Experimental _ Numerical Analysis of Classical Bleed Slot System

Document Sample
Experimental _ Numerical Analysis of Classical Bleed Slot System Powered By Docstoc
					  Experimental and numerical analysis of a classical bleed slot
            system for a turbocharger compressor

Subenuka Sivagnanasundaram, Stephen Spence and Juliana Early
School of Mechanical and Aero Space Engineering, Queen’s University of Belfast, UK

Bahram Nikpour
Cummins Turbo Technologies, Huddersfield, UK


This paper presents an insight into the performance and inducer flow field of a
turbocharger centrifugal compressor incorporating a classical bleed slot system with
various slot positions. The bleed slot is popular widely used map width enhancement
method in the turbocharger industry. Even though the technique has been used for a
long time, further study is still required to gain a better understanding of the flow
structures inside the compressor stage due to the bleed slot. Therefore, this detailed
study of the bleed slot system has been performed through a careful CFD analysis
validated against experimental data. Three different positions of the bleed slot opening
on the impeller shroud have been analysed and discussed in terms of their impact on the
map width of the compressor. Discussion focuses on the slot flow variation throughout
the speedline, the inducer inlet swirl near the surge flow condition and its impact on map
width enhancement. The paper also demonstrates the capability of CFD analyses to
predict the surge point.


              Absolute flow velocity (m/s)
              Meridional velocity (m/s)
              Swirl velocity (m/s)
              Ratio of specific heats
              Choke flow rate (kg/s)
              Surge flow rate (kg/s)
              Total pressure (MPa)
              Static pressure (MPa)
              Total Temperature (K)
              Static Temperature (K)
              Tip speed (m/s)
              Relative flow velocity (m/s)
              Incidence angle (deg)
              Blade angle (deg)
              Flow angle (deg)
              Compressor efficiency
              Non-dimensional distance between the wall and the 1st node of the mesh
  CFD         Computational Fluid Dynamics
  CTT         Cummins Turbo Technologies
  GGI         General Grid Interface
  MW          Map Width
  MCA         Mass Circumferentially Averaged
  ND          Non-Dimensional
  PR          Total-to-Total Pressure Ratio
  SST         Shear Stress Transport
  QUB         Queen’s University Belfast


The turbocharger compressor designer’s focus is on achieving a wider map with higher
pressure ratio while simultaneously improving the efficiency characteristic to satisfy the
development of modern internal combustion engines. However, the compressor map is
limited by two flow phenomena at higher and lower flow rates, called choke and surge
flow conditions respectively. The choke flow condition is the mass flow rate at which the
flow velocity becomes sonic, whereas surge occurs due to the complete reversal of the
flow at lower flow rate and higher pressure ratio. Surge is a system phenomenon and it
is not acceptable to operate the compressor in this region. Therefore, it is challenging to
design a compressor to satisfy the requirement of higher pressure ratio as well as
increased map width.

             Figure 1 Effect of swirl component on incidence at inducer inlet

A common cause of surge is the occurrence of stalled flow due to excessive incidence at
the leading edge of the impeller blades, although it can also result from stall in the
diffuser. At the target operating condition the inducer inlet is usually designed in such a
way that flow enters the inducer with zero or close to zero incidence. However, at off-
design conditions the flow approaches the blade leading edge at an incidence angle
which can be positive or negative. The incidence angle i’ depicted in Figure 1, which is
associated with the extremes of low flow rate, is termed as a positive incidence angle. If
the incidence is large enough the flow separates from the suction side of the blade after
the leading edge, resulting in localized flow recirculation and reduced pressure ratio,
leading the system to surge. As illustrated in Figure 1, introducing a positive swirl
component to the inlet flow tends to reduce excessive positive incidence, i’’.
Consequently it reduces the possibility of flow separation on the suction side of the blade
and moves the onset of surge to a lower mass flow rate. Imparting a positive swirl
component in the inducer inlet, however, reduces the work input and therefore results in
a lower pressure ratio for a given impeller speed.

In an attempt to enhance the compressor map width and overcome the problems with
surge, a number of approaches have been considered in the past. The bleed slot casing
treatment is a popular method that has been widely investigated [1-4]. In the
turbocharger industry, a very common design practise is to adopt a classical bleed slot
system, which was first investigated by Fisher [1] in the late 1980s. The authors’
investigation showed that, with the inclusion of a classical bleed slot, there was a
significant improvement in surge flow since the inducer flow was stabilized considerably.
Recently, Xinqian et al [5] published an analysis of a non-symmetric bleed slot which
reduced the surge flow by as much as 10% compared to a symmetric bleed slot. This
investigation analysed the non-symmetric flow characteristics of the impeller due to the
presence of the downstream scroll. In addition, the work published by Yamaguchi [6]
revealed that an inclusion of cavity vanes improved the map width by 20% compared to

a bleed slot system with no cavity vanes where no penalty of the efficiency was found.
This was performed by introducing counter swirl at the inducer inlet with an appropriate
cavity vane vane design. Mohtar et al [7] analyzed the compressor performance with a
classical bleed slot system. In this analysis, the slot opening to the impeller was
completely replaced by a series of holes that injected the flow into/out of the impeller.
No noticeable improvement in the map width was achieved due to this design. However,
the author showed that the combination of the slot opening holes with the Super MWE
techniques improved the map of the compressor significantly. The Super MWE technique
was first investigated and published by Nikpour [2] and the addition of this to a classical
bleed slot produced a significant impact on the map width in terms of improving the
surge flow rate. The Super MWE technique allows better mixing of the recirculated slot
flow with the main inlet flow at surge by introducing it further upstream from the inducer
inlet. As a complete alternative to the classical bleed slot, reference [8] reports a state of
the art high pressure ratio compressor for a helicopter engine using a series of fixed
vanes at the inducer tip to facilitate inducer flow recirculation. As technology has
improved, the ability to provide cheap and durable variable geometry inlet guide vanes
and diffusers has also been investigated [9 and 10].

   In the present study of compressor map width enhancement and performance, CFD
analysis has been performed for a full stage turbocharger compressor model. The
investigation has been carried out for the baseline compressor model with different
locations of the bleed slot systems. The detailed analysis of two different slot locations
has been presented in this paper by comparison with the baseline slot. This shows the
impact on the compressor map width and performance, and the influence of the variation
of slot flow and incidence angle on the compressor performance. At or near surge, the
incidence angle variation has been analysed by computing the variation of swirl and axial
velocity components at the leading edge of the impeller blades due to the recirculating


   The centrifugal compressor stage used for this investigation as the baseline
configuration was a commercially available turbocharger unit produced by Cummins
Turbo Technologies (CTT). This compressor stage consisted of various components
including an inlet duct, shroud bleed slot with an annular cavity, noise baffle, impeller,
pinched vaneless diffuser and asymmetric scroll. The compressor is from a turbocharger
used on a heavy duty diesel engine of approximately 400 HP (298 kW) output. The
standard bleed slot, which has a width of 3 mm, was placed just after the inducer throat
of the impeller at an angle of 45°. The position of the baseline slot was developed
through a series of tests carried out at CTT (not published in open literature). The exact
geometry details of the impeller have not been divulged for reasons of commercial
confidentiality; however the impeller has an outer diameter of approximately 100 mm.
The inducer of the impeller, where the map width enhancement slot is positioned, has a
typical design and therefore it is expected that the findings of this research are relevant
to most similar centrifugal impeller designs.


In this study the numerical analysis has been performed for a full stage of the
compressor using ANSYS CFX and ICEM CFD software. A structured mesh has been
generated for the inlet duct, bleed slot with an annular cavity, impeller and diffuser
whereas the scroll has an unstructured mesh. For this analysis, the required mesh
quality has been achieved through a grid independence study which revealed that the
solutions were grid independent when the model consisted of around 10–12 million
elements in total. Figure 2 represents the schematic of the full stage flow passage. The
further details of the mesh generation including the quality of the mesh and the
computational domain set-up have been described in reference [11].


                                         GGI       Frozen rotor                 GGI

                                       Noise Baffle

                                                                   Frozen rotor

                                       Frozen rotor                       Impeller

                       Primary Inlet                    Diffuser
                           Duct          Annular Cavity

                Figure 2 Schematic of full stage compressor flow passage

For the CFD set-up the inlet boundary conditions have been defined as the total
atmospheric pressure and temperature. Two different types of outlet boundary conditions
have been used: the average static pressure from the choke point to the mid-map and
the mass flow rate from the mid-map to the surge point. The convergence criteria have
been set to a value of 1E-4 of RMS residuals. Since a steady state analysis has been
carried out it is usually difficult to obtain a converged solution at stall due to the
transient effects. Therefore, in this analysis, the blend factor or gradient relaxation has
been applied, which helps to obtain a converged solution at the stall condition. Even
though the simulation satisfied the set residual value, the results were only accepted
when the efficiency and inlet mass flow also became constant and the mass flow
imbalance at each component reached zero.

In this investigation, the surge flow has been defined as the point when the static
pressure begins to fall with decreasing mass flow (the gradient of the static pressure line
becomes positive). However, in some cases the total pressure may be falling with
decreasing mass flow even though the static pressure has not yet begun to fall.


Figure 3 represents the schematic of the turbocharger test facility that was available in
the Queen’s University of Belfast. On the turbine side of the turbocharger, the air and
propane gas was supplied to the combustion chamber and then the combusted hot air
was used to drive the turbine and so the compressor. On the compressor side, the
ambient air was drawn in to the compressor and discharged to the atmosphere through
the discharge line. On the air supply line before the combustion chamber and on the
discharge line after the compressor, two orifice plates were placed to measure the
turbine and compressor air flow rates. The mass flow rate has been calculated in
accordance with ISO 5167. In order to obtain more confidence in the measurements, a
number of pressure and temperature transducers were installed at the inlet and exit of
the compressor and the average value of these readings was obtained.                In this
experimental analysis, the choke flow has been attained by fully opening the discharge
valve and the surge flow has been determined through the operator’s observation of the
audible cyclic flow reversals in the test system at the surge point. The flexibility of the
test rig allowed refined control of the compressor flow rate near surge. The consistency
of the surge point measurement was ensured through the repeatability of the baseline

compressor test. The repeatability tests took place up to 4 times at two intermediate
speeds on different occasions and the measurement of the surge mass flow rate was
found to be repeatable within ±0.5%.

Figure 3 Schematic of turbocharger test facility available in the Turbomachinery Laboratory at QUB

         Inlet section                  Slot opening


                   Figure 4 Compressor housing of the turbocharger unit tested

Since this study focuses on an investigation of various bleed slot configurations the
standard housing of the compressor unit was split after the slot opening as shown in
Figure 4 and the modification was only made to the front part of the housing, referred to
as an inlet section. The modified inlet section was assembled with the rear part of the
split housing. The purpose of this split was to avoid a repeat manufacturing of the scroll
section of the housing. A bell-mouth was also designed, manufactured and assembled
with the housing in order to provide a uniform flow at the inducer inlet, to align
effectively with the boundary conditions in the CFD model. The complete assembly of the
compressor housing and bell-mouth has also been shown in Figure 4.

There was one difference between the compressor housing geometry used in this study
and that for a typical turbocharger application. A turbocharger compressor housing
normally includes a series of supporting struts in the bleed slot cavity to physically
support the portion of the shroud that is upstream of the bleed slot. However, since
these struts could influence the flow and in some way act like vanes in the bleed slot
cavity, this investigation sought to remove their influence. Because the front portion of
the shroud still needed to be mechanically supported in the absence of the struts, a
series of 6 small round pins were incorporated in the experimental hardware to support
the shroud while causing minimum disturbance to the flow. The pins had a diameter of 3
mm which was small by comparison with the dimensions of the struts cast into the
original commercial turbocharger housing. The pins were not included in the CFD model.


For this study, the baseline compressor was tested at four different speeds (48%, 68%,
87% and 100%) using the experimental setup described above. However, the CFD
analysis was only performed for three higher speeds. The pressure ratio characteristics
and the efficiency of the baseline compressor model from both the CFD analysis and the
experiments are shown in Figure 5 and Figure 6. The mass flow rate and efficiency are
non-dimensionalised by considering the highest stage mass flow and efficiency from the
experimental data as 100%. As illustrated, the pressure ratio obtained from the CFD
results is in good agreement with the experimental results at all three speeds.
Importantly, it should be noticed that the surge point is predicted effectively.
                                        Baseline Exp
                                        Baseline CFD
      Pressure ratio (t-t)





                                    0   20             40       60             80    100       120
                                                            ND Mass flow (%)

 Figure 5 Pressure ratio – mass flow characteristics of the baseline compressor with baseline slot

The compressor total-to-total isentropic efficiency, (                   ), has been calculated using the
following conventional equation:

where       and      are the total temperatures at the inlet and exit of the experimental
compressor and PR is the pressure ratio (total-to-total). In general, the efficiency
prediction using CFD analysis has good agreement with the test data at the choke side of
the map, however, there is more discrepancy towards the surge condition at the lower
speed. Yet, the prediction has a similar trend to the test results and shows a good
representation of the efficiency characteristic.


     ND Efficiency (t-t)



                                                                          Baseline Exp
                                                                          Baseline CFD
                                                               0            20           40           60           80         100          120
                                                                                                  ND mass flow (%)
                                                                    Figure 6 Efficiency of the baseline compressor with baseline slot

From the CFD results, the predicted slot flow of the baseline model throughout the speed
lines has also been analysed and illustrated in Figure 7. The slot flow rate is given in
terms of a percentage of the overall stage flow rate where the stage flow rates are non-
dimensionalised using the maximum stage flow rate from the experiment as 100%. On
the plot, the positive Y-axis values represent inflow at/near choke conditions and the
negative Y-axis values represent the recirculating flow at/near surge conditions. The
study shows that the slot inflow at choke conditions increases with increasing speeds.
Therefore, it is expected to have an efficiency penalty due to higher mixing of the slot
inflow with the impeller main inlet flow, and higher frictional losses in the bleed passage
due to an increased flow rate and velocity. On the other hand, at/near surge flow, as the
speed increases the amount of the recirculating flow decreases so that the loss
occurrence would be opposite to the choke flow conditions.
                     ND slot flow (%of Stage mass flow)

                                                                    0            20        40            60          80        100         120


                                                                                                                              100% speed
                                                           -40                                                                87% speed
                                                                                                                              67% speed
                                                                                                ND Stage mass flow (%)
                                                                        Figure 7 Slot flow variation against the main inlet flow

Having established confidence through validation of the baseline CFD model with
experimental measurements, further investigations have been carried out by modifying
the baseline bleed slot system. The position of the slot was moved towards and away
from the main blade leading edge and is referred to as MWE1P and MWE2P respectively
where the baseline model is called MWEB. This study was mainly performed to gain a
detailed understanding of the flow field and its impact on compressor performance due
to various slot positions of a classical bleed slot system. An analysis for the same
compressor model with various slot widths has been already published in reference [12].




      Pressure ratio (t-t)







                                      0   20   40        60            80   100         120
                                                    ND Mass flow (%)
 Figure 8 Pressure ratio – mass flow characteristics of the compressor by moving the slot position
                              towards the main blade leading edge

Figure 8 and Figure 9 depicts the comparison of the pressure ratio – mass flow
characteristics and efficiency predictions between MWEB and MWE1P. In this case, the
CFD analysis was performed for 87% and 100% speeds since the experimental data
showed that the slot modifications only have an impact on these speeds.

When varying slot position, the shroud static pressure would be the major factor that
determines the performance of the bleed slot system. While the static pressure along the
impeller shroud generally increases going from the inlet to exit, there is an initial drop in
static pressure between the leading edge and the throat close to choke. The difference in
static pressure between the inlet of the impeller and the slot opening on the impeller
shroud is the driving force of the slot inflow at/near choke and recirculation at/near
surge flow conditions. Therefore, as the slot position is moved towards the leading edge
it is expected that the surge flow would be higher than the baseline case, which is
confirmed by the experimental measurements at both speeds. However, at 87% speed,
the CFD predicted the surge flow rate to be the same as the baseline model. This
suggests that the CFD does not have the capability of accurately predicting small
changes in the surge flow. It is interesting to note that at 100% speed both the
experiments and CFD show lower pressure ratio compared to the baseline case.


          ND Efficiency (t-t)






                                      0   20   40          60          80   100         120
                                                    ND Mass flow (%)

  Figure 9 Efficiency prediction of the compressor by moving the slot position towards the main
                                         blade leading edge


      ND slot flow (%of main inlet flow)
                                                 0      20        40              60            80     100     120

                                                                                         100% speed

                                                                           87% speed

                                                               68% speed
                                                                           ND Main inlet flow (%)

                                                                Figure 10 Slot flow variation

When the slot position is shifted towards the main blade leading edge the shroud static
pressure at the slot opening to the impeller would be lower than the baseline line model
so that, at the choke flow condition, the pressure difference between the impeller main
inlet (usually atmospheric condition) and the slot opening would be higher which drives a
higher slot inflow in this case. On the other hand, at surge, this modified configuration
would allow a lower amount of recirculation due to smaller driving pressure force (static
pressure difference between the slot opening and the impeller main inlet – the pressure
at the slot opening would be higher than the impeller inlet condition). This has been
clearly predicted in Figure 10.



         Pressure ratio (t-t)







                                                  0.0   20.0      40.0            60.0       80.0     100.0   120.0
                                                                         ND Mass flow (%)
Figure 11 Pressure ratio – mass flow characteristics of the compressor with the slot position moved
                             away from the main blade leading edge

In the next step of this study, another configuration was obtained by moving the of the
baseline slot further away from the main blade leading edge and similar analysis has
been performed as for the previous model. The pressure ratio and efficiency
characteristics for the MWE2P configuration are shown in Figures 11 and 12. At the
100% speed, the CFD and experiment showed similar performance. At higher flow rates,
moving the slot position further downstream would affect the slot inflow due to higher
shroud static pressure. This was clearly demonstrated with experimental data through
reduced choking mass flow rate; however, the CFD did not identify this phenomenon. At
87% speed the CFD analysis predicted no improvement in terms of reducing the surge
flow rate. However, a small deterioration was apparent in the experimental
measurements. This study showed that moving the slot position further away from the
leading edge of the blade only improved the surge flow rate at the highest speed,
although with slightly lower choke mass flow rate.



       NDEfficiency (t-t)






                                                 0.0   20.0        40.0         60.0           80.0        100.0   120.0
                                                                            ND Mass flow (%)

  Figure 12 Efficiency – mass flow characteristics of the compressor with the slot position moved
                             away from the main blade leading edge

The experimental data as illustrated in Figure 12 shows that this model can have a
higher efficiency at lower speeds (48% and 67%) on the choke side of the map. This is
due to the lower slot inflow (Figure 13) which leads to a reduced loss and downstream
blockage formed due to the mixing of the slot inflow with the main inlet flow. At the
highest speed, there was a notable efficiency drop near the choke mass flow condition
and also a reduction in the maximum flow.         As shown in Figure 13 the MWE2P
configuration at both speeds, 87% and 100%, had a higher slot flow recirculation
at/near surge flow. This means that the inducer flow would be higher than the baseline
case for a particular stage flow. Therefore, this increased inducer flow, especially at
100% speed, and stabilized the inducer which brought an improvement in surge flow. An
increased inducer flow is more effective in terms of reducing the surge flow at higher
speed. For example, at 87% speed, this modified configuration recirculates about 3%
higher flow rate than a baseline model but no surge flow improvement has been
achieved. The results indicate that the modifications to the slot position only have an
impact at higher speeds.

      ND slot flow (%of stage mass flow)

                                                  0     20          40             60           80          100     120


                                                                                        100% speed

                                                                          87% speed
                                                                68% speed
                                                                          ND Stage mass flow (%)
                                                        Figure 13 Slot flow variation against stage flow

Figure 14 Velocity contour with super-imposed velocity vectors in the bleed slot passage (MWEB at

Figure 15 Velocity contour with super-imposed velocity vectors in the bleed slot passage (MWE2P
                                           at choke)

Figure 16 Velocity contour with super-imposed velocity vectors in the bleed slot passage (MWEB at

 Figure 17 Velocity contour with super-imposed velocity vectors in the bleed slot passage (MWE2P
                                            at surge)

Two localized recirculation regions (Region 1 and Region 2) were found. Region 2
appeared to be similar in all three configurations. For example, these flow characteristics
of the bleed slot system are shown for MWEB and MWE2P in Figure 14 and Figure 15 at
the choke flow condition. These two localized recirculation regions contributed to losses,
producing a negative impact on compressor overall performance; these flow structures
appeared unavoidable with the current design of bleed slot. Importantly, an interesting
flow characteristic was found with the MWE2P configuration. Thus, when the slot position
is moved further downstream the part of the slot opening on the shroud experiences
negative pressure differential and, as a consequence, the near-wall flow along the
shroud enters into the slot passage through this part of the slot opening. This is
illustrated in Figure 15.This flow characteristic creates blockage in the slot opening and
so the slot inflow at choke is reduced (Figure 13).

On the other hand, at the surge flow condition, Region 2 exists in all three cases, but the
effect is more severe in the case of MWE2P.Examples are illustrated for MWEB and
MWE2P in Figure 16 and Figure 17 at the surge condition. Thus, Region 2 has actually
affected the impeller passage as well as the slot passage. Even though the localized
recirculation, Region 2, still exists within the slot and/or impeller passages, shifting the
slot position away from the main blade leading edge continuously increases the bleed

flow out through the slot (Figure 13). For example, the MWE2P configuration recirculates
5 – 7% more flow through the slot than is achieved with the baseline configuration. The
effective slot recirculation is achieved as a result of higher shroud static pressure at the
slot opening as the slot position is moved further downstream from the leading edge. As
already discussed, an increase in slot recirculation increases the inducer flow and so
stabilizes the inducer at reduced stage flow rates. The flow field study of the bleed slot
system with various slot location emphasizes the importance of an optimized position of
the bleed slot in order to obtain the widest possible map.

                           60                                          MWE2P
       Map width (%)

                           50                                                  MWEB
                                       50000     70000                90000          103500
               Figure 18 Map width calculated from measured data due to different slot positions

The map width obtained from the measured data for each speed line for all three slot
positions is compared in Figure 18. In general, no considerable map width improvement
has been measured at 48%, 68% and 87% speeds. However, at the highest speed, the
baseline is better than the MWE1P configuration. The MWE2P configuration shows similar
map width as the baseline even though the surge flow condition was significantly
improved. The reason is that there was a similar reduction in choke flow at this speed,
which offset any improvement at the surge condition.



       Incidnence (deg)





                                 0.0       0.2     0.4                0.6      0.8            1.0
                                                  Hub-to-shroud span

   Figure 19 Incidence angle variations at choke from hub-to-shroud at the maximum
                                speed of the compressor
In general, the formation of higher incidence angles is one of the major factors for the
stall initiation in the compressor inducer. Therefore, an analysis of incidence angle
distribution at both choke and surge conditions, and the variation of inlet swirl that was
carried through with the recirculating flow has been performed for the three models

(baseline bleed slot and two different slot positions) These distributions are shown in
Figure 19, Figure 20 and Figure 21 for the maximum speed of the compressor since
varying the slot position had most impact at this speed. The incidence angle has been
computed using the circumferentially mass flow averaged velocity components just
upstream of the main blade leading edge from the CFD analysis.

First of all, at the choke flow condition as shown in Figure 19, the model with the slot
position closer to the leading edge (MWE1P) predicted similar incidence angle variations
to the baseline model. This is because, as illustrated in Figure 10, these two models
predicted similar slot inflow and stage flow so that the inducer flow and incidence has to
be the same in both cases. However, the model with the slot position further
downstream (MWE2P) shows reduced incidence angle over the outer 80% of the span
compared to the baseline slot. A reduced inflow through the slot due to higher shroud
static pressure meant that the inducer passed a higher amount of flow. As a result of
this, incidence is reduced due to an increased axial velocity component. At this point, it
should be noticed that there was not good agreement between the predicted and
measured choke flow rates for the MWE2P configuration at the highest speed and
therefore the predicted incidence distribution in Figure 19 may not be a good
representation of the actual inducer flow for this particular configuration.



      Inlet swirl (m/s)






                                 0.0   0.2    0.4             0.6          0.8     1.0
                                             Hub-to-shroud span

Figure 20 Inlet swirl variation at surge from hub-to-shroud at the maximum speed of the


                          12.0                           Reduced incidence with
                                                         increased recirculation
      Incidnence (deg)






                                 0.0   0.2    0.4             0.6          0.8     1.0
                                              Hub-to-shroud span

Figure 21 Incidence angle variations at surge from Hub-to-shroud span at the maximum
                                speed of the compressor

As is evident in Figure 11, the MWE2P slot position clearly has a beneficial effect on
extending the surge point to lower flow rates. The predicted distribution of inlet swirl
that was introduced by the slot recirculation flow at surge is plotted in Figure 20. As the
slot positions were moved further downstream of the leading edge an increased inlet
swirl was produced. Two main reasons are responsible for this increased swirl:
    1. An increased amount of recirculated flow
    2. An increased amount of work input to the flow that is recirculated through the
        bleed passage
As a consequence of an increased swirl and an increased axial velocity component, a
significantly reduced incidence angle was achieved at the outer 50% of the span with the
furthest downstream slot position (MWE2P) compared to other two models. This model
also predicted a negative incidence between 85 and 95% span due to the concentration
of swirling flow in this region (Figure 22). From both Figure 20 and Figure 21, it is
apparent that the influence of the recirculated swirling flow can extend to about 50% of
the outer span by moving the slot position further downstream (MWE2P) whereas the
other two models have an impact on a smaller proportion of the hub-to-shroud span.

      Hub-to-shroud span


                           0.4                                     concentration of
                                                                  recirculating flow
                           0.2                                      due to MWE2P

                                 50     70       90       110        130        150    170        190   210
                                                                Axial velocity (m/s)
                                      Figure 22 Axial velocity variations at surge (100% speed)


The paper presents an experimental and numerical analysis of a classical bleed slot
system for a turbocharger centrifugal compressor. The purpose of this study was to gain
an insight into the inducer flow field of the classical bleed slot system with various slot
positions. The study was performed for three different slot positions where a baseline
slot was modified in order to achieve two other models: one model was with the slot
position closer to the main blade leading edge and the other one was with the slot
position further downstream. This analysis gives a better understanding of the impact of
the bleed slot position on map width improvement or deterioration and the compressor
performance. For most of the performance range very good agreement was achieved
between the CFD and the experimental results, with particularly good prediction of the
surge flow rate. Use of the CFD gave a better understanding of the behaviour of the slot
flow and how it impacted the inducer flow field near the surge condition. In particular,
moving the slot position further downstream increased the amount of flow recirculation
which was shown to improve leading edge incidence over the outer half of the blade
span and produced a beneficial effect on the surge mass flow rate.


The authors are grateful to ANSYS Inc for access to and technical support with the CFD
and grid generation software used for this research study. The authors also would like to

express their gratitude to Cummins Turbo Technologies for supporting this project and
providing extensive help.


1.    Fisher, F.B., (1988), “Application of Map Width Enhancement Devices to
      Turbocharger Compressor Stage”, Society of Automotive Engineers, Inc.
2.    Nikpour, B., (2004) “Turbocharger compressor flow range improvement for future
      heavy duty diesel engines”, THIESEL 2004 Conference on Thermo- and Fluid
      Dynamic Processes in Diesel Engines.
3.    Hunziker, R., Dickmann, H P., and Emmrich, R., (2001), “Numerical and
      experimental investigation of a centrifugal compressor with an inducer casing bleed
      system”, IMechE Pro Instn Mech Engrs, Vol 215 Part A.
4.    Ishida, M., Sakaguchi, D., and Ueki, H., (2006) “Effect of pre-whirl on unstable flow
      suppression in a centrifugal impeller with ring groove arrangement”, ASME Turbo
      Expo 2006, GT2006-90400.
5.    Xinqian, Z., Yangjun, Z., Mingyang, Y., Takahiro, B., and Hideaki, T., (2010),
      “Stability improvement of high-pressure-ratio turbocharger centrifugal compressor
      by asymmetric flow control – Part 2: Non-axisymmetric self recirculation casing
      treatment”, ASME Turbo Expo 2010, GT2010-22582.
6.    Yamaguchi, S., Yamaguchi, H., Goto, S., and Nakamura, F., (2002), “The
      development of effective casing treatment for turbocharger compressor”, IMechE
      Turbocharger and Turbocharging.
7.    Mohtar, H., Chesse, P. and Chalet, D., 2010, “Effect of a map width enhancement
      system on turbocharger centrifugal compressor performance and surge margin”,
      Proc. IMechE Vol. 225 Part D: J. Automobile Engineering
8.    Barton, M. T., Mansour, M. L., Liu J. S., and Palmer D.L, (2006), "Numerical
      Optimization of a Vaned Shroud Design for Increased Operability Margin in Modern
      Centrifugal Compressors", Journal of Turbomachinery, ASME, Vol 128, pp. 627-631.
9.    Kim, Y., Engeda, A., Aungier, R., and Amineni, N., (2002), "A compressor stage with
      wide flow range vaned diffusers and different inlet configurations", IMechE, Vol 216,
      Part A.
10.   Engeda, A., (2001), "The unsteady performance of a centrifugal compressor with
      different diffusers", Proceedings of the Institution of Mechanical Engineers, Journal
      of Power and Energy, Part A, v 215, n 5, 2001, p 585-599.
11.   Coppinger, M., and Swain, E., (2000), "Performance prediction of an industrial
      centrifugal compressor inlet guide vane system", IMechE Paper, Vol 214 Part A.
12.   Sivagnanasundaram, S., Spence, S., Early, J., and Nikpour, B., (2011), “An impact
      of various shroud bleed slot configurations and cavity vanes on compressor map
      width and the inducer flow field”, ASME Turbo Expo 2011, GT2011-22154.
13.   Japikse, D., 1996, “Centrifugal compressor design and performance”, Concepts
      NREC, ISBN 0-933283-03-2.


Shared By: