See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/319166714

Conceptual Design of an Axial Turbocharger Turbine

Conference Paper · June 2017

DOI: 10.1115/GT2017-64825

CITATIONS READS
7 2,769

4 authors, including:

Apostolos Pesyridis Antonio Ferrara

Brunel University London DACA-I Powertrain Engineering
112 PUBLICATIONS   2,111 CITATIONS    1 PUBLICATION   7 CITATIONS   

SEE PROFILE SEE PROFILE

Raffaele Tuccillo
University of Naples Federico II
79 PUBLICATIONS   913 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Special Issue "Progress in Turbomachinery Research for Organic Rankine Cycle and Supercritical CO2 Waste Heat Recovery Systems" View project

Organic Rankine Cycle Technology Workshop - April 7th, London View project

All content following this page was uploaded by Apostolos Pesyridis on 27 November 2017.

The user has requested enhancement of the downloaded file.

 Proceedings of ASME Turbo Expo 2017: Gas Turbine Technical Conference and

DRAFT Exposition
GT2017
July 26-30, 2017, Charlotte, NC, USA

GT2017-64825

CONCEPTUAL DESIGN OF AN AXIAL TURBOCHARGER TURBINE

Apostolos Pesiridis Antonio Ferrara Raffaele Tuccillo Hua Chen

Institute of Energy Futures, University of Naples University of Naples National Laboratory
Brunel University London Federico II, Federico II, of Engine
London, United Kingdom Naples, Italy Naples, Italy Turbocharging
Technology, Tianjin,
China

ABSTRACT INTRODUCTION
Despite engine turbocharging being a widespread technology, The modern automotive industry is under strict regulations to
there are still drawbacks present in current turbocharging reduce emissions in order to comply with continuously updated
systems stemming from the apparent mismatch between the emissions standards. Downsizing of the internal combustion
periodic operation of a piston engine operating in conjunction engine is a technical solution that has become widely accepted
with an essentially steady-state, rotordynamic machine by the industry. Turbocharging is a cost effective method of
(turbocharger). The primary issue remains the provision of downsizing an engine whilst reducing exhaust gas emissions,
adequate transient response thereby suppressing the issue of reducing fuel consumption while maintaining performance. For
turbocharger lag (turbo-lag) or the poor initial response of the these reasons, turbocharging is becoming the most widely
turbocharger to driver-commanded, engine operating point adopted technology in the automotive market. In 2012, 32% of
changes due to its inertia. Another problem is engine- passenger and commercial vehicles sold had a turbocharger
turbocharger matching and operation under pulsating conditions installed, and this is predicted to increase to 40% by 2017 [1].
in the exhaust manifold and generally unsteady engine With this in mind and with the given drawbacks of
operating conditions. The exhaust flows of internal combustion conventional mechanical turbochargers in terms of transient
engines are characterized by pulsating flows at constant engine response, one way of mitigating turbo-lag is to introduce mixed
speeds (local pulsating effect) as well as “global” unsteadiness or axial flow turbines for turbochargers. Axial turbines are a
during engine transient events. Because of the volute volume bolder move with the idea in both cases being to replace the
and the length of the flow path, this unsteadiness generates a conventional radial turbine. The axial flow turbine has an
phase shift between mass flow, temperature and pressure at intrinsically lower inertia than the radial flow one, as stated by
rotor inlet, and a stronger circumferential variation of the rotor the research works of Rahnke, 1988 [2] and Bauer et al 2012
inlet condition than in steady flow conditions. The shift and the [3]. Furthermore, a turbocharger turbine coupled with an engine
variation increase the losses in the turbine, resulting in lower will work under pulsating flow conditions since the pulse
turbine efficiency. turbocharging system is the turbocharging system of choice in
The current paper develops original concept work carried modern automotive turbocharging. As a result, the turbine will
out at Brunel University London to develop an innovative fluid- work at highly variable blade speed ratios (U/Co). Turbine
dynamic design for an axial turbine for turbocharger efficiency at such conditions is normally poor, making it
application. An axial flow turbine coupled with a specially- difficult to extract energy and provide rapid acceleration.
designed, outflow volute, arranged in a non-classical way, are Improving the turbine efficiency at low U/Co conditions would
the target of this work. CFD analysis and 1D simulation of an produce benefits in both transient and steady state performance
engine coupled with the innovative turbine have been of the turbocharger and engine. In this sense, the axial turbine
performed to highlight the design potential. promises higher efficiencies at low U/Co than the radial one [3].
To further improve turbine performance a uniquely arranged
volute placed downstream of the rotor was designed. This
arrangement was chosen for the following reasons:

1 Copyright © 2017 by ASME

 The absence of a volute upstream of the turbine, means Brunel University of London has been employed: the model
that the inlet flows to the stator and the rotor are largely simulates a SI engine coupled with a Honeywell turbocharger
circumferentially uniform, under both steady and pulsating (GT1548) and it returns the same torque and power curves
flow conditions, so reducing the mixing losses that take place in released by Ford. The Table 1 shows the inlet total pressure and
conventional turbocharging turbines. temperature, the mass flow rate, the rotational speed and the
Since the flow path through the stator in the new turbine is outlet static pressure of the original radial flow turbine for
much shorter than that through a volute, pulsating flow energy different engine operations at maximum load.
can be better preserved increasing mean bulk flow energy,
while the flow condition at rotor inlet will more quickly Table 1-Design input data
synchronise with that at turbine inlet [4]. In other words, both
stator and rotor will work quasi-steadily with little unsteady Engine Speed
(max. load) 6000 5000 4000 3000 2000
effects [5] as would occur in conventional turbines, and this [rpm]
should improve turbine efficiency under such unsteady flow Mass flow 0.114 0.115 0.093 0.073 0.045
condition. [kg/s]
The new turbine also offers apparent advantages under Inlet Total 2.27 2.38 1.97 1.67 1.32
different unsteady flow conditions, namely, under engine Pressure [bar]
transient acceleration events, where it would take less time for Outlet Static 1.37 1.46 1.28 1.17 1.07
Pressure [bar]
engine exhaust to reach the rotor of the new turbine because of
Power [kW] 8.6 8.11 6.01 4.14 1.51
the inherently shorter flow path required in the much smaller,
resulting exhaust manifold and turbine volute volume between Speed [rpm] 137992 146199 132084 115917 87655
the engine and the rotor. Inlet Total
Moreover, the possibility of recovering the static pressure temperature 1050 1090 1066 1040 1039
in the downstream system can lead to an increase in total-to- [K]
static efficiency. Furthermore, it is possible to virtually increase
the expansion ratio of the rotor, so reducing the outer diameter
at the same blade speed.
The work described hereinafter details the authors’ SPECIFIC SPEED ANALYSIS
methodology for the design of the concept axial inflow turbine. One of the purposes of this project is to replace the
Although the final goal of this activity is to improve the conventional radial turbine of an existing turbocharger with a
unsteady and transient response of a turbocharged engine, a new axial turbine in a similar arrangement as proposed
first validation of the design effectiveness compares the engine Lotterman and Kares, 2014 [6]. Generally, an axial machine is
performance with a conventional radial flow turbocharging suitable for higher flow rates, while the radial turbine is able to
turbine and with the innovative turbine arrangement. process in a single stage a greater enthalpy drop. A preliminary
analysis of the specific speed at the different operating points
NOMENCLATURE obtained from the engine simulation is needed in order to
Symbols confirm the feasibility of the use of an axial turbine in place of
𝜌𝑡 = Total Density [kg/m3] 𝑛 =Rotational Speed [rpm] a radial turbine.
𝑀 = Mach Number [-] 𝑚̇ = Mass Flow Rate [kg/s] The suitability of each type of machine to a specific
𝛾 = Specific Heat Ratio [-] 𝛥ℎ𝑡,𝑖𝑑 =Ideal Total Enthalpy application depends on it specific speed and diameter defined
𝑐𝑝 = Specific Heat [kJ/kg/K] Drop [kJ/kg] by [7]:
𝑟 = Radius [m] 𝛥ℎ𝑡𝑠,𝑖𝑑 = Ideal Total to Static
𝑚̇
𝑣𝑥 = Axial Velocity [m/s] Enthalpy Drop [kJ/kg] 𝜔√
𝜌
𝑣𝑟 = Tangential Velocity [m/s] 𝜌 = Static Density [kg/m3] 𝑁𝑠 = (1)
𝐷𝑚 = Mean Diameter [m] 𝛥ℎ𝑡,𝑖𝑑 3/4
𝑣 = Velocity [m/s]
U = Blade Speed [m/s] 𝑁𝑠 = Specific Speed [-]
𝜓 = Azimuth Angle [rad] 𝐷𝑠 = Specific Diameter [-] 𝐷𝑚 𝛥ℎ𝑡𝑠,𝑖𝑑 1/4
𝐷𝑠 =
A= Cross Section Area [m2] 𝑇𝑡 =Total Temperature [K] √𝑚̇⁄𝜌 (2)
R= Universal Gas Constant [kJ/kgK] 𝑇 =Static Temperature [K]
Co = Isentropic Speed [m/s]
In this case, both Ns and Ds have been calculated for the
Abbreviations
CFD = Computational Fluid Dynamics
different engine operating conditions at full load in Table 1, by
evaluating the total-to static enthalpy drop according to the
DESIGN INPUT DATA predicted inlet and outlet conditions. Therefore, the Ns and Ds
The target engine selected for the application is a medium- values identify a point on the Balje diagram [8], (Fig. 1) for
small size gasoline engine i.e., the 1.6-litre Ford EcoBoost. In each operating condition considered. Indeed, the Balje diagram
order to obtain reasonable input values, a 1D model, developed [8] gives an expectation of the turbine efficiency as a function
by University of Naples Federico II in cooperation with the of Ns and Ds. The expected operating points for the project

2 Copyright © 2017 by ASME

turbine (plotted dots highlighted in Fig.1) fall within a good
efficiency zone (red circle, 70-80%) for the purpose of a Overall Geometry of the Turbine
65
turbocharger turbine. This means that the idea of replacing the
radial turbine with an axial one appears to be reasonable for this 52.45 52.38 51.93 53.62
55 51.48
application.
45 41.42 40.56 40.58

Diameter (mm)
39.65 38.46

35

25

11.89 12.73 13.47 13.04

15 10.06

5
1000 2000 3000 4000 5000 6000 7000
[rpm]
Data at Engine Speed
Mean Diameter Blade Height Max Diameter
Figure 2- Turbine overall geometry

Figure 1-Balje Diagram

The maximum rotor diameter changes slightly especially
DESIGN INPUT CHOICE from 5000 rpm to 2000 rpm so it can be concluded that the
Since a turbocharger turbine operates across a wide operating design point choice does not affect the maximum diameter of
envelope, the aim of the designer is to obtain a very flexible the machine for a fixed reaction degree and blade speed ratio.
machine, which is able to effectively in this region i.e., with The blade height instead grows with the engine speed and it
good efficiency. In the preliminary design phase, a design point reaches a max value for the 5000-rpm regime. Since the
has to be chosen to start the design process: once a design is maximum diameter influences the inertia of the turbine and the
obtained, successive analyses will be necessary to understand if blade height influences the efficiency of the machine (the gap
the machine can work properly in a wide range of points, and, between the tip of the blade and the casing has a higher
if not, design modifications will need to be made. In order to incidence as the height blade decreases) the design point could
make a choice, it is interesting to analyze how the geometry be set at 5000 rpm in order to reduce the losses without
broadly changes. To make this, a 1D-simulation code was weighing on the inertia of the machine. Furthermore, the choice
developed. Using this code, a preliminary calculation of the of the design point at 5000 rpm allows definition of a larger
axial turbine geometry has been made at each of the engine turbine inlet area: in this way, it can adapt to a wide range of
operating points. Since the purpose is to analyze how the mass flow rates as shown in the Table 2.
geometry changes with the design point parameters, the
reaction degree and the blade speed ratio have been fixed while Table 2 – Inlet area for different design points
the inlet and outlet conditions, the mass flow rate and the speed Engine Speed[rpm] 6000 5000 4000 3000 2000
have been varied for the range of expected operating 2
Inlet Area [mm ] 1661 1626 1584 1514 1308
conditions. The results are reported in Fig.2.
Furthermore, to avoid an oversized turbine, a good practice is
to set the design point not at the maximum load but at 75% of
the load: in this way, the design point chosen, does not lie on
the boundaries of the operating range of the engine. From the
1D simulation of the engine, the following results are obtained
in Table 3.

3 Copyright © 2017 by ASME

 Table 3- Design point data PRELIMINARY DESIGN
Inlet Total Pressure [bar] 2.35 The preliminary design requires the definition of the maximum
Inlet Total Temperature [K] 1097 diameter of the machine, since a limit is imposed by the
diameter of the compressor for good dynamic matching. The
Outlet Static Pressure [bar] 1.45
maximum diameter of the GT1548 compressor is 48.6 mm and
Mass Flow rate [kg/s] 0,113 the turbine should be of a nearly similar value. However, lower
Rotational Speed [rpm] 145669
efficiencies pertain to lower maximum diameters as well as
lower inertia. Therefore, for the preliminary design, a trade-off
analysis was carried out to establish the maximum diameter. A
It is worth noting that in the original turbocharger the outlet 1D commercial turbomachinery code was used (Axstream) to
static pressure is the one at the turbine exit while, in the perform the preliminary calculation by setting as input data the
modified turbocharger, the outlet static pressure is the one at the values of the design point previously chosen and the Craig &
outflow of the volute since it is arranged downstream of the Cox model for loss correlation [10]. This allowed identification
turbine. of a domain for the possible solutions for the single stage
It should be pointed out that the volute, is necessary to machine. In Fig.3, the the results are plotted as total-to-static
collect the exhaust gases from the turbine and to direct the efficiency vs maximum diameter. The aforementioned limit on
gases to the discharge duct. For this to be achieved, it is maximum diameter imposes a compromise solution
necessary to convert the flow from the axial to radial direction characterized by good efficiency characteristics without
prior to discharge. As stated above, the volute can be exploited increasing the inertia. For the solution selected, the code returns
to recover static pressure, like a diffuser: in this way, it is the velocity triangles for the stator and the rotor at mid-span,
possible to accept a higher kinetic energy at the rotor discharge hub and tip section as well as the blade size (Fig.4). The 15
(so an apparently lower total-to-static efficiency). To stator blades have an axial length of 7 mm and a 9 mm mean
investigate this possibility, the turbine was designed with a height while the 14 rotor blades have an axial length of 10 mm
rotor outlet static pressure lower than the pressure at the outlet and an 11.6 mm mean height. The machine has a mean
of the original turbine, based on the expected pressure recovery diameter of 36.5 mm and a maximum diameter of 48.4 mm
through the discharging diffuser and volute. A reasonable (Fig.5).
definition of the rotor outlet pressure has been reached by
assuming a pressure recovery coefficient of 50% [9] and a
volute exit Mach number of 0.2 (lower than the Mach number
at the exit of turbine since the flow decelerates) and by
assuming isentropic flow conditions. Table 4 shows the effect
of reducing the rotor outlet static pressure for a fixed blade
speed ratio: the overall size of the turbine is reduced with a
lowering of its inertia.

Table 4 – Effect of reducing the discharge static pressure

original revised
Figure 3 – Turbine efficiency versus maximum diameter
Static Pressure Outlet Turbine [bar] 1.45 1.39
Total-to-static Efficiency [%] 74.07 73.01
Blade Speed Ratio [-] 0.55 0.52
Rotor Mean Diameter [mm] 38.56 37.56
Rotor Blade Height [mm] 13.41 13.47
Stator Mean Diameter [mm] 38.56 37.56
Stator Blade Height [mm] 12.05 12
Maximum Rotor Diameter[mm] 51.97 51.03
2
Inlet Area [mm ] 1623.6 1588.6

Figure 4 – Mid-span section velocity triangles

4 Copyright © 2017 by ASME

 are solved in this reference frame. TurboGrid has been used to
generate the meshes of the two fluid domains. In this
preliminary calculation, it has been decided to use a fine mesh
composed of 246,000 hexahedral elements for the stator and
360,000 hexahedral elements for the rotor: the meshes are
composed of structured blocks arranged in an unstructured
manner [11]. The solver used, CFX, allows using a stage
method (mixing-plane) to interface the two domains:
circumferential averaging of the fluxes through bands on the
interface was performed. This method accounts for time-
averaged interactions, but not transient interactions. It leads to a
steady-state solution and requires computational efforts,
suitable for a preliminary calculation, with a good accuracy
Figure 5 – Geometric detail of annular section [12]. The turbulence model selected was the k-ω that gives
acceptable results for turbomachinery problems as suggested in
3D DESIGN literature. The boundary conditions, suitable for this kind of
Once the preliminary design phase was completed, the design
problem, are total pressure and temperature at the inlet of the
process proceeded with the construction of the blade profile.
stator and static pressure at rotor outlet. The design point
The ANSYS BladeGen software was employed, and the blade
provided these values as well as the rotational speed of the rotor
profiles were defined by using the blade angles and the annulus
domain. The calculations were carried out by considering
sizes provided by the preliminary 1D analysis carried out
different numbers of blades: the original one (15 stator and 14
earlier (Fig.6).
rotor blades), a configuration with 11 stator blades and 12 rotor
blades with little increase in blade thickness and a configuration
with 19 stator blades and 20 rotor blades with a reduction of the
blade thickness. In Table 5, they are compared in terms of
overall performance and flow conditions. The configuration
with a higher number of blades led to a higher efficiency, with a
mass flow rate closest to the project value. The reduction of the
maximum Mach number indicated improved flow guidance
provided by the stator, which approaches the rotor blade row
with the correct incidence angle (Fig.7).

Table 5 - Overall performance of different configurations

Configuration Original 11/12 19/20
Tts Efficiency [%] 70.9 69.8 74
Ttt Efficiency [%] 85.2 82.3 82.8

Figure 6 – Rotor blade profile & 3D models Power [kW] 14.9 15.1 13
Mass Flow Rate [kg/s] 0.134 0.153 0.114
CFD PRELIMINARY CALCULATION
A CFD preliminary calculation was necessary to evaluate the
design effectiveness and to calculate the flow conditions at the
exit of the machine that are the input data for the design of the
volute. Since the rotor inlet conditions are circumferentially
uniform and the outlet ones could be considered in the same
manner, so that the periodicity of the rotor blade passages can
be assumed and therefore a single stator and rotor blade
passage analysis was chosen. The stator and the rotor passages
were considered as two separate domains: the stator fluid
domain is considered in a fixed reference frame while the
reference of the rotor domain is rotating and the flow equations

5 Copyright © 2017 by ASME

 pressure, the inlet axial velocity, the inlet density and the inlet
tangential velocity. These fluid dynamic properties were
obtained from the CFD preliminary analysis as a mass flow
average of the properties distribution at the outlet of the turbine.
The r2 and A2 were fixed to obtain a certain pressure recovery.

Table 6 – Pressure Recovery for Different Exit Diameters

Inlet Width [mm] 11,9 11,9 11,9 11,9 11,9
Inlet Radius [mm] 18,25 18,25 18,25 18,25 18,25
Outlet Width [mm] 11,9 11,9 11,9 11,9 11,9
Outlet Radius [mm] 24 30 34 40 50
Press. Recovery [%] 21,6 50,6 62,5 72,6 82,6
Figure 7- Flow velocity distribution at rotor blades
The pressure recovery calculated using the isentropic flow
VOLUTE DESIGN equations is only an approximation since friction losses,
The designed axial machine was coupled with a diffuser-type, boundary layer effects, diffusion and curvature effects are not
volute system for collecting the gas flow from the rotor considered. The diffuser with an outlet radius of 34 mm seems
directing it to discharge at its outlet, recovering the static to be satisfactory, in terms of pressure recovery potential for a
pressure at the same time. To shift the flow from axial to radial given size.
directions, a diffuser of the type represented in Fig.8 has been The volute was preliminarily designed assuming a uniform
chosen. It slows down the flow by reducing the tangential flow with varying azimuth angle [13]; the mass flow rate at
component of flow velocity (conservation of angular each section of the volute can be expressed in the following
momentum). It also reduces the axial velocity component way:
𝜓
thanks to the increase in the annular cross sectional area [9]. 𝐴𝜓 𝜌𝜓 𝑣𝑟𝜓 = 𝐴0 𝜌0 𝑣0𝑟 ( ) (4 )
The preliminary design was carried out using the isentropic 360
Where A0, ρ0, v0r are the cross-sectional area, the density and
flow equations in order to obtain a preliminary indication of
the tangential velocity at the inlet of the volute ( exit from the
both diffuser size and pressure recovery that could be achieved.
diffuser). Aψ, ρψ, vrψ are the cross-sectional area of the volute,
the density and the tangential velocity at a generic azimuth
angle, respecitvely. The free vortex law is used to compute the
tangential component of the velocity:
𝑟0
𝑣𝑟𝜓 = 𝑣𝑟0 (5 )
𝑟𝜓
The term rψ is the distance between the centroid of the
generic cross section and the center of the volute. The cross-
sectional area is calculated as:
𝜓
𝐴0 𝜌0 𝑣0𝑟 ( )
𝐴𝜓 = 360 (6 )
Figure 8 – Axial to Radial Exhaust Diffuser 𝑖 𝜌𝜓 𝑣𝑟𝜓
Where the density 𝜌𝜓 can be calculated with the isentropic
In the second phase, a CFD analysis was carried out to
flow equations assuming that the flow velocity at the cross
provide more reliable results. The system of isentropic flow
sectional-area can be calculated as:
equations used with conservation of energy was the following: 𝑣𝑟𝜓
𝑟1 𝑣𝜓 = (7 )
𝜌2 𝐴2 𝑣𝑥2 = 𝜌1 𝐴1 𝑣𝑥1 ; 𝑣𝑟2 = 𝑣𝑟1 ∗ 2
; 𝑣2 = √𝑣𝑥2 2
+ 𝑣𝑟2 (3.a) 𝑡𝑎𝑛(𝛼𝜓 )
𝑟2
drψ
The term 𝛼𝜓 is the flow angle and it is equal to r .A
(−𝛾−1)
1 ψ dψ
𝑣2 𝑣22 𝛾−1
𝑀2 = ; 𝑇2 = 𝑇𝑡 − ; 𝜌2 = 𝜌𝑡 (1 + ( ) 𝑀22 ) (3.b) relation between the centroid radius rψ and the azimuth angle
√𝛾𝑅𝑇2 2𝑐𝑝 𝛾
must be assigned: this can be done with a Bezier polynomial of
the third order [13]. At this point, it is possible to calculate each
The geometric variables r1 and A1 are assigned (from the intermediate section of the volute. The tongue are presented a
geometry of the rotor) as well as the total temperature, the total problem from the design point of view at the intersection

6 Copyright © 2017 by ASME

between the cross sectional area at 0° and the cross sectional
area at 360°. The resultant design is illustrated in Fig.9.

Figure 9 – Volute tongue design

The initial CFD investigation of the volute, indicated that

Figure 11 – Symmetric volute configurations tested
the tongue region gave unsatisfactory performance for
application with an axial rotor. Looking at the velocity vectors In Fig.11 different designs of the symmetric volute are
in Fig.10 in a sectional plane orthogonal to the axis of the shown. The different shapes are obtained by changing the shape
volute, it can be observed that the flow does not follow cross- and the area of both the duct outflow and the cross section area
sectional contours but partial back-flow occurs instead. with respect to the first trial volute. The exit area has been
Furthermore, the flow impacts on the tongue indicating calculated by using the isentropic-flow equations. In Fig.12 the
increased total pressure losses and a local acceleration of the solutions are compared in terms of size, pressure recovery and
flow. fluid dynamic inertia.

System Size
180
151.72 146 151.72 151.72
146 139 146
150 137
Length (mm)

116 112 113

120 107
90
54 54 54
60 41
30 25
30
0
Figure 10 – Streamlines and velocity vectors in a classic volute
1 2 3 4 5 6
volute
This phenomenon occurs because the flow at the exit is Height Width Axial Length
almost purely radial so the spiral shape of the volute is counter- Figure 12 – Volute performance comparison
productive since it works better when the flow has a strong
The optimum solution was a compromise between size,
tangential component. In this case, the flow does not follow the
spiral contours, impacting on the wall of the volute especially pressure recovery and fluid dynamic inertia represented by the
fluid volume trapped by the volute. The results, are obtained
on the tongue, with total pressure losses and a lower pressure
from a CFD analysis using a medium size tetrahedral mesh for
recovery occurring as a result.
the volute.
Since the flow has an almost symmetrical distribution at the
exit, a symmetric volute could be designed. The target of this
new design was to collect the flow from the turbine and recover
the static pressure.

7 Copyright © 2017 by ASME

 between two successive simulations fell below 1%, the
optimum mesh density was considered as having been reached.

Figure 13-Streamlines in the chosen volute design

Figure 15 – Stator mesh refinement

Figures 13 shows clear streamlines through the chosen

In Fig.15 an example of the mesh density rise for the stator
volute: as the flow exits radially the volute initially splits the
domain is shown: an increase to the number of elements along
flow in two streams which are then merged near the top at the
the streamwise and spanwise directions, as well as of the
outlet Figure 14 includes the rotor relative positioning in
boundary layer. For the stator and rotor fluid-domain a multi-
relation to the streamline field.
block, structured mesh has been used while for the short
diffusing section and volute outlet domain an unstructured
tetrahedral mesh has been used (Fig.16).

Figure 16 – Volute & diffuser mesh

The same analysis has been made for the rotor, the diffuser
Figure 14- Streamlines in the chosen volute design and the volute domain. The optimum number of elements for
each domain is shown in Table 7. Once again a single passage
TURBINE CFD ANALYSIS of the machine has been simulated and presented.
The turbine was composed of four sub-domains: a stator, a
rotor, a diffuser and a volute outlet fluid domain. These were Table 7 – Optimized Mesh Element Number for Each Domain
singularly considered and interfaced such that a different mesh Domain Elements
was built for each of them in order to choose the optimum Stator 105946
number of elements and the best element shape for each of the Rotor 148990
subdomains that compose the entire turbine fluid domain. To
Diffuser 155678
find the optimum number of elements, a mesh sensitivity
Volute 524670
analysis has been conducted on each subdomain. For each
subdomain, the number of elements was increased by 50% for
From the CFD simulation, it is possible to see that the flow
each successive calculation, monitoring the mass flow rate and
conditions are basically uniform in each stator passage due to
the total pressure at the outlet: when the percentage difference
the uniformity of the inlet conditions (Fig.17).

8 Copyright © 2017 by ASME

 CFD calculation are shown using as boundary conditions the
data of the engine operating points from the 1D engine
simulation. The results show that the turbine works efficiently
in a wide range. An efficiency increase of over 5% shows that
the diffuser-volute has been correctly designed.

Total to Static Efficiency

81

Total To Static Efficiency [%]

79 79.24
78.48 78.34
77 77.57
76.14
Figure 17 – Stator static pressure distribution 75
73 73.73
73 72.21 72.49
On the other hand, the non-uniformity of the rotor outlet 71 71.23
conditions causes little difference among the several rotor 69
passages (Fig.18). Comparing simulation results for one 67
passage of the machine with the one for the entire rotor, it is
65
possibile to state that the choice of modeling one passage of the
1000 2000 3000 4000 5000 6000 7000
machine allows a reduction of the computational time without [rpm]
affecting the accuracy of the solution appreciably (Table 8). Data at Engine Speed
Figure 19 – Efficiency improvement of new volute design compared to
conventional volute

The turbine map of the designed system has also been

obtained numerically. For each turbine speed, an inlet total
temperature and an outlet static pressure have been fixed (in
accordance with the values obtained from the 1D engine
simulation). The inlet total pressure has been varied to simulate
different expansion ratios until the chocking condition was
reached. Therefore, for each expansion ratio, the mass flow rate
obtained through CFD has been corrected. In this way a
characteristic curve at constant speeds was calculated. This
process is repeated for different turbine speeds.
Once the turbine map was obtained, the last step was to
Figure 18 – Rotor static pressure distribution
perform a simulation of the engine with the old turbine replaced
Table 8 – Results comparison by the new one in the (Ricardo Wave) engine model.
3D Model Entire One In Figures 20 and 21 the turbine map, as calculated
Rotor Passage originally and processed using Ricardo Wave is shown. The
System Total-to-static Efficiency[%] 78.84 78.87 performance of the engine, boosted by the new turbocharger, is
System Total to Total Efficiency[%] 80.19 80.18 presented and compared with the performance of the engine
Turbine Total-to-static Efficiency[%] 73.42 73.48 boosted by the original, radial turbocharger in Figure 22.
It is worth to highlight that the calibration parameters of the
Turbine Total to Total Efficiency[%] 83.96 84.08
engine have not been changed and neither has the threshold
System Pressure Recovery [%] 55.4 55.58
pressure of the waste-gate been modified. Obviously, the
Volute Pressure Recovery [%] 8.5 8.57 performances of the engine boosted with the new turbine, could
Diffuser Pressure Recovery [%] 51 51.46 be further improved by recalibrating the engine parameters.
Power [kW] 13.348 13.374 Looking at the power curves, Fig.22, it is clear that the new
solution does not penalize the power output of the engine: this
The rotor and the diffuser domain are interfaced in the same is important since one of the objectives was to replace the radial
way, using the mixing-plane (stage) method implemented in turbine with an axial turbine, maintaining the performances of
Ansys-CFX. The CFD results show that the choice of modeling the engine at a good level.
one machine passage is justified. In Fig.19 the results of the

9 Copyright © 2017 by ASME

 the same pressure boost as the original one. One of the aims of
the project can be therefore considered as reached: to produce a
turbine able to ensure the same performance as the
conventional one.

150
Engine Power

Power [kW]
100
Turbine with new
volute
50 Radial turbine

Axial turbine with

0 conventional volute
1000 3000 5000 7000
Figure 20 – Turbine map constructed numerically Engine speed [rpm]
Figure 22 – Comparison engine performance: Power

260 Engine Torque (Maximum Load)

Engine Torque (Nm) 240

220

200
Turbine with innovative volute
180
Original Turbine
160
1000 2000 3000 4000 5000 6000 7000
Engine Speed (rpm)
Figure 21 – Turbine map constructed numerically (in Figure 20) Figure 23 – Comparison engine performance: Torque
extrapolated here using Ricardo Wave
200000
In this way, the engine operates in the same way but with an
Turbine Speed [rpm]

axial turbine substituting the radial turbine to improve its 150000

transient response. In some points, at high engine speeds, the
power is higher than the original (axial) one as a result of the 100000
increased expansion ratio. Fig.23 shows that, the axial turbine Turbine with
with the new, innovative volute ensures a greater torque from 50000 innovative volute
2000 rpm with a curve that is quite flat which points to Original Turbine
improved drivebility. 0
In Fig.24, a comparison between the turbine speeds is 1000 2000 3000 4000 5000 6000 7000
shown. As it is possible to observe, the new turbine speed Engine Speed [rpm]
curves follow quite closely the original turbine speed. This is
Figure 24 – Turbine speed comparison
expected since the turbine has been designed to work under the
same conditions as the original one, so ensuring almost the The compressor pressure ratio for each case is plotted in
same boost level. Obviously, there are some little differences Figure 25. The pressure ratio is very similar in behaviour to the
because the turbocharger works in different operating points as original, but with a generally slightly higher level of boost
previously stated. Fig.24 confirms that the new solution ensures provided in some regimes. This happens at high speeds where

10 Copyright © 2017 by ASME

the pressure ratio provided by the turbine is greater than the turbine of an existing turbocharger with an axial turbine. The
original one as well as the power output. The figure is a aim was to reduce the turbo-lag phenomenon by lowering the
confirmation that the pressure ratio could be maintained but inertia of the machine while maintaining an otherwise
achieved more rapidly further proving the new design concept. acceptable performance level. Moreover, an innovative layout
Figures 26 and 27 conclude with the final design of the of the turbine-volute has been analyzed: in the designed
entire system shown (Fig. 26) and the uniform pressure turbine, the volute was placed downstream of the axial flow
distribution of the proposed new volute (Fig.27). rotor. The idea was to enhance the performance of the machine
2 preserving inlet flow kinetics and thermal inertia by requiring a
shorter manifold coupling to the engine. Furthermore, the
Compressor Pressure Ratio

1.8 efficiency of the turbine can be improved by recovering static

pressure in the downstream system, designing it as a diffuser.
1.6 The design of the machine developed from a preliminary
1.4 Turbine with phase to justify the choice of an axial machine and was
innovative volute followed up by a detaile design and 3D CFD analysis through
1.2 Original Turbine both a single passage as well as full-stage analysis.
The conventional volute was replaced by a symmetric
1 volute offering superior performance for axial turbocharger
0 1000 2000 3000 4000 5000 6000 7000 turbines. Finally, the performance map of the the system has
Engine Speed [rpm] been numerically constructed. This map has been used in the
Figure 23 – Compressor Pressure Ratio Comparison engine model to evaluate the impact of the new turbine design.
The axial and original radial turbines were eventually evaluated
in terms of engine and turbine performance.
The CFD results have shown that the innovative
arrangement works to a satisfactory level. The axial turbine has
a high efficiency in a wide range of operation and the
downstream volute section allows static pressure recovery,
thereby increasing the total-to-static efficiency of the turbine.
The engine simulation results show that the modified
turbocharger, based on the modified axial turbine system, can
ensure the same performances as the original one in terms of
power and torque of the engine for a more compact, higher
response system with efficiencies superior to both the original
radial turbine as well as the original axial turbine.
Figure 26 – 3D Draw of the Entire System Future work for this project includes a rotordynamic
analysis of the new turbocharger with the slightly elongated
shaft in place of the original radial turbine, full transient engine
data and a re-calibrated engine to operate optimally with the
new turbine in place.

REFERENCES
[1] Reports and Reports, 2014, "Turbocharger Market by
Vehicle Types (Passenger and Commercial), Fuel Types
(Diesel and Gasoline) and Geography - Global Trends and
Forecasts to 2017", Markets and Markets.
[2] Rahnke C.J., 1988, "Axial Flow Automotive
Turbochargers" ASME 85-GT-123.
Figure 27 – Pressure Distribution at Design Point.
[3] Bauer, Balis and Devis, 2012 , "The next generation of
CONCLUSIONS gasoline turbo technology" Internationales Wiener
In this work, an innovative turbine for turbocharger application Motorensymposium.
has been proposed, analysed, designed and computionally [4] Baines,2010,"Overview of pulsating flows in
tested. The objective of the project was to replace the radial turbochargers".

11 Copyright © 2017 by ASME

[5] Hakeem I., Costall A., Su C., and Martinez-Botas A.,2007,
"Effect of volute geometry on the steady and unsteady
performances on mixed flow turbines" in Proceedings of
the Institution of Mechanical Engineers Part A Journal of
Power and Energy 221(4):535-549 June 2007.
[6] Lotterman J., and Kares V.,8 Jul 2014, "Axial Turbine".
Patent US 8769950 B2.
[7] Baskharone, 2006, Principles of Turbomachinery in Air-
Breathing Engines, Cambridge.
[8] Balje O.E.,1981, Turbomachinery: A guide to Design,
Selection and Theory.
[9] Aungier R.H., 2006, Turbine Aerodynamics.
[10] Cox H., and Craig H.,1970, "Performance Estimation Of
Axial Flow Turbines" Instn Mech Engres, Vol 185 32/71,
pp. 407-424.
[11] Ansys, 2012, TurboGrid: High quality mesh for
turbomachinery.
[12] Ansys, CFX Theory Guide, 2012.
[13] Abidat M.,Hamidou M., Hacheml M., and Hamel A.,2006,
"Design and flow analysis of radial and mixed flow
turbine volutes" vol. ECCOMAS CFD.
[14] Xie L., Zhang Y., Zhuge W., Chen L. and Zhang S.,2011,
"Effects of Pulsating Flow Conditions on Mixed Flow
Turbine Performance" ASME paper GT2011-45164.
[15] Zhou C., Himmel C.,and Hodson H.,2013, "The Effects of
Trailing Edge Thickness on The Losses of Ultra-High Lift
LP Turbine Blades", ASME paper GT2013-94029.
[16] Ramezani A., Remaki L., Blanco J., and Antolin J.,2014,
"Efficient Rotating Frame Simulation in Turbomachinery"
ASME paper GT2014-25101.
[17] Roclawski H., Gugau M., Langecker F., and Bohle M.,
2014, "Influence of Degree of Reaction on Turbine
Performance for Pulsating Flow Conditions" ASME paper
GT2014 - 25829.
[18] Schoenenborn H., and Ashcroft G., 2014, "Comparison of
Non-Linear and Linearized CFD Analysis of the Stator-
Rotor Interaction of a Compressor Stage", ASME paper
GT2014-25256.
[19] Gwehenberger T., 2009, "Single-Stage High-Pressure
Turbocharging", ASME paper GT2009-59322.
[20] Casartielli E., Mangani L., Hanimann L., Mokulys H., and
Mauri S., 2013, "Development of a Nover Mixing Plane
Interface Using a Fully Implicit Averaging for Stage
Analysis", ASME paper GT2013-94390.
[21] Feng Z.,Li J., and Sun,2009, "Effects of Tip Clearance on
Unsteady Flow Characteristics in an Axial Turbine
Stage",ASME paper GT2009-59828.
[22] Arnold S., Martin, Slupski, and Kanigowski,2001,
"Variable Geometry Turbocharger" Patent US 6269642
B1.

12 Copyright © 2017 by ASME

View publication stats