Vous êtes sur la page 1sur 16

DieselNet Technology Guide � Engine Intake Charge Management

DieselNet.com. Copyright � Ecopoint Inc. Revision 2017.11


Turbocharger Fundamentals
Hannu J��skel�inen, Magdi K. Khair

Turbocharging Challenges
Fixed Geometry Turbochargers
Variable Geometry Turbochargers
Compressor Map Width Enhancement
Multiple Compressors
Assisted Turbocharging
Boosting Systems
Turbocharger Durability and Materials
Turbocharger Bearings
Abstract: Turbochargers are centrifugal compressors driven by an exhaust gas
turbine and employed in engines to boost the charge air pressure. Turbocharger
performance influences all important engine parameters, such as fuel economy,
power, and emissions. It is important to understand a number of fundamental
concepts before moving on to a more detailed discussion of turbocharger specifics.
Turbocharger Construction
Turbocharger Compressor
Basic Principles of Compression Process
Compressor Maps
Turbocharger Turbine
Turbine Energy Extraction
Turbine Performance
Turbocharger Construction

A turbocharger consists of a compressor wheel and exhaust gas turbine wheel coupled
together by a solid shaft and that is used to boost the intake air pressure of an
internal combustion engine. The exhaust gas turbine extracts energy from the
exhaust gas and uses it to drive the compressor and overcome friction. In most
automotive-type applications, both the compressor and turbine wheel are of the
radial flow type. Some applications, such as medium- and low- speed diesel engines,
can use an axial flow turbine wheel instead of a radial flow turbine. The flow of
gases through a typical turbocharger with radial flow compressor and turbine wheels
is shown in Figure 1 [Schwitzer 1991].

[schematic]
Figure 1. Turbocharger construction and flow of gases
(Source: Schwitzer)
Center-Housing. The turbine-compressor common shaft is supported by a bearing
system in the center housing (bearing housing) located between the compressor and
turbine (Figure 2). The shaft wheel assembly (SWA) refers to the shaft with the
compressor and turbine wheels attached, i.e., the rotating assembly. The center
housing rotating assembly (CHRA) refers to SWA installed in the center-housing but
without the compressor and turbine housings. The center housing is commonly cast
from gray cast iron but aluminum can also be used in some applications. Seals help
keep oil from passing through to the compressor and turbine. Turbochargers for high
exhaust gas temperature applications, such a spark ignition engines, can also
incorporate cooling passages in the center housing.

[photo]
Figure 2. Sectional view of turbocharger
Sectional view of an exhaust gas turbocharger for a gasoline engine showing
compressor wheel (left) and turbine wheel (right). The bearing system consists of a
thrust bearing and two fully floating journal bearings. Note the cooling passages.
(Source: BorgWarner)
Turbocharger Bearings
Bearings. The turbocharger bearing system appears simple in design but it plays a
key role in a number of critical functions. Some of the more important ones
include: the control of radial and axial motion of the shaft and wheels and the
minimization of friction losses in the bearing system. Bearing systems have
received considerable attention because of their influence on turbocharger friction
and its impact on engine fuel efficiency.

With the exception of some large turbochargers for low-speed engines, the bearings
that support the shaft are usually located between the wheels in an overhung
position. This flexible rotor design ensures that the turbocharger will operate
above its first, and possibly second, critical speeds and can therefore be subject
to rotor dynamic conditions such as whirl and synchronous vibration.

Seals. Seals are located at both ends of the bearing housing. These seals represent
a difficult design problem due to the need to keep frictional losses low, the
relatively large movements of the shaft due to bearing clearance and adverse
pressure gradients under some conditions.

These seals primarily serve to keep intake air and exhaust gas out of the center
housing. The pressures in the intake and exhaust systems are normally higher than
in the turbocharger�s center housing which is typically at the pressure of the
engine crankcase. As such, they would primarily be designed to seal the center
housing when the pressure in the center housing is lower than in the intake and
exhaust systems. These seals are not intended to be the primary means of preventing
oil from escaping from the center housing into the exhaust and air systems. Oil is
usually prevented from contacting these seals by other means such as oil deflectors
and rotating flingers.

Turbocharger seals are different from the soft lip seals normally found in rotating
equipment operating at much lower speeds and temperatures. The piston ring type
seal is one type that is often used. It consists of a metal ring, similar in
appearance to a piston ring. The seal remains stationary when the shaft rotates.
Labyrinth-type seals are another type sometimes used. Generally turbocharger shaft
seals will not prevent oil leakage if the pressure differential reverses such that
the pressure in the center housing is higher than in the intake or exhaust systems.

Turbocharger Compressor

A radial flow compressor stage is composed of two sections, the impeller or �wheel�
and the cover or �housing�, Figure 3. Filtered air enters through the center of the
compressor cover and proceeds through the inducer into the inducer (1) of the
compressor wheel. As air proceeds through the compressor wheel it makes a 90� turn
thus changing its flow from an axial to a radial direction. Air exits the
compressor wheel at the exducer (2), enters a narrow stationary diffuser and then
passes through to a volute or scroll from which it is discharged from the
compressor.

Figure 3. Cross section of radial flow compressor


Compressor Wheel. The critical components of the compressor wheel are the blades.
These blades have three regions:

the leading edge is a sharp pitch helix designed for scooping air in and moving it
axially;
blades are curved to change the direction of the airflow from axial to radial and
at the same time to accelerate it to a high velocity;
blades terminate in a trailing edge which is designed to propel air radially out of
the compressor wheel�defined by pitch, angle offset from radial and/or back taper
or back sweep.
The shape of the compressor blade profile from the impeller inlet to the outlet can
significantly impact performance and durability. Generating an appropriate shape
can be a compromise between manufacturing cost and compressor performance. A
significant development in manufacturing of compressor impellers was the extension
of flank milling (milling with the side of the cutter) to the production of
arbitrary surfaces [Wu 1986]. Previously, flank milling was limited to the
production of ruled surfaces�surfaces that were defined by a limited number of
curves and arbitrary surfaces required the use of point milling (milling with the
tip of the cutter)�a very time consuming and costly process for volume production.
Arbitrary surfaces allow the blade profile to be tailored more precisely to the
demands of the flow to provide better performance and durability rather than the
need to limit the surface to a few well defined mathematical curves as ruled
surfaces do [Holset 2004].

The compressor wheel blade trailing edge or blade tip can either extend radially
from the center of the wheel�s hub (as do the two impellers in Figure 9 below) or
have a backward curvature, Figure 8. Forward curved blades are also possible but
are rarely used in turbochargers. Radial impeller blades were very common in the
past due to ease of manufacture and strength considerations�the blade tip was
relatively unaffected by centrifugal forces resulting from high rotational speeds.
However, efficiency is lower with radial blades and modern turbochargers will
typically use a backward curved impeller to maximize compressor efficiency.
However, backward curvature can be limited by manufacturing considerations when
impellers are produced by casting and by strength considerations. The backward
curvature means that blade tips can be subjected to significant bending forces at
high rotational speeds with the bending stress increasing with backward sweep
angle.

Selection of blade tip or trailing edge angle is a trade-off between performance


and manufacturing considerations. Figure 4 shows some of the performance trade-
offs. The backward curved blade has the highest efficiency but tends to provide the
lowest pressure ratio for a given tip speed that drops quickly [Hanlon 2001].
Backward curved impeller blades require higher tip speeds compared to radial blades
to maintain a given pressure ratio.

Figure 4. Effect of blade tip angle on compressor performance


Turbocharger compressors often benefit from a higher number of blades to provide
good flow guidance along the impeller. However, if full length blades are used, the
optimum number of blades can lead to blockage of flow at the impeller inlet. To
avoid this problem, splitter blades that start part way through the impeller can be
used, Figure 5.

[photo]
Figure 5. Compressor impeller with full length and splitter blades
(Source: BorgWarner)
Blades have extremely close tolerances with the housing to minimize backflow.
Production of cast impellers and housings to tight tolerances can be a significant
challenge. In some cases, the stationary housing can be coated with an abradable
coating that is softer than the blade material. Upon initial start-up, the coating
is worn away enough so that the impeller blade can pass freely and clearances are
kept to a minimum. This can provide a significant benefit in turbocharger
efficiency as evidenced by the improvement in engine fuel economy below 1300 rpm
shown in Figure 6 and by lower transient particulate emissions (not shown) [Sharp
1999].

[chart]
Figure 6. Influence of turbocharger abradable coating on fuel economy
Compressor wheel attachment to the shaft can be via a through-hole (through-bore
wheel) or a blind hole (boreless wheel). In boreless designs, attachment of the
wheel to the shaft can be via a threaded connection or welding. In thrubore
designs, a nut at the end of the shaft is commonly used.

[schematic]
Figure 7. Thrubore and boreless compressor wheels
Boreless compressor wheels can eliminate the area of high stress at the interface
between the wheel and shaft bore at the axial location where wheel diameter is
largest. However, the axial length of the wheel increases due to the need to attach
the shaft to the wheel while avoiding shaft protrusion beyond the z-plane (Figure
7). This can increase the footprint of the turbocharger and, because of the longer
distance from the bearing support to the wheel�s center of mass, make the rotor
dynamic effects more challenging. Also, alignment of the shaft and impeller as well
as rotating assembly balancing can become more difficult and expensive.

Diffuser. The diffuser can be vaneless as shown in Figure 3 or vaned, Figure 8


[Heywood 1988]. Early compressor diffusers used a series of divergent nozzles to
convert kinetic energy to pressure via volume expansion.

[schematic]
Figure 8. Segment of a radial compressor and vaned diffuser plate
Vanes shorten the flow path through the diffuser, reducing frictional losses and
controlling the radial velocity component of the gas. Due to lower friction, head
and efficiency are enhanced. However, off-design operation rapidly changes the
incidence angle to the vanes and flow separation occurs, resulting in a reduced
operating range [Hanlon 2001]. Vaned diffusers are common in marine and generator
set applications where a high pressure ratio is often required and a wide flow
range is unnecessary.

Vaneless diffusers are used when some efficiency can be sacrificed to reduce
manufacturing costs, when size is not a criterion, or if a wide range of flow rates
at a given pressure ratio is required. Vaneless diffusers are used in most
automotive type applications. A vaneless diffuser provides a significantly lower-
pressure recovery compared with a vaned diffuser of the same diameter. The diffuser
walls can also be designed with a nonlinear area increase with radial distance. In
any design it is important to ensure smooth internal surfaces in the diffuser to
reduce frictional losses [Siuru 2003].

Volute. A spiral-shaped volute or scroll collects the flow from the diffuser and
passes it to the compressor outlet. Figure 9 shows two volutes and their
corresponding compressor wheels. The cross-sectional area increase of volute from
a, to b, to c helps convert kinetic energy into potential energy, or air velocity
into pressure [Schwitzer 1991]. Together with increased pressure, the compressed
air experiences an increase in its temperature�to over 200�C in some cases�and
therefore a reduction in its density. An intercooler or aftercooler can be used to
cool the compressed air and increase its density before it enters the engine. Note
the different connection to the piping at the exit from the volutes.

[photo]
Figure 9. Two volute cover designs for radial compressor
An important objective of volute design is to achieve a uniform flow at the volute
exit. This is usually attained at the design flow-rate only, so that at off-design
conditions the volute is either too small or too large and a pressure distortion
develops circumferentially around the volute. At low flow-rates the pressure
increases with azimuth angle, while at high flow-rates the pressure decreases.
These circumferential pressure distortions can be transmitted back to the impeller
discharge and even as far back as the impeller inlet. The pressure distortions
reduce the compressor�s performance and have a direct impact on diffuser and
impeller flow stability [Carter 2009].
Basic Principles of Compression Process

Before discussing various performance aspects of turbocharger compressors, it is


important to understand a few basic principles.

The ideal compression process can be considered to be isentropic, Figure 10. For an
ideal gas, a good approximation for a turbocharger compressor, the change in
enthalpy is a function only of temperature so this Figure 14lso reflects the
enthalpy change. Due to various inefficiencies, the actual process consumes more
energy than the ideal.

[chart]
Figure 10. Ideal and actual compression process
Isentropic efficiency of a compressor can be defined as

?isentropic = hisentropic/(hin-hout) (1)

Inefficiency will show up as a higher than ideal compressor outlet temperature for
a given pressure ratio, Figure 11. Turbocharger efficiency can have a direct impact
on the capacity of the intercooler with less efficient compressors requiring higher
intercooling capacity to maintain a given intake manifold temperature.

[chart]
Figure 11. Effect of compressor efficiency on outlet temperature
The energy behind the pressure increase in a centrifugal compressor is supplied by
the impeller via acceleration of the flow to very high velocities. Figure 12
clarifies various components of these velocities for backward, radial and forward
curved impeller blades [Golloch 2005].

[schematic]
Figure 12. Velocity triangles for backward, radial and forward curved impeller
blades
In Figure 12, C1 and C2 are the absolute velocity vectors (i.e., relative to the
stationary frame of the compressor) of the air entering and exiting the impeller.
U1 and U2 are the impeller blade tip speed vectors at these same locations. W1 and
W2 are velocity vectors of the air flow entering and leaving the impeller relative
to the impeller:

W = C - U (2)

Pressure generation inside the compressor consists of several steps: (1) kinetic
energy is first supplied to the air by means of the rotor which accelerates the air
to a high speed, (2) as the air passes between the blades of the impeller, the
cross-sectional flow area between the blades increases which causes some of the
kinetic energy imparted to the flow by the impeller to be converted into static
pressure within the impeller even while additional kinetic energy is being added,
(3) after the air exits the impeller, no more energy is transferred into the air
flow. It is now the role of the diffuser and volute to convert the remaining
kinetic energy into static pressure as efficiently as possible. This process is
further illustrated in Figure 13 where the subscripts 1, 2 and 2� correspond to the
impeller inlet, diffuser inlet and volute inlet as shown in Figure 3 [Merker 2012]
[Hanlon 2001].

[schematic]
Figure 13. Centrifugal compression process in coordinates of specific enthalpy (h)
and specific entropy (s)
The total enthalpy change through the compressor can be described by the Euler
equation:
?ht = 1/2 [(C22 - C12) + (-W22 + W12) + (U22 - U12)] (3)

It can be shown that using Equation (2), the pressure increase across the impeller
can be written as

P2 - P1 = [G/2gc] � (C2TU2 - C1TU1) (4)

where C2T and C1T are the tangential components of C2 and C1, Figure 12. Figure 14
shows how pressure ratio can vary with impeller tip speed, U2 [NATO 2007]:

[chart]
Figure 14. Effect of impeller tip speed on pressure ratio
A parameter that is common in fan engineering and occasionally used for
turbocharger compressors [Engels 2002] is the dimensionless pressure number, ?. One
way to write this is:

? = 2?ht/U22 (5)

The value of the pressure number relates the total enthalpy change in the
compressor to the tip speed of the compressor impeller. Different values of ?
reflect different compressor design philosophies and a trade-off can exist
between ? and turbocharger efficiency.

Compressor Maps

A common tool used in selecting a turbocharger compressor for a given application


is a compressor map. A compressor map provides the operating envelope of the
compressor over all stable operating points. For turbocharger compressors, the x-
axis is normally in terms of flow while the y-axis is in terms of pressure. In
order to construct such a map, it is necessary to first determine the pressure
characteristics of the compressor over a range of compressor speeds and flows.
These characteristics are then plotted on a single graph to create the compressor
map.

As shown in Equation (4), the pressure developed by a radial flow centrifugal


compressor is proportional to the product of C2T and U2. Using this relationship,
Equation 1 and Figure 12, one can deduce the shape of the ideal flow versus
pressure curves for various blade tip angles. In an impeller with radial blade
tips, as flow decreases (i.e., W2 decreases) at a constant impeller speed, the
magnitude of C2 decreases and C2T is unaffected making the ideal pressure ratio
versus flow characteristic flat, Figure 15b. With backward curved blades, the
magnitude of C2 increases and C2T also increases making the ideal pressure ratio
highest at low flows Figure 15c [Hanlon 2001].

Figure 15. General shape of pressure ratio versus flow characteristics


(a) forward, (b) radial, (c) backward curved impeller blades
In addition to pressure ratio, flow range is another important consideration that
needs to be taken into consideration when generating a compressor map. The lowest
stable flow possible from a centrifugal compressor at a particular rotational speed
is determined by the surge limit while the highest flow by choking.

As flow is reduced at a constant impeller speed, W2 decreases proportionately,


causing the flow angle a2 to decrease. In a vaneless diffuser, this creates a
longer flow path through the diffuser which increases frictional losses to the
diffuser walls. Also, on the inlet side of the impeller, W1 decrease causing the
flow angle a1 to decrease. At low enough flows, the low angle of a1 can cause flow
separation to occur on the low pressure side of the blade at the impeller inlet
that can further reduce flow into the impeller. This combined with the high
frictional losses in the diffuser can lead to a situation where the pressure
generation in the diffuser falls below the delivery pressure and the flow in the
compressor can reverse. This reverse flow continues until the resistance through
diffuser drops enough for air discharge to be continued. This process is repeated
in a cyclic fashion and is referred to as surge. The point on the pressure versus
flow curve at a given impeller speed where surge occurs is referred to as the surge
limit. The surge limit can also be thought of as the maximum of the pressure versus
flow curve at a given impeller speed. Prolonged operation during compressor surge
can damage the compressor.

Vanes in the diffuser can shorten the flow path through the diffuser to reduce
frictional losses and control the radial velocity component with the net result
that pressure generation and efficiency are improved. However, off-design operation
changes the incidence angle to the vanes and flow separation can occur, leading to
a reduced operating range for the compressor.

At the other end of the pressure versus flow curve, as flow increases, a point can
be reached where W becomes too high for the flow to remain attached to the impeller
blades and separation can occur within the passages between the impeller�s blades.
This separation reduces the effective flow area and eventually a point is reached
where the Mach number reaches 1 and flow through the impeller becomes choked. The
point on the pressure versus flow curve at a given impeller speed where choking
occurs is referred to as the choke limit. In practice, it is common to define the
choke limit in terms of compressor efficiency. For example, one manufacturer
defines the choke limit is the point on the pressure versus pressure curve where
compressor efficiency drops below 58%.

Using the above information, it is now possible to understand a compressor map, an


example of which is shown in Figure 16. Commonly, contours of constant compressor
efficiency are included on the compressor map. In addition to the surge and choke
limits, compressor maximum speed is fixed�usually by material limitations. These
three curves define the usable envelope of pressure ratio and flow that the
compressor can achieve.

Figure 16. Example compressor map


BorgWarner EFR 9180 turbocharger
Corrected Parameters. In order to account for differences in ambient conditions,
the compressor map can be expressed in terms of variables that have been corrected
to standard temperature and pressure. Corrected mass flow is normally expressed as

Corrected mass flow = Actual mass flow � (v?)/d (6)

where

? = Tactual/Treference (7)

d = Pactual/Preference (8)

This was done for the compressor map of Figure 16. Some manufacturers use corrected
volume flow instead:

Corrected volume flow = Actual volume flow / (v?) (9)

Several other corrected parameters may also appear on compressor maps including:

Corrected speed = Actual speed / (v?) (10)


Corrected torque = Actual torque / d (11)

Corrected power = Actual power / (dv?) (12)

The reference temperature and pressure used by different manufacturers is not


entirely consistent and the user should be careful to check what these conditions
are.

A compressor map can also be expressed in nondimensional variables. A dimensional


analysis of a centrifugal compressor reveals that

[equation] (13)

For high Reynolds number, the Reynolds number effect tends to be weak. For a fixed
geometry compressor and constant gas properties, this can be simplified to

[equation] (14)

The variable groupings in this simplification are no longer non-dimensional. In


this representation, the compressor speed parameter could have the dimension of rps
(revolutions per second) / vK, while the mass flow could be expressed in units of
(kg/s) vK / MPa. Figure 17 shows a compressor map expressed in such simplified
variable groupings derived from the dimensional analysis. Figure 17 also plots
compressor efficiency versus the flow parameter for a series of compressor speeds;
a representation that is sometimes used. It should be noted that variables and
units used in compressor maps from different manufacturers can vary.

[chart]
Figure 17. Compressor map plotted using simplified variable groupings from
dimensional analysis
(Source: Holset Turbochargers)
A parameter that is sometimes referred to in relation to compressors is the
structural trim. For a compressor (see Figure 3 for variable definitions):

Trimstructural = (d1/d2)2 (15)

As can be seen in Figure 18, decreasing trim shifts the compressor map to the left.
It is also important to note that a trim decrease lowers the flow rate at choking
conditions for a given impeller speed.

[chart]
Figure 18. Effect of compressor trim on compressor map
Equation (15) applies to compressor wheels where the trailing edge is parallel to
the axis of rotation. In cases where it is not, the aerodynamic trim can be used
[Houst 2015]:

Trimaero = (d1/d2,RMS)2 (16)

And where d2,RMS is defined as:

d2,RMS = v[(d22 + d2,tip2)/2] (17)

where:
d2 = diameter of the wheel at root end of the exducer (i.e., where the hub and
trailing edges meet)
d2,tip = diameter of the shroud edge of the wheel at the exducer (i.e., where the
shroud and trailing edges meet)

The ratio of the area of the volute inlet to its distance from the compressor
shaft, the so called A/R ratio (see Figure 39) can also be referred to in relation
to compressors. Compressor performance is relatively insensitive to A/R ratio.
Directionally, larger A/R housings are used for low boost pressure and smaller A/R
housings for high boost pressure applications. The A/R ratio is much more critical
with respect to turbines (see discussion later).

Other parameters that can be used to discuss compressor characteristics include the
annulus area and the vaneless diffuser annulus area. Definitions for these can be
found elsewhere [Houst 2015].

Figure 19 illustrates how the airflow requirements at constant engine speed and
load would appear on a typical compressor map.

[chart]
Figure 19. Compressor map overlayed with constant load and constant speed airflow
characteristic from a 4-stroke diesel engine
Figure 20 overlays the full load compressor performance curves for several
different applications from the 1980s on a compressor map with a dimensionless flow
parameter [Bosch 1986]. It should be noted that turbocharger control has changed
considerably and the characteristics for more modern applications would be
considerably different, especially for the truck and passenger car applications.

Figure 20. Compressor map showing example engine performance lines


Figure 21 shows the range of the compressor map that would be utilized between no
load and full load for and older application where boost pressure is
limited�possibly with a wastegate [Henein 1985]. At very low speeds and loads, the
compressor would produce no boost.

[chart]
Figure 21. Performance characteristics of turbocharger compressor
Turbocharger Turbine

Types of Turbine Geometry


There are three types of turbines suitable for exhaust gas turbochargers: radial-
flow, axial-flow and mixed-flow, Figure 22 [Mathis 2003]. The mixed flow turbine
has characteristics between the radial and axial turbines. The high-pressure flow
enters the turbine at the inlet (3�) and is guided to the stator inlet (3�) where
vanes turn the flow in a direction tangential to the rotor. The flow leaves the
stator vanes and enters the rotor blades (3), which turn the flow back in the
opposite direction, extracting energy from the flow. The flow leaving the rotor
blades (4), now at a lower pressure, passes through a diffuser where a controlled
increase in flow area converts dynamic head to static pressure. After the diffuser,
the flow exits to the discharge conditions.

[schematic] [schematic] [schematic]


Figure 22. Three possible exhaust gas turbine geometries
Variable Geometry Turbochargers
The purpose of the inlet is to guide the flow from the supply to the stator vanes
with a minimum loss in total pressure. A radial turbine will include a volute or
scroll to help achieve this objective. The stator or nozzle induces a swirl
component to the flow so that a torque can be imparted to the rotor blades. Stators
can be equipped with numerous curved airfoils called vanes that turn the flow in
the tangential direction. Many radial-inflow turbines for turbochargers often have
no vanes in the stator. To achieve good efficiency over a wide range of inlet flow
conditions, variable-geometry stators can be used, typically with either pivoting
stator vanes (Figure 23) or a variable width nozzle (Figure 24) [Carter 2009].

[schematic]
Figure 23. Swing Vane Variable Geometry Turbine
[schematic]
Figure 24. Moving nozzle or moving wall type variable geometry turbine
The rotor extracts energy from the flow, converting it to shaft power. The rotor
blades are attached to a rotating disk that transfers the torque of the rotor
blades to the turbine output shaft. Like the stator, the rotor has a number of
individual curved airfoils called rotor or turbine blades. The blades are angled to
accept the flow from the stator with minimum disturbance when the turbine is
operating at design conditions. The flow from the stator is then turned back in the
opposite direction in a controlled manner, causing a change in tangential momentum
and a force to be exerted on the blades. The flow leaving the turbine rotor can
have a significant amount of kinetic energy. If this kinetic energy is converted to
static pressure in an efficient manner, the turbine can be operated with a rotor
discharge static pressure lower than the static pressure at diffuser discharge.
This increases the turbine power output for a given inlet and discharge conditions.

Radial Flow Turbines. The choice of turbine type depends on a variety of factors
including turbine efficiency, packaging requirements and manufacturing cost.
Generally, for engine applications of about 1000 hp/turbocharger or less (i.e.,
turbine wheel diameter of about 130 mm or less), cost and simplicity are critical
factors and radial turbines are typically the most attractive option. In these
applications, maximum exhaust temperatures tend to be higher and radial-flow
turbines are the better choice. Compared to the axial-flow turbines, there is a
much larger difference between the rotor inlet relative and absolute velocities for
the radial-flow turbine that results in a lower relative inlet total temperature at
the design point. Also, due to the decrease in rotor speed with radius, the
relative total temperature decreases toward the root of the radial-flow turbine
blades. This is a major advantage for applications where turbine rotational speed
can potentially be limited by temperature-dependent material properties. Due to
bending stress considerations in the rotor blades, a radial-flow turbine would use
a radial blade tip profile at the inlet to eliminate bending loads. This combined
with the decreased temperature in the high-stress blade root areas allows the
radial-inflow turbine to operate at significantly higher wheel speeds than an
axial-flow turbine to provide an appreciable increase in turbine efficiency for
high-pressure-ratio applications [Mathis 2003]. In this size range, radial turbines
also offer better performance because of tip clearance considerations and there is
no need for a separate nozzle ring which simplifies manufacturing and reduces cost
when compared to axial turbines.

Axial Flow Turbines. Axial turbines, Figure 25, are typically used in applications
that exceed about 1000 hp/turbocharger (i.e., turbine wheel diameter of more than
about 130 mm) where they offer an efficiency and size advantage over radial flow
turbines. Examples of applications that can use an axial flow turbine include
medium and low speed diesel engines. For many of these large engines, maximum
turbine inlet temperatures and pressure ratios are lower than for smaller high
speed engines and the blade speed of an axial wheel is not constrained by stress
considerations. The radial-flow turbine would be at a size disadvantage. The use of
radial blade tips on a radial flow turbine to eliminate bending stress limits the
ratio of the tangential component of the absolute inlet velocity to blade tip speed
(CT/U) to 1 or less while an axial flow turbine can have CT/U > 1 with only a small
impact on efficiency. This would result in a larger and heavier radial-inflow
turbine than axial flow turbine when a given power and rotational speed is required
[Mathis 2003].

[schematic]
Figure 25. Turbocharger featuring radial flow compressor and axial flow turbine
MAN Diesel & Turbo Series TCA
Axial turbines typically need a relatively large exhaust diffuser and collector for
maximum efficiency which can make them more difficult to package for a single
turbocharger installation where space is limited. However, if two turbines are used
in a series arrangement such as for a turbocompound or dual turbocharger
installation, an axial turbine can bring both manufacturing and packaging benefits.
When the exhaust from a radial turbine is directed straight into an axial turbine,
a considerably simplified piping arrangement can be used, Figure 26 [Greszler 2008]
[Gobert 2007].

[schematic]
Figure 26. Turbo-compound utilizing axial flow turbine connected to a turbocharger
with a radial flow turbine
Axial turbine manufacture poses particular challenges. For large axial turbines,
individual blades are often mounted on the hub in what is known as the �fir tree�
root arrangement. However, such a technique would be too costly for small wheels
where one-piece, blade, disk and hub arrangements (�blisk�) are used. However, this
requires complex tooling and limits the choice of blade shapes more so than with
radial turbines.

Large axial turbines also commonly use a �damping wire� to dampen blade vibration.
To incorporate this wire, a hole in each blade at about 50-75% span is required
which would be impractical on a small axial wheel being produced in high volumes.
Potential vibration issues therefore have to be tackled through basic blade design
that leads to compromises in performance.

Mixed Flow Turbines. Choosing between radial-flow and mixed-flow is less


straightforward. The original motivation for mixed flow turbine wheels for
turbochargers was the reduction of turbine mass and inertia to improve engine
transient response. A mixed-flow turbine wheel can handle a higher gas flow than a
radial turbine of equivalent diameter. Later it was discovered that mixed-flow
turbines can also allow both improved aerodynamic blade loading�leading to greater
efficiency�and the ability to �adjust� the turbine�s peak efficiency point to a
higher or lower engine speed that can be useful in maximizing pulse energy from the
exhaust. However, the merits of a mixed-flow turbine must be weighed against the
more complex tooling required to produce it. Furthermore, the turbine housing
requires more space which could be an issue in a vehicle�s congested engine
compartment [Holset 2004a][L�ddecke 2012]. Audi�s 3rd generation of TFSI engines
(EA888) are an example of a modern application that uses mixed flow turbines. The
1.8 L engine uses an IHI mixed flow turbine while the 2.0 L (Generation 3B) uses a
Continental RAAX mixed flow turbine, Figure 27.

[]
Figure 27. Continental�s RAAX (RAdial AXial) turbine for Audi�s 2.0 L Generation 3B
TFSI engine
Constant Pressure and Pulse Turbines
Constant Pressure Turbocharging. The volume of the piping between the exhaust
manifold and the turbine housing plays a very important role in the dynamic
operation of the turbocharger. By enlarging this volume as well as that of the
manifold, pulsations from the exhaust valves discharging into the manifold are
dampened and the pressure at the turbine inlet is much more constant. This approach
is referred to as constant pressure turbocharging. While this method ensures the
continuous flow of exhaust gas over the turbine blades, it does not maximize the
use of exhaust energy in the turbine because the pressure peaks are damped out. It
can also result in higher exhaust back pressure which can reduce engine volumetric
efficiency.

In the past, constant pressure systems were common in two-stroke engine


applications [Obert 1968]. They continue to be used primarily in highly rated
engines for constant speed and load and in marine engines.

Pulse Turbocharging. A second approach makes use of the pulsating nature of exhaust
emanating from each of the exhaust valves and is called pulse turbocharging or
pulse conservation. A common design for in-line engines from 4 to 6 cylinders
couples the front cylinders in a single manifold that leads to one side of the
connecting flange, and the rear cylinders in another single manifold leading to the
other side of the same connecting flange, Figure 28 [Obert 1968]. This method is
preferred for applications where engine response is important�including most on-
highway applications�since exhaust energy utilization is maximized and exhaust back
pressure is lower.

[schematic]
Figure 28. Pulse turbocharging
In modern engine designs utilizing pulse turbocharging and where one turbocharger
is served by 4 or 6 cylinders, it is common to use a twin scroll turbocharger,
Figure 29. In twin scroll designs, the two volutes extend around the entire
circumference of the rotor.

[schematic]
Figure 29. Cross-section of a twin scroll turbine
An example for a 4 cylinder engine is shown in Figure 30. The exhaust inlet is
split into two parts where one part carries exhaust from cylinders one and four,
while the other from cylinders two and three. Pulse turbocharging also prevents
exhaust pulses exiting the cylinder from one cylinder forcing exhaust gas back into
another cylinder when exhaust valve open events overlap [Grosse 2000]. Greater part
load efficiencies are possible and more air is available at full load. Fuel
consumption reductions of 15 to 20% are possible.

[schematic]
Figure 30. Twin scroll turbocharger for a 4-cylinder engine
Exhaust gas recirculation in a pulse turbocharging system can be a challenge and
pulse energy can be optimized if EGR is recirculated to the intake manifold from
only one side of the two-piece exhaust manifold. This would create a significant
difference in the exhaust flows reaching the turbine from the two exhaust
manifolds. One solution is to use an asymmetrical twin scroll turbine, Figure 31.
This approach is used by Daimler in its engines intended for US EPA 2010 and Euro
VI applications.

[schematic]
Figure 31. Asymmetrical twin scroll turbine
Dual volute turbines are another way to capture exhaust pulse energy. Unlike the
twin scroll design, the individual volutes in dual volute turbines only extend part
way around the rotor�s circumference, Figure 32. The challenge with dual volute
turbines is that different pressures in two adjoining blade channels when the rotor
moves over one of the separating walls can lead to critical blade vibrations. Thus,
twin scroll designs are usually preferred.

[schematic]
Figure 32. Dual volute turbine
However, one advantage of the dual volute turbine is that it is much easier to
combine the benefits of a variable geometry nozzle into a pulse turbine, Figure 33
[Sauerstein 2009]. While a twin scroll turbine could in principle also incorporate
a variable geometry nozzle, for example by using a movable divider wall separating
the twin scrolls, implementation is much more challenging [Anschel 2011].

[schematic]
Figure 33. Dual volute turbine with variable geometry nozzle
(Source: BorgWarner)
Pulse Converter. A third way of connecting the exhaust manifold to the turbine
housing is called the pulse converter. Figure 34 describes one embodiment of the
pulse converter method showing a venturi box that converts exhaust kinetic energy
to potential energy or pressure, which is similar to that in the constant pressure
system.

[schematic]
Figure 34. Pulse converter
Turbine Energy Extraction

Gas flow through a turbine is considered to be an adiabatic expansion process.


Ideally, this process is isentropic as shown in the T-s diagram of Figure 35. It
should be noted that as with a compressor, assuming that the exhaust is an ideal
gas is reasonable. For an ideal gas, enthalpy, h, is only a function of
temperature, h(T), and the temperature differences in Figure 35 also reflect
enthalpy differences. The actual energy transfer in a turbine is smaller than the
isentropic value due to irreversibilities in the flow. The actual process is marked
by an increase in entropy and is also represented in Figure 35. The enthalpy change
associated with an entropy increase is less than that for an isentropic process.
The ratio of the actual to isentropic enthalpy change reflects the degree of
entropy increase and is referred to as the isentropic (sometimes adiabatic)
efficiency:

?isentropic = (hin - hout) / hisentropic (18)

It can also be shown that the energy transfer to the turbine wheel can be expressed
as

?hactual = U3 CT3 - U4 CT4 (19)

[chart]
Figure 35. Ideal and actual expansion process across a turbine
Losses. The difference between the ideal and actual enthalpy change across the
turbine is made up of losses from various sources. These sources include the inlet,
stator, rotor, diffuser and exit [Mathis 2003].

Inlet losses primarily arise from frictional and turning effects. The inlet should
be made as large as the packaging restraints allow; reducing velocities and
minimizing losses. Axial inlets are short and have relatively low velocities and as
such have low frictional losses but often suffer from turning losses due to flow
separation along their outer diameter. Longer axial inlets with more gradual
changes in outer diameter tend to reduce the turning losses and prevent separation,
but adversely impact turbine envelope. Tangential entry inlets such as those found
in radial flow turbines tend to have higher losses due to the tangential turning
and acceleration of the flow. The spiral flow path also tends to be longer,
increasing frictional losses.

Stator losses arise primarily from friction within the vanes, secondary flows
caused by the flow turning, and exit losses due to blockage at the vane trailing
edges.

Rotor losses are similar to those for stators but with a few additional components
to account for tip clearance and windage losses. Turbine rotors operate with a
small clearance between the tips of the blades and the turbine housing. Flow leaks
across this gap from the high-pressure side of the blade to the low-pressure side,
causing a reduction in the pressure difference at the tip of the blade. This
reduces the tangential force on the blade, decreasing the torque delivered to the
shaft. Windage losses arise from the drag of the turbine disk.

Losses in the diffuser arise from sources similar to those in other flow passages,
namely, friction and flow turning.
As for exit losses, in a single stage turbocharger turbine they equal the exit
kinetic energy of the flow unless used in an additional turbine stage or in some
other manner. If the kinetic energy of the flow exiting the diffuser is used in
later stages, the exit losses are zero.

Dimensionless Parameters. To make comparisons between different turbines easier, a


number of variable groupings arising from dimensional analysis are sometimes used
[Mathis 2003].

Specific Speed. The specific speed of a turbine is defined as

[equation] (20)

where Q2 is the volumetric flow rate through the turbine at rotor exit conditions.
The specific speed is used to relate the performance of geometrically similar
turbines of different size. In general, turbine efficiency for two geometrically
similar turbines at the same specific speed will be approximately equal. However,
small differences can exist due to clearance and Reynolds number effects.
Maintaining specific speed of a turbine is a common approach to scaling a turbine
to different flow rates.

Specific Diameter. The specific diameter is defined as

[equation] (21)

where dtip is the tip diameter of either a radial in-flow or axial flow turbine
rotor. Specific diameter and specific speed are used to correlate turbine
performance.

Blade-Jet Speed Ratio. Turbine performance can also be correlated against the
blade-jet speed ratio, which is a measure of the blade speed relative to the ideal
stator exit velocity. The ideal stator exit velocity, C0, is calculated assuming
the entire ideal enthalpy drop across the turbine is converted into kinetic energy:

[equation] (22)

The blade-jet speed ratio is then calculated from

[equation] (23)

The value of blade speed at the mean turbine blade radius is typically used for
axial turbines while for radial-inflow turbines, the rotor tip speed is used.

Turbine Performance

Figure 36 shows a typical performance map for a radial flow turbine. The pressure
ratio represents the ratio of inlet to outlet pressure. One important feature to
note is that mass flow typically reaches a limiting value that is unique to each
rotational speed. This is a result of choking. The choked flow limit increases as
turbine speed decreases.

[chart]
Figure 36. Typical radial-flow turbine map
Rarely is an entire map such as that in Figure 36 shown. Rather, turbine
performance around peak turbine efficiency is more commonly found, Figure 37. When
the pressure ratio versus mass flow characteristics at a number of turbine speeds
are plotted in this way, they often form a curve that is almost continuous. A line
drawn through these curves at the peak efficiency for each turbine speed is
sometimes referred to as the turbine swallowing capacity and represents the desired
operating curve of the turbine. In order to better match the turbine swallowing
capacity to the flow from the engine, it is important to select a suitable A/R
ratio for the turbine (see below). Further adjustments can be made by controlling
the swallowing efficiency with a wastegate or variable geometry turbine.

[chart]
Figure 37. Turbine swallowing capacity (lower curve) and efficiency (upper curve)
Figure 37 also shows measured turbine efficiency (the product of mechanical and
isentropic efficiency) at a number of different turbine speeds as a function of
turbine pressure ratio. Another way to present turbine efficiency is to plot it
against blade speed ratio, Figure 38. The peak turbine efficiency is typically
found near a blade speed ratio of about 0.7 for radial flow turbines and around 0.4
for axial flow turbines.

[chart]
Figure 38. Turbine isentropic efficiency as a function of blade speed ratio
A/R Ratio. An important parameter to match the swallowing capacity of the turbine
to the flow from the engine is the A/R ratio. It is defined as the cross-sectional
area of the inlet of the scroll over the distance of the centroid of that area to
the center of the turbine shaft, Figure 39.

Figure 39. Illustration and definition of A/R ratio


Figure 40 illustrates the effect of the A/R ratio on the swallowing capacity of a
radial flow turbine. A small A/R ratio increases the tangential velocity at the
turbine wheel tip for a given exhaust flow and thus provides higher turbine and
compressor rotational speeds. This can improve air flow at low engine speeds that
can be used to enhance low speed engine torque�especially if torque is smoke
limited.

Figure 40. Effect of A/R Ratio on Swallowing Capacity of an Exhaust Gas Turbine
Garret GT4508R turbocharger with 85 turbine trim
Figures 38 shows the effect of A/R ratio on smoke emissions at engine speeds below
peak torque speed. The turbocharger with the smaller A/R ratio helped reduce smoke
emissions at these two speeds by a substantial margin. Turbocharger lag is also
reduced and thus engine transient response improved with a smaller A/R ratio.
However, a small A/R ratio turbine will choke at lower flow rates. To maximize
engine power, a larger A/R ratio would be more beneficial.

[chart]
Figure 41. Effect of turbine A/R ratio on smoke from medium-duty DI diesel engine
For turbines, the trim is also an important parameter that influences flow
capacity. A larger trim will handle a higher flow rate and result in less
backpressure but will recover less exhaust energy and increase turbocharger lag.

References

Anschel, P., R. Chandramohanan, 2011. �Simplified Variable Geometry Turbocharger


With Variable Nozzle�, US Patent Application 2011/0232282 A1,
http://www.google.com/patents/US20110232282
Bosch, 1986. �Automotive Handbook�, Society of Automotive Engineers, Warrendale,
PA, 2nd Edition
Carter, J., et al., 2009. �Turbocharging technologies for heavy-duty diesel
engines�, In: "Advanced direct injection combustion engine technologies and
development: Diesel engines" (Volume 2), Ed. H. Zhao, Woodhead Publishing, Great
Abington, UK
Engels, B., 2002. �Lifetime prediction for Turbocharger Compressor Wheels - Why Use
Titanium?�, BorgWarner Knowledge Library,
http://turbos.bwauto.com/tools/download.aspx?t=document&r=107&d=109
Gobert, U., et al., 2007. �Turbo compressor system for internal combustion engine
comprising two serially placed turbo units with their rotation axes essentially
concentric�, US Patent 7,287,379 (Volvo Lastvagnar AB),
http://www.google.com/patents/US7287379
Golloch, R., 2005. �Downsizing bei Verbrennungsmotoren�, Springer
Greszler, A., 2008. �Diesel Turbo-compound Technology�, ICCT/NESCCAF Workshop,
Improving the Fuel Economy of Heavy-Duty Fleets II, February 20, 2008,
http://www.nescaum.org/documents/improving-the-fuel-economy-of-heavy-duty-fleets-
1/greszler_volvo_session3.pdf
Grosse, J., 2000. �New Engines For Renault in Quest For Green Machine�, FT
Automotive World, January 2000
Hanlon, P.C. (ed.), 2001. �Compressor Handbook�, McGraw-Hill
Henein, N.A., D.J. Patterson, 1985. �Combustion Engine Economy Performance and
Emissions�, Lecture Notes to Ford Tractor Operations, October 16, 1985
Heywood, J.B., 1988. �Internal Combustion Engine Fundamentals�, McGraw-Hill, New
York
Holset, 2004. �Machined-from-solid compressor wheels�, HTi Magazine, 2, 3
Holset, 2004a. �Choosing the right turbine for the automotive turbocharger�, HTi
Magazine, 2, 4-5
Houst, V., et al., 2015. �Functionally asymmetric two-sided turbocharger wheel and
diffuser�, US Patent application 2015/0063980 A1 (Honeywell),
http://google.com/patents/US20150063980
L�ddecke, B., et al., 2012. �On Mixed Flow Turbines for Automotive Turbocharger
Applications�, International Journal of Rotating Machinery, 2012, Article ID
589720, doi:10.1155/2012/589720,
http://downloads.hindawi.com/journals/ijrm/2012/589720.pdf
Mathis, D.M., 2003. �Fundamentals of Turbine Design�, In: Handbook of
Turbomachinery, Second Edition, Ed. E. Logan, Jr. and R. Roy, CRC Press,
doi:10.1201/9780203911990.ch7
Merker, G.P., et al., 2012. �Combustion Engines Development Mixture Formation,
Combustion, Emissions and Simulation�, Springer, doi:10.1007/978-3-642-14094-5
NATO, 2007. �Performance Prediction and Simulation of Gas Turbine Engine Operation
for Aircraft, Marine, Vehicular, and Power Generation�, Final Report of the RTO
Applied Vehicle Technology Panel (AVT) Task Group AVT-036.RTO, Technical Report TR-
AVT-036, http://www.cso.nato.int/pubs/rdp.asp?RDP=RTO-TR-AVT-036
Obert, E.F., B.H. Jennings, 1968. �Internal Combustion Engines�, International
Textbook Co., Scranton, Pennsylvania, 3rd Edition
Sauerstein, R., et al., 2009. �The Dual-Volute-VTG from BorgWarner � A New Boosting
Concept for DI-SI engines�, BorgWarner KnowledgeLibrary,
http://turbos.bwauto.com/tools/download.aspx?t=document&r=530&d=623
Schwitzer, 1991. �Introduction To Turbochargers�, Schwitzer Turbochargers,
Indianapolis, Indiana
Sharp, C.A., Khair, M.K., Gorel, A., 1999. �The Effect of A Turbocharger Clearance
Control Coating on The Performance and Emissions of a 2-Stroke Diesel Engine�, SAE
Technical Paper 1999-01-3665, doi:10.4271/1999-01-3665
Siuru, W.D., 2003. �Automotive Superchargers and Turbochargers�, In: Handbook of
Turbomachinery, Second Edition, Ed. E. Logan, Jr. and R. Roy, CRC Press,
doi:10.1201/9780203911990.ch13
Wu, C.Y., 1986. �Multiple cutter pass flank milling�, US Patent 4,596,501 (Pratt &
Whitney Canada Inc.), www.google.com/patents/US4596501
###

Vous aimerez peut-être aussi