in turbulence level with angular position at all radii. It is also evident that the level of turbulence is substantially higher at the lower radii near the trailing edge of the blade, with a maximum difference of up to 40% between [R.sub.p] = 42 mm and [R.sub.p] = 62 mm. These differences reduce with angular position so that near the leading edge they become almost similar.

[FIGURE 9 OMITTED]

Preliminary results at smaller distances, [H.sub.dp], to the discharge port showed that the axial flow was affected significantly by the discharge port opening and closing as the rotor turns. It was, therefore, important to understand how these changes in flow took place, and for this reason similar measurements to those of Figure 8 were made at other axial distances from the discharge port; they are presented in Figures 10 through 12. The closest distance to the discharge port that was possible to reach for LDV measurements was [H.sub.dp] = 59 mm (Figure 10), which shows the maximum effect of the discharge port opening and closing. Near the root of the rotor at [R.sub.p] = 42 mm (Figure 10a), the axial mean and rms velocities are similar to those at [H.sub.dp] = 73 mm, with a uniform peaked-like profile across the chamber. As the radius increases (Figures 10b and 10c), the velocity variation from the trailing edge to the leading edge becomes steeper, with the peak value shifting toward the trailing edge and again similar to those obtained at [H.sub.dp] = 73 mm. However, the flow near the leading edge of the rotor at this axial distance is significantly different from that at [H.sub.dp] = 73 mm so that a massive increase in axial mean velocity can be seen with a maximum value of around 22 m/s (or 5.6[V.sub.p]) near the tip of the rotor. This large increase in axial velocity coincides with the opening of the discharge port (around an angle of 60[degrees]; Figures 10a and 10b) as the high pressure air rushes into the discharge port.

[FIGURE 10 OMITTED]

The turbulence velocity fluctuation at all radial locations follows the mean flow variation and shows similar variation to that observed at [H.sub.dp] = 73 mm, except near the leading edge, where the discharge port opening and closing caused significant change in the mean axial flow. The rms velocity values are very high and even higher than the mean values; the instantaneous variation of velocity in this region (not presented here) was very unstable, with large fluctuations. The high level of turbulence in this region (Figures 10b and 10c) is caused by very high shearing flow of rapid acceleration (during port opening) and deceleration (fast expanding flow), which also indicates that the flow can also be highly three-dimensional during the expansion process. Comparison between Figures 10a, 10b, and 10c suggests that the effect of the discharge port opening is almost negligible near the root of the rotor, [R.sub.p] = 42 mm, and becomes more pronounced as [R.sub.p] increases with the maximum effect at [R.sub.p] = 63.2 mm near the tip of the rotor. These results suggest that information on other velocity components is essential to describe the flow characteristics fully; as mentioned before, this is one of the objectives of the next phase of measurements. It is also expected that the flow complexity would be more at closer distances to the discharge port with larger pressure fluctuation (or perhaps a pressure wave).

From the results presented in Figures 8 through 10, it is evident that the axial flow was affected significantly by the presence of the discharge port, especially near the leading edge during its opening. To understand the flow development with the axial distance, [H.sub.dp], detailed measurements were made at different axial positions; a sample of these results are presented in Figures 11 and 12, which show the angular variation of the axial mean and rms velocities for different distances from the discharge port at a radial position near the root of the rotor, [R.sub.p] = 46 mm in Figure 11, and near the tip of the rotor, [R.sub.p] = 63.2 mm in Figure 12.

[FIGURE 11 OMITTED]

Near the root of the rotor at [R.sub.p] = 46 mm and [H.sub.dp] = 75 mm from the port, the mean velocity profile shows a uniform peaked profile (Figure 11a), as described above. As the flow approaches the port, the velocity profile becomes slightly steeper, with a tendency for the peak location to move closer to the trailing edge. The major effect of the port presence can be seen close to the leading edge of the rotor at [H.sub.dp] = 58 mm (Figure 11d), where a substantial decrease in mean velocity can be seen; this can be due to the fact that the flow passage of the port cavity is situated at higher radial location than [R.sub.p] = 46 mm. Apart from this, no other major effect can be seen at this radial location. The results also show no significant change in the turbulence velocity with [H.sub.dp] at this radial location; the trend is that the rms level is highest near the trailing edge and reduces to a minimum near the leading edge of the rotor. It should be noted that the angles of the flow passage change at different axial locations, which is due to the helical nature of the rotors.

[FIGURE 12 OMITTED]

On the other hand, near the tip of the rotor at [R.sub.p] = 63.2 mm, the effect of the discharge port opening is much more significant as the flow approaches the port, in particular, close to the leading edge of the rotor. At [H.sub.dp] = 75 mm (Figure 12a), the mean velocity profile shows the expected pattern, as described above, with maximum velocity near the trailing edge and reduction to a minimum and uniform level near the leading edge. As the flow approaches the port, similar velocity profiles can be seen from the trailing edge to the point where the velocity becomes minimum and uniform. From this point onward, the mean flow is subjected to a major change due to the opening of the port so that at [H.sub.dp] = 71 mm, near the leading edge, a small but distinguishable increase in mean velocity can be seen. Closer to the port at [H.sub.dp] = 65 mm (Figure 12c), this increase in mean flow becomes considerable so that the maximum velocity near the leading edge becomes 1.5 times larger than the peak value near the trailing edge. The increase and decrease in the mean velocity is due to the opening and closing of the discharge port with the opening being close to an angle of 60[degrees], where the mean velocity starts to increase. The turbulence velocity fluctuations follow the mean flow variation and increase considerably with a magnitude comparable to the mean velocity due to the reasons already given above. A similar trend in both mean and rms velocities, but broader and stronger, can be seen closer to the port at [H.sub.dp] = 58 mm (Figure 12d). As discussed previously, this high level of change in the mean and turbulence velocities in this region suggests that the flow during port opening time is expanding very rapidly as the high pressure air tries to rush out of the port in all directions, causing an unstable and highly three-dimensional flow.

Finally, limited measurements at a rotational speed of 750 rpm were obtained to quantify the effect of the rotational speed; they are presented in Figure 13. These are similar to those of Figure 8 for 1000 rpm. The results show the same trend as those with 1000 rpm in both the mean and rms velocities. The maximum mean axial velocity at the trailing edge was 5.3 m/s (or 1.75[V.sub.p]) (Figure 13d). Within the measured range, the results show that a good degree of scaling exists between the normalized values of the mean and rms velocities.

[FIGURE 13 OMITTED]

The next phase of flow measurements will aim to characterize the flow within the female rotor at the exit of the discharge port and at the inlet of the suction port and measure the tangential mean and rms velocities within the working chamber of the male and female rotors. A special optical window is required to do the latter measurements. This has been designed, and preliminary testing is already in progress.

CONCLUSION

Angle-resolved axial mean flow and turbulence characteristics were successfully performed with high spatial and temporal resolution inside the working chamber of the male rotor of a screw compressor using laser Doppler velocimetry at different radial and axial positions and at operating rotor speeds of 750 and 1000 rpm. A summary of the main findings follows.

1. Averaged velocity values resolved over different angular (temporal) resolutions revealed that both the axial mean and rms velocities remained unchanged for an angular resolution of up to 2[degrees].

2. Chamber-to-chamber axial mean and turbulent flow variations were found to be very similar, with differences up to 10% near the leading edge of the rotor.

3. The velocity results showed that, in general, the mean axial flow within the working chamber reduced from the trailing edge toward the leading edge with velocity values larger than the rotor pitched velocity, [V.sub.p], in particular near the trailing edge region by up to 1.64 and 1.75 times for rotational speeds of 750 and 1000 rpm, respectively.

4. A similar trend was observed in the rms velocity fluctuations, with maximum levels behind the rotor and minimum in front of it; the turbulence levels near the root of the rotor were up to 40% higher than those near the tip of the rotor.

5. The effect of discharge port opening on the axial mean and rms velocities was found to be significant near the leading edge of the rotors, giving rise to high mean and rms velocities of the order of 4.2[V.sub.p] as the high-pressure airflow expanded quickly and produced a complex, unstable, and three-dimensional flow with very steep velocity gradients.

6. The effect of discharge port opening was less effective on the flow near the root of the rotor but increased with radius, [R.sub.p], so that the maximum impact of the port opening was observed near the tip of the rotor.

ACKNOWLEDGMENTS

Financial support from EPSRC under Research Grant EP/C541456/1, from Trane, Lotus, and Gardner Denver is gratefully acknowledged. The authors would like to thank Prof. I. Smith for his valuable discussions and helpful suggestions, and also Michael Smith, Jim Ford, and Grant Clow for valuable technical support during the course of this work.

REFERENCES

Arcoumanis, C., and J.H. Whitelaw. 1987. Flow mechanics of internal combustion engines: A review. Proceedings of IMechE 201(C1):57-74.

Arcoumanis, C., R. Martinez-Botas, J.M. Nouri, and C.C. Su. 1997. Performance and exit flow characteristics of mixed flow turbines. Int. Journal of Rotating Machinery 3(4):277-93.

Arcoumanis, C., R. Martinez-Botas, J.M. Nouri, and C.C. Su. 1998. Inlet and exit flow characteristics of mixed flow turbines. Paper 98-GT-495, ASME International Gas Turbine and Aeroengine Congress and Exhibition, June 2-5, Stockholm.

Distelhoff, M.F.W., A.J. Marquis, J.M. Nouri, and J.H. Whitelaw. 1997. Power and concentration measurements in a stirred tank with different impellers. The Canadian Journal of Chemical Engineering 75:641-52.

Hockey, R.M., and J.M. Nouri. 1996. Turbulent flow in a baffled vessel strirred by a 60 pitched blade impeller. Chem. Eng. Science 51:4405-21.

ISO. 1996. ISO 1217:1996, Displacement Compressors--Acceptance Tests. Geneva: International Organization for Standardization.

ISO. 1991. ISO 5167-1:1991, Measurement of Fluid Flow by Means of Pressure Differential Devices. Geneva: International Organization for Standardization.

Kampanis N., C. Arcoumanis, R. Kato, and S. Kometani. 2001. Flow, combustion and emissions in a five-valve research gasoline engine. Society of Automotive Engineering paper 2001-013556.

Kovacevic A., N. Stosic, and I.K. Smith. 2000. The CFD analysis of a screw compressor suction. Int. Compressor Eng. Conf., Purdue, p. 909.

Kovacevic A., N. Stosic, and I.K. Smith. 2002. CFD analysis of screw compressor performance. In Advances of CFD in Fluid Machinery Design, ed. R.L. Elder, A. Tourlidakis, and M.K. Yates. London: Professional Engineering Publishing.

Liu, C.H., J.M. Nouri, C. Vafidis, and J.H. Whitelaw. 1990. Experimental study of flow in a centrifugal pump. Fifth Int. Symp. on Applications of Laser Techniques to Fluid Mechanics, July, Lisbon.

Nouri, J.M., D. Guerrato, N. Stosic, A. Kovacevic, and C. Arcoumanis. 2006. Cycle resolved velocity measurements within a screw compressor. 18th Int. Compressor Eng. Conf., Purdue, paper C011.

Schleer, M., and R. Abtahi. 2006. Clearance effects on the evolution of the flow in the vaneless diffuser of a centrifugal compressor at part load condition. Proc. ASME Turbo Expo 6(B):981-91.

Stosic, N. 1998. On gearing of helical screw compressor rotors. Proc IMechE 212(C):587-94.

Stosic, N., E. Mujic, A. Kovacevic, and I.K. Smith. 2006. The influence of discharge ports on rotor contact in screw compressor. 18th Int. Compressor Eng. Conf., Purdue, paper C010.

Zaidi, S.H., and R.L. Elder. 1993. Investigation of flow in a radial turbine using laser anemometry. Paper 93-GT-55, ASME Int. Gas Turbine and Aeroengine Congress and Exhibition, May, Cincinnati, Ohio.

Received February 5, 2007; accepted October 19, 2007

Jamshid M. Nouri, PhD Diego Guerrato Nikola Stosic, Dr(Sci), CEng Constantine Arcoumanis, PhD, CEng

Jamshid M. Nouri is Assistant Dean, Diego Guerrato is a doctoral student, Nikola Stosic is Chair of Positive Displacement Compressor Technology, and Constantine Arcoumanis is Dean at the School of Engineering and Mathematical Sciences, City University, London, UK.