Content uploaded by Dezong Zhao
Author content
All content in this area was uploaded by Dezong Zhao on Mar 22, 2016
Content may be subject to copyright.
Performance testing of an electrically
assisted turbocharger on a heavy duty
diesel engine
E. Winward (1), J. Rutledge (2), J. Carter (3), A. Costall (4), R. Stobart (1), D.
Zhao (1), Z. Yang (1)
1: Dept. Aeronautical and Automotive Engineering, Loughborough University, UK
2: Research and Advanced Engineering, Caterpillar Inc.
3: BorgWarner Turbo Systems, UK
4: Imperial College London, UK
Abstract
An electrically assisted turbocharger was designed, built and tested by a consortium
consisting of Caterpillar Inc., BorgWarner Turbo Systems, Imperial College London
and Loughborough University. The Electric Turbocharger Assist (ETA) device was
based on a BorgWarner BV63 variable turbine geometry (VTG™) turbocharger with
a bearing housing that was extended to accommodate a switched reluctance (SR)
electrical machine. The ETA device was evaluated over a range of steady state and
transient engine conditions with the ETA providing electric assist or electric
regeneration. The response of the single stage ETA turbocharger matched a
production 2-stage turbocharger on the same engine.
1. Introduction
Electric Turbocharger Assist (ETA) describes a turbocharger device in which the
turbocharger operation is assisted by means of an electric motor/generator which is
mechanically coupled to the turbocharger shaft. Such devices are also referred to
as a hybrid turbocharger or turbocharger motor/generator. Electrification of the
turbocharger can be applied to both standard fixed or variable geometry turbines.
There have been a number of research programs which have investigated the
integration of an electric machine with an exhaust gas turbocharger. In many
concepts the electric machine is built into the turbocharger bearing housing (1-5).
There is at least one example where the electrical machine is overhung on the
compressor side (6).
An ETA device such as that described in this paper has two operating modes: 1)
Electrical assist mode (herein referred to as motoring) in which the electric motor
provides additional steady state boost pressure or accelerates the rotor during
transients to give faster boost pressure rise rate and improved engine transient
response (1,4,6-10). 2) In electrical regenerative mode (herein referred to as
generating) the ETA can be used to recover part of the otherwise waste heat of the
exhaust by generating electricity from excess turbine power (2,11,12). A battery
can be used to supply the electrical power to the ETA when motoring and store
energy when the ETA is generating.
Electrically assisted turbochargers provide an additional degree of freedom for
engine air path control and have the potential to improve the engine-turbocharger
matching and resizing of the turbine for optimal efficiency (maintaining transient
performance via motoring assist) (12,13). These benefits can potentially give
thermal efficiency gains of 10% (6). The ETA as a technology can be an enabler for
engine down-sizing, engine down-speeding, improving power density and reducing
parasitic losses.
This paper describes the ETA hardware which was developed and tested by a
consortium consisting of Caterpillar Inc., BorgWarner Turbo Systems, Imperial
College London and Loughborough University and which was co-funded by Innovate
UK (formerly the Technology Strategy Board UK). The paper presents results from
engine steady state and transient testing of the ETA on a 7 litre off-highway diesel
engine.
2. Description of ETA Hardware
The turbocharger used for this project was a modified BV63 from BorgWarner Turbo
Systems which has a variable turbine geometry (VTG™). The electric machine was
a switched reluctance topology supplied by Magnomatics. The integration of the
electric machine with the turbocharger was carried out by BorgWarner. The core
assembly can be seen in Figure 1. In order to accommodate the electric machine,
the rotor shaft length is increased significantly compared to the standard
turbocharger. The electric rotor has a larger diameter than the bearings and
therefore requires the bearing housing to be split into two halves. It was found that
a vertical split line as seen in Figure 1 was preferable to a horizontal split line.
Nevertheless, careful control of the concentricity of the two halves is required to
maintain rotor shaft motion within limits.
The switched reluctance machine has the advantage of a simple rotor structure
without any magnetic material, which results a high rotor temperature capability
and no concern about the high cost of rare-earth materials. However, compared to
permanent-magnet machines, the SR machine typically has lower power density
and greater noise signature.
There are a considerable number of different switched reluctance motor topologies
which can be defined by the number of phases and the associated numbers of
stator and rotor poles. The choice of topology depends on many factors including
the starting capability, torque ripple, iron core losses, converter cost, power
switching losses and commutation frequency.
Figure 1: Integration of electric machine with turbocharger
Single phase and two phase machines generally have poor starting capability but
they have an advantage in reduced switching frequency compared to designs with
more phases. They also require fewer components in the power converter and have
a lower assembly cost due to fewer machine coils.
A commonly used SR topology is 3-phase machine with 6 stator poles and 4 rotor
poles (e.g. 6/4 type) and this was selected for the ETA, Figure 2. This machine is
capable of starting from any position in either direction. This was selected based on
the balance of requirements spanning torque ripple, iron core losses, converter cost,
power switching losses and commutation frequency. The SR motor was designed to
give a peak electrical power in excess of 5kW in both motoring and generating. This
enabled evaluation of both modes of operation of the prototype ETA on the test
engine over a useful electrical power range.
Figure 2: 6/4 type switched reluctance electrical machine
The electric machine’s rotor is clamped between turbine wheel and compressor
wheel. This has the advantage of simple assembly, not requiring press- or shrink-fit
of the electrical rotor to the shaft. Lubrication oil was used to cool the stator of the
electrical machine. This eliminates the risk of oil and water mixing through casting
porosity and reduces the hoses required.
A position sensor-less control method was devised to control the switched
reluctance electrical machine, this method avoided the requirement to package a
motor angular position sensor within the ETA device. This sensor-less control
enabled electrical operation of the machine between 40kRPM and 130kRPM (14).
Figure 3 summarises the electrical operating envelope of the ETA with sensor-less
control which could be practically explored on the test engine at Loughborough
University.
Figure 3: ETA position sensor-less control operating envelope
In Figure 3 the vertical axis represents the % of the maximum achievable torque of
the SR electrical machine in both motoring (+ve) and generating (–ve). The ETA
was controlled by setting a % of the maximum achievable torque. The speed of the
ETA was therefore a balance between the ETA % torque demand and engine
conditions (speed, load, air system settings, etc.). It was found that the ETA could
not be operated at high motoring electrical power demand set points at lower
turbocharger speeds (40 to 70kRPM) as the motor torque exceeded compressor
needs at those speeds. The other points within the operating envelope could be
reached under steady state conditions using a combination of varying engine speed,
load, and air system settings.
3. Gas Stand Testing
A ‘gimballed’ version of the ETA device bearing housing that allowed bearing
housing reaction torque to be measured was manufactured to be used for gas stand
testing at Imperial College London. This gas stand testing allowed the electrical
machine to be characterised both electrically and in terms of machine efficiency.
This electrical characterisation was an important part of developing the position
sensor-less control method mentioned above. The electrical machine also allowed
the turbine performance to be mapped at much lower flow rates than is possible on
a conventional gas stand (points at which turbine power is insufficient to overcome
bearing and other losses). This allowed extension of the turbine performance maps
using measured data, rather than extrapolation, into areas of operation seen during
engine transient load shedding. More information on the testing arrangement can
be found in the literature (15).
-100
-80
-60
-40
-20
0
20
40
60
80
100
30 40 50 60 70 80 90 100 110 120 130 140
Motoring / Generating set point (%)
Motor Speed (kRPM)
Position Sensorless Control
Demonstrated
Motoring
Operating
Envelope
Demonstrated
Generating
Operating
Envelope
Cannot hold
ETA in this
area even at
idle
4. ETA Installation
The engine testing of the device was carried out both at BorgWarner in the US (for
standard turbocharger hardware sign off testing) and at Loughborough University in
the UK (for controls development and performance evaluation). The testing at
Loughborough was carried out by fitting the device to a 7 litre displacement tier 4
off-highway diesel engine in place of the production 2-stage turbochargers. This
engine is equipped with high pressure loop cooled EGR and is used in such
applications as excavators, wheel loaders and gen sets. The ETA was mounted
remotely from the exhaust manifold for ease of installation of the prototype
turbocharger, Figure 4. This mounting solution led to some heat loss between the
exhaust manifold and the turbine which reduced the available energy in the exhaust
but also reduced the thermal load into the ETA and avoided excessive mechanical
load on the engine exhaust manifold. This installation solution also allowed initial
testing with a conventional (non-electrified) BorgWarner BV63.
The ETA was equipped with several K-type thermocouples which measured the
temperature at several locations inside the electrical machine stator. The oil supply
(bearing lubrication and stator cooling) to the ETA was independent of the engine
with a specially designed oil cart used, Figure 4. This allowed precise control of the
oil temperature and pressure. The test cell was equipped with a motoring dyno that
was capable of rapid load application to allow assessment of engine transient
response. Engine fuel consumption, air mass flow and exhaust emissions were all
measured.
Figure 4: ETA fitted to test engine in Loughborough University test cell
ETA
Oil Cart
5. Engine Steady State ETA Results
5.1 ETA winding temperatures (non-electrified)
Initial testing of the ETA prototype device focussed on resolving some issues with
winding wiring, winding insulation, and device sealing. Once these issues were
overcome tests were run to understand the sensitivity of electrical machine winding
temperatures to engine parameters such as load and exhaust temperatures without
any electrical assistance. Figure 5 presents results of the winding temperature
measurements compared to related parameters.
Figure 5: ETA winding temperature sensitivity (1800 RPM)
The winding temperature shown in Figure 5 is the measured temperature at one of
the end winding positions. The end windings are furthest from the cooling oil and
showed higher temperatures than other measurement points. Figure 5 shows a
generally reasonably linear relationship between end winding temperature and
engine load/exhaust temperature at the mid region but departs from this at the
Winding Temperature
Engine Power
Winding Temperatures vs Engine Power
Winding Temp Upper Limit
Winding Temperature
Engine Torque
Winding Temperatures vs Engine Torque
Winding Temp Upper Limit
Winding Temperature
Turbine Inlet Temperature
Winding Temperatures vs Turbine Inlet
Temp
Winding Temp Upper Limit
Winding Temperature
Turbine Power
Winding Temperatures vs Turbine Power
Winding Temp Upper Limit
Winding Temperature
Compressor Outlet Temperature
Winding Temperatures vs Compressor
Outlet Temp
Winding Temp Upper Limit
Winding Temperature
Exhaust Mass Flow
Winding Temperatures vs Exhaust Mass
Flow
Winding Temp Upper Limit
higher and lower ends. It is believed that the lower end departure from the linear
trend is due to the windings being heated by the leak off from the bearing oil circuit.
The intermittent departures from the linear temperature at the top end were
believed to be due to exhaust gas blow-by from the turbine leaking across either
the horizontally split bearing housing or the turbine bearings. Higher levels of
exhaust gas blow-by into the turbocharger oil return line were seen with the ETA
device than with production turbochargers.
An example of a sudden increase in winding temperature at higher loads can be
seen in Figure 6. In this case the end winding temperature starts to increase very
quickly once high load has been reached. The sudden increase in the rate of end
winding temperature does not coincide with the preceding increase in load, instead
it is delayed by nearly one minute. This suggests that the higher temperatures seen
at high loads is not solely due to the higher exhaust temps but some other effect,
such as the increased blow-by rates detailed earlier.
Figure 6: Load independent winding temperature increase (1800 RPM)
A correlation analysis was carried out to understand which factors affect the
winding temperature of the non-electrified ETA, Figure 7. Many of the parameters
that correlated with end winding temperature were not unexpected (engine load,
exhaust temps etc.) but it was also seen that increased bearing oil and cooling oil
flow rates led to increased end winding temperatures while increased bearing and
stator oil supply pressures led to reduced winding temperatures. The higher
winding temps from higher cooling flow rates may be due to the effect of greater
heating of the windings from increased splash from heated bearing oil (end
End Winding Temp
Torque
End Winding, Torque and Exhaust Time vs Time
0 450 900 1350 1800 2250 2700
Rate of Change of
winding temp
Turbine Inlet Temp
Time
Turbine Inlet Temp Rate of Change of winding Temperature
Torque End Winding Temp
windings are close to drainage side of bearings). The unexpected correlations is a
clear indication of the complexity of the ETA as a system to cool.
Figure 7: End winding 1 temperature correlation
Once the non-electrified temperature limits of the device had been established,
work was done to implement the ETA sensor-less control method throughout the
operating speed range of the machine. The position-sensor-less controls were
demonstrated under steady state conditions over the area shown in Figure 3.
5.2 ETA winding temperatures (electrified)
The effects of ETA electrical assist on winding temperatures was also investigated.
A series of sensitivity tests were carried out to assess rates of heating and cooling.
An 1800 RPM example of this is shown in Figure 8. This shows a series of tests
where the electrical machine was continuously motored at a constant low % torque
demand at different engine torque levels. The motoring continued until a
predetermined temperature was reached.
It was seen that at higher starting loads the winding temperature was already at a
higher temperature (as per the trend in Figure 5) and that it increased at a higher
rate from this point (i.e. the 300 Nm case took 40 seconds less time to reach the
upper temperature limit than the 100 Nm case took for the same temperature rise).
This faster rate of electrical heating is believed to be due to a combination of the
higher torque points having higher electrical power for the same motor torque
setting (due to the higher turbocharger speed) and increased winding resistance
from higher starting temperatures. The rate of cooling was also seen to be slower
at the higher exhaust temperatures. This clearly suggests that prolonged use of the
ETA at higher loads may lead to temperature issues.
No. Channel
1 Engine Speed
2 Engine Torque
3 ETA Speed
4 Sta tor Temp 1
5 Sta tor Temp 2
6 Winding Temp 1
7 Winding Temp 2
8 Winding Temp 3
i9 Oil Cart Bearing Pressure
h10 Oil Cart bearing flow rate
h11 Oil Cart bearing oil supply temp
i12 Oil Cart Stator Pressure
h13 Oil Cart Stator F low Rate
14 Oil Cart Stator oil supply temp
15 ETA Oil Return Temp
16 Exh Manifold Temp
17 Exh Temp Turbine Out
18 Air Mass Flow
19 Fuel mass Flow
20 Ambient Air temp
21 Compressor inlet temp
22 Compressor outlet t emp
23 Exhaust Mass Flow
0 5 10 15 20 25
-1
-0.8
-0.6
-0.4
-0.2
0
0.2
0.4
0.6
0.8
1
channel of variable
R Value
Corrleaiton for ETA-End-Winding-A1
9
12 15
10
11
13 14
Channel Number
R Value
Correlation for end winding temp 1
21
Figure 8: ETA electrical heating rate comparison at different loads
The winding temperatures were also monitored during engine performance testing.
An example is shown in Figure 9 where it can be seen that over a series of 6 block
load tests (load quickly increased at a constant desired engine speed) that the
effect of a few seconds of motoring assistance accumulates over each assistance
event leading to a noticeable divergence in winding temperature from an identical
case ran without the electrical machine active. This suggests that more intermittent
and shorter duration use of the device may lead to less temperature problems.
Figure 9: Effect of ETA block load assistance on winding temps
Winding Temp
Exhaust Temp
0 100 200 300 400 500 600
Rate of change of
winding temp
Time (second)
300 Nm 200 Nm 100 Nm 0 Nm
0 100 200 300 400 500 600
TurboSpeed
Time (second)
300 Nm 200 Nm 100 Nm 0 Nm
0
100
200
300
400
500
600
700
800
900
Torque Load
Winding Temperature over
repeated Block Loads
1
2
3
4
5
6
7
8
9
10
11
75
80
85
90
95
100
0100 200 300 400
ETA
Power
Winding
Temperature
Time
Winding Temperature
Torque Load
Winding Temp - No ETA Power
ETA Power
Effect of ETA
assistance on
winding temps
5.3 Effect of ETA motoring and generating on engine AFR and EGR
In order to assess the effect of the ETA on the steady state operation of the engine,
a range of ETA power swings (running the device at a series of motoring and
generating % torque demands) were carried out. These were run with the other air
system actuators either locked at a fixed position or attempting to control to a set-
point. The addition or subtraction of extra torque to the turbo-shaft had a
significant effect on the engine air system. The effect of a limited amount of ETA
motoring and generating power on the engine Air Fuel Ratio (AFR) and Exhaust Gas
Recirculation (EGR) achievable ‘space’ (with EGR and VTG vane position actuators
fixed) is illustrated in Figure 10. This shows that motoring the ETA will tend to
increase engine AFR while generating will reduce both AFR and EGR. The engine
particulate and NOx emissions will obviously also be affected by any changes in AFR
and EGR rate that the ETA drives.
Figure 10: Effect of ETA motoring and generating on engine AFR and EGR
(1800 RPM 40% load)
5.4 Effect of ETA electrical power on engine pumping efficiency
The other benefit of motoring is that it will increase the inlet manifold pressure with
no requirement for more turbine power (i.e. waste gate closure) that could lead to
an increase in exhaust pressure (although the increased air flow will increase
turbine flow leading to a possible increase in turbine expansion ratio and exhaust
manifold pressure).
An increase in inlet manifold pressure relative to exhaust manifold pressure will
lead to a reduction in engine pumping losses. Pumping losses can be defined as the
85%
90%
95%
100%
105%
110%
115%
55% 70% 85% 100% 115% 130% 145%
Normalised EGR
Normalised AFR
No ETA Power
Vane Pos 3 -
EGR Swing
Vane Pos 2 -
EGR Swing
Vane Pos 1 -
EGR Swing
EGR / VTG
Envelope
85%
90%
95%
100%
105%
110%
115%
55% 70% 85% 100%115% 130%145%
Normalised EGR
Normalised AFR
40% Motoring
Vane Pos 3 - EGR
Swing - 40% Mot
Vane Pos 2 - EGR
Swing - 40% Mot
Vane Pos 1 - EGR
Swing - 40% ETA
Mot
EGR / VTG
Envelope
85%
90%
95%
100%
105%
110%
115%
55% 70% 85% 100% 115% 130% 145%
Normalised EGR
Normalised AFR
40% Generating
Vane Pos 3 - EGR Swing - 40% Gen
Vane Pos 2 - EGR Swing - 40% Gen
Vane Pos 1 - EGR Swing - 40% ETA
Gen
EGR / VTG Envelope
EGR / Vane Pos Envelope - 40% Mot
EGR / Vane Pos Envelope - 40% Gen
‘pumping efficiency’ which is the ratio of the 720 deg. Indicated Mean Effective
Pressure (IMEP) to the 360 deg IMEP for a 4 stroke engine. This is plotted in Figure
11 against exhaust pressure for the data from Figure 10. The improved pumping
efficiency for a given exhaust pressure when the ETA is motoring can clearly be
seen. However, there is a reduction in pumping efficiency associated with
generating. This illustrates that the ETA is capable of decoupling exhaust pressure
from inlet manifold pressure which can be used to reduce engine pumping losses.
Figure 11: Effect of ETA power on engine pumping efficiency (1800 RPM 40%
load)
5.5 Sweeps of ETA power at different VTG and EGR valve positions
The ETA has also been seen to affect the engine efficiency in others ways. Motoring
has been seen to increase the engine friction due to the increased cylinder pressure
from increased air mass in the cylinder. Generating has been seen to lower AFR
which then leads to slower combustion, which reduces the engine closed cycle
efficiency.
Some of the ETA effects can be counter-acted through adjustments to other air
system actuators. Closing the VTG vanes will help maintain a higher AFR during
generating, reducing the gross indicated efficiency penalty. This will however be at
the cost of increased pumping losses. Figure 12 and Figure 13 show normalised
plots of engine brake, mechanical, pumping, closed cycle efficiency as well as turbo
speed and AFR against normalised ETA electrical power (-ve: generating, +ve:
motoring, 100% being the maximum ETA power achieved at a mid VTG and EGR
point). The data was taken at 1800 RPM 40% load. The data is plotted at two vane
positions and two EGR valve positions. The effect of the different engine air system
actuator settings on the engine’s response to ETA motoring and generating can be
seen.
Engine Pumping Efficiency
Exhaust Pressure
Pumping Efficiency vs Exhaust Pressure
No ETA Power 40% gen 40% mot
Figure 12: Normalised engine pumping eff., mechanical eff., closed cycle
eff. vs ETA power (1800 RPM 40% load)
From Figure 12 and Figure 13 it can be observed that there is a clear pumping and
mechanical efficiency penalty from having a more closed VTG vane position but it
can be seen that there is significant AFR and closed cycle efficiency benefit from the
more closed vanes, especially under higher levels of generating.
The brake efficiency plots in Figure 13 show the combined effects of ETA power on
engine mechanical, pumping and indicated efficiencies. It can be seen that a more
closed vane position gives lower brake efficiencies under non-motoring or motoring
cases (due to a combination of lower pumping efficiency from more restrictive
turbine and slightly lower mechanical efficiency from higher cylinder pressure).
As the ETA generates there is a drop in AFR. The more closed vane case sees a
higher minimum in AFR due to:
(i) The higher initial AFR.
(ii) A smaller proportional reduction in compressor power due to the
electrical power extracted from the turbine being a smaller portion of
the overall higher turbine power (more closed vanes gives a greater
expansion ratio).
(iii) The higher turbo speeds also appear to be of benefit as the
turbocharger stays in a more efficient part of the compressor map.
95%
96%
97%
98%
99%
100%
101%
102%
103%
-120% 0% 120%
Normalised Pumping Efficiency (%)
Normalised Motoring and Generating Power (%)
Pumping Efficiency vs ETA Power
More closed
vanes
99.0%
99.2%
99.4%
99.6%
99.8%
100.0%
100.2%
100.4%
100.6%
-120% 0% 120%
Normalised Mechanical Efficiency (%)
Normalised Motoring and Generating Power (%)
Mech Efficiency vs ETA Power
More closed
vanes
90%
92%
94%
96%
98%
100%
102%
104%
106%
-120% 0% 120%
Normalised Closed Cycle Eff. (%)
Normalised Motoring and Generating Power (%)
Closed Cycle Eff. vs ETA Power
More open vanes, less open EGR
More open vanes, more open EGR
More closed vanes, less open EGR
More closed vanes, more open EGR
More closed
vanes
Figure 13: Normalised engine AFR, turbo speed and brake Eff. vs ETA
power (1800 RPM 40% load)
The higher minimum AFR leads to higher closed cycle efficiency which outweighs
the lower mechanical and pumping efficiencies of the more closed vane case. This
leads, under high generating power levels, to equivalent or better brake efficiencies
compared to the more open case. The particulate emissions penalty associated with
running to such low AFRs with the more open vane cases is also significant.
The results presented in Figure 12 and Figure 13 illustrate that careful selection of
the ETA’s turbocharger is required to find one that can operate well both in
motoring and generating mode.
6 Engine Transient ETA Results
A key potential benefit of ETA technology is the potential to improve transient
response. This can be used to either:
90%
92%
94%
96%
98%
100%
102%
104%
-120% 0% 120%
Normalised Brake Efficiency (%)
Normalised Motoring and Generating Power (%)
Brake Eff. vs ETA Power
More open vanes, less open EGR More open vanes, more open EGR
More closed vanes, less open EGR More closed vanes, more open EGR
More closed
vanes
More closed
vanes
60%
70%
80%
90%
100%
110%
120%
130%
140%
150%
-120% 0% 120%
Normalised AFR (%)
Normalised Motoring and Generating Power (%)
AFR vs ETA Power
More closed
vanes
70%
80%
90%
100%
110%
120%
130%
140%
-120% 0% 120%
Normalised Turbo Speed (%)
Normalised Motoring and Generating Power (%)
Turbo Speed vs ETA Power
More closed
vanes
(i) Improve engine load acceptance to improve productivity.
(ii) Improve engine load acceptance to allow engine down-sizing or down-
speeding for improved fuel consumption.
(iii) Allow an engine to use a turbocharger matched for Brake Specific Fuel
Consumption (BSFC) rather than response and meet the transient
requirements using the ETA.
A range of transient tests were carried out on the ETA to quantify the potential
response benefits. These included:
Actuator step response tests.
Engine block load acceptance tests.
Simulated application working cycles.
The first two tests were relatively simple discrete transient events and could be run
without having the ETA and engine controls fully integrated. The final one required
a fully developed control system for the ETA, VTG and EGR.
6.1 Actuator step response tests
The actuator response tests involved carrying out comparable turbo speed step
changes using the ETA motor and the VTG vanes. The response of turbo speed and
other parameters was compared. The turbo speed results are shown in Figure 14.
Figure 14: ETA vs VTG vane position turbo speed response test
-6
-4
-2
0
2
4
6
0
40
80
120
160
200
240
45 50 55 60 65 70 75 80 85
Time (second)
TC Speed(kRPM)
Initial turbo speed (kRPM)
ETA and VTG response time comparison - TC
Speed
TC Speed at test start (ETA)
TC Speed at test end (VTG)
TC Speed at test start (VTG)
TC Speed at test end (ETA)
Time 5-95% - ETA
Time to 50% - ETA
Time 5-95% - VTG
Time to 50% - VTG
1.90 1.68 1.96
2.96 2.95
3.30
0
0.5
1
1.5
2
2.5
3
3.5
50 65 80
5-95% Response Time (second)
Initial Turbo Speed (kRPM)
ETA and VTG response
time comparison - TC
Speed
Time 5-95% - ETA Time 5-95% - VTG
0.63
0.44 0.40
0.67 0.69 0.75
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
50 65 80
5-50% Response Time (second)
Initial Turbo Speed (kRPM)
ETA and VTG response
time comparison - TC
Speed
Time to 50% - ETA Time to 50% - VTG
From Figure 14 it can be seen that the 5-95% turbo speed response time achieved
by ETA motoring assist was on average 1.2 seconds faster than that achieved by an
equivalent VTG vane step change. The 5-50% response time was faster by an
average of 0.2 sec and therefore the ETA advantage was mostly over the second
half of the step change. It can also be seen that the benefit was lower at lower
turbo speeds. This was thought to be due to the lower available motor electrical
power at those speeds.
An example of one of the tests (65 kRPM starting speed) is presented in Figure 15
and shows the ETA power, vane position and exhaust manifold pressure. Exhaust
manifold pressure increased significantly with closing the vanes while less so with
the ETA motoring assist. Thus, for the same step change in turbo speed, VTG will
result in increasing pumping losses (Figure 12). The reduced pumping loss penalty
on acceleration is therefore an additional benefit of the ETA.
Figure 15: 65 kRPM actuator response test
6.2 Engine block load test results
Engine block load tests were also carried out involving running the engine at low
load at a fixed speed and then increasing the load on the engine to a pre-
determined level as quickly as the dyno will apply it. The amount by which the
engine speed drops upon load acceptance (speed drop) and the time taken for it to
recover to its prior speed (recovery time) are measures of an engine’s transient
capability.
The performance of the baseline engine with its production 2-stage turbocharger
was set as an initial target for the ETA. It was found that the BorgWarner BV63
single stage turbocharger with VTG of the same aero specification as the ETA was
not capable of matching the 2-stage response. The non-electrified single stage had
worse speed drop and recovery time. This can be seen in Figure 16 which also
shows that the 2-stage case had slightly greater initial boost, which would increase
engine response. Despite lower initial boost the ETA was capable of matching the 2-
stage recovery time but did have slightly worse engine speed drop (which is more
affected by initial boost than rate of boost rise).
Block load tests were run over a range of other speeds and loads looking at the
effects of different ETA power levels, different assist durations, different initial
conditions and other factors. An example of some of the results is shown in Figure
4 4.5 5 5.5 6 6.5 7 7.5 8
ETA Motoring
Electrical Power
Vane Position
(% closed)
Time (sec)
65kRPM starting TC speed - ETA Power
and Vane Position vs Time
Vane position
ETA Power
0
20
40
60
80
100
120
140
160
0 5 10 15
Exhaust Manifold Pressure (kPa)
Time (sec)
Exhaust Manifold Pressure Response
Response to Vane
Closure
Response to ETA
power increase
17 where the effects of ETA electrical power on block load performance at 1400
RPM are shown.
Figure 16: ETA block load response vs 2-stage engine response
Figure 17 shows that the ETA is more effective at reducing recovery time than
speed drop and that the incremental response benefit from increased ETA power
levels starts to decrease (35% ETA power gives large benefit compared to no assist
case but 70% power does not double this). This is due to the engine speed
recovery becoming less air system constrained with increased levels of ETA power,
making any further boost increase with higher ETA power less effective.
Figure 17: 1400 RPM block load sensitivity to ETA power
Engine Speed (RPM)
Time (s)
Speed
Torque (Nm)
Time (s)
Torque
Boost (kPa)
Time (s)
Boost
2 stage baseline Single stage ETA
Single stage - no ETA
40%
50%
60%
70%
80%
90%
100%
600 650 700 725
Speed Drop (% of no
ETA assist case)
Block Load
35% ETA 70% ETA 90% ETA
40%
50%
60%
70%
80%
90%
100%
600 650 700 725
Recovery Time (% of
no ETA assist case)
Block Load
Engine Speed (RPM)
Time (s)
ETA Power Effect on 1400 RPM Block
Load Performance
No ETA Assist
ETA assist - 35%
ETA assist - 70%
ETA assist - 90%
6.3 Engine working cycle test results
Working cycles were run once a control system for the ETA had been integrated
with the engine controller. In the simplest control system implementation, ETA
transient motoring assist was triggered based on inlet manifold conditions and
other engine related parameters. The control system also had a battery state of
charge (SoC) maintenance algorithm which regulated ETA motoring and generating
in steady state conditions. This was required to assess effects on cycle efficiency of
maintaining sufficient battery state of charge to allow transient assistance.
One of the working cycles used to evaluate the ETA was a so called trenching test
in which the engine speed demand is fixed whilst the engine torque is constantly
changing rapidly over the course of several minutes. Figure 18 shows a range of
engine responses (engine speed, ETA electrical power, turbo speed, intake Mass Air
Flow (MAF) and the intake Manifold Air Pressure (MAP)) for two trenching tests, one
run with ETA assistance the other without.
Figure 18: Trenching test working cycle (with and without ETA assist at an
engine speed demand =1800 RPM)
The results presented in Figure 18 show very clearly how ETA motoring assist
improved the engine transient response relative to the no ETA assist case. In this
example the ETA is providing motoring assist during rapid load acceptance and then
generating briefly between. The increased acceleration of the turbocharger with
motoring assist and the increased MAP rise rate are clear and these contribute to
the improved engine speed response.
Utilisation of the ETA over a working cycle requires the battery SoC be maintained.
Figure 19 compares two tests that employ the same transient ETA control method
as illustrated in Figure 18; one cycle with battery SoC maintained (using a controls
strategy) and one where the battery is allowed to discharge. This shows that
battery SoC can be maintained but that it is at the cost of a 0.6% fuel penalty over
the working cycle.
Figure 19: SFC penalty of battery state of charge maintenance
Several different control systems were developed in simulation and tested on the
ETA equipped engine at Loughborough University (16-21). This work highlighted
that there exists very strong interaction between the ETA electrical power, VTG and
EGR that requires careful control system design and system energy management.
0.0%
0.1%
0.2%
0.3%
0.4%
0.5%
0.6%
0.7%
050 100 150 200 250 300 350
Cumulative fuel consumption
increase for charge maintenance
(%)
-0.01
0
0.01
0.02
0.03
0.04
050 100 150 200 250 300 350
Cumulative Battery
Energy Discharge
Battery discharged
Battery charged
-2
-1
0
1
2
3
4
5
6
7
050 100 150 200 250 300 350
ETA Power
Time
No battery SoC maintenance
battery SoC maintenance
Generating
Motoring
7 Conclusions
The presented work has demonstrated that operation of the ETA offers an
additional degree of freedom for air system control and can be used to influence the
engine pumping losses. Steady state motoring operation of the ETA will lead to
improved SFC through pumping benefits while steady state generating operation of
the ETA can lead to an SFC penalty through a combination of increased pumping
losses and the effects of reduced air fuel ratio on combustion. Prolonged or high
engine load operation of the ETA device is currently limited by stator winding
insulation temperature limitations. These cooling issues are attributed to the use of
oil for cooling, the design of the ETA cooling system and excessive blow-by at high
loads.
It has also been shown that the ETA device is capable of enabling improved engine
transient response (>60% reduction in block load recovery time) compared to a
variable geometry turbocharger through the ability to more quickly increase inlet
manifold pressure (allowing a faster increase in engine fuelling). Diminishing
benefits from increased ETA power were observed and illustrate that the there are
other controlling factors on engine transient response beyond the rate of boost
increase (i.e. initial boost level at start of transient, system inertia etc.) that start
to become the limiting factors. The ETA device allowed the block load recovery time
performance of a 2-stage turbo to be matched with a single stage turbocharger.
The ETA device has shown response benefits during on engine testing of a
simulated working cycle (reduced speed drop and faster recovery time). A fuel
penalty of 0.6% is seen when the battery state of charge is maintained over the
working cycle compared to allowing the battery to discharge.
This work illustrates the potential of ETA technology but highlights both the
challenges of cooling the electrical machine and the fuel penalty from maintaining
battery state of charge over a working cycle.
8 Future Work
BorgWarner has used the experience gained with the ETA prototype described in
Section 2 to develop an improved electric turbocharger concept, the BorgWarner
eTurbo™, Figure 20. This is a permanent-magnet machine with a higher power
density. This design incorporates a new rotor structure which allows improved
tolerances and optimal stiffness for best rotordynamic behaviour. The stator is
cooled by engine coolant (rather than oil) which allows greater power output whilst
maintaining temperatures within the limit of the winding material.
Figure 20: Core assembly of eTurbo™
Reference List
[1] Balis, C., Middlemass, C., and Shahed, S. M., “Design and Development of e-
Turbo for SUV and Light Truck Applications”. In 9th Diesel Engine Emissions
Reduction (DEER) Conference, Technical Session 4: Waste Heat Utilization, 2003.
[2] Hopmann, U., and Algrain, M. C., “Diesel Engine Electric Turbo Compound
Technology”. In Proceedings of the SAE Future Transportation Technology
Conference, paper 2003-01-2294, 2003
[3] Arnold, S., Balis, C., Barthelet, P., Poix, E., Samad, T., Hampson, G., and
Shahed, S. M., Garrett Electric Boosting Systems (EBS) Program. Final report, grant
DEFC05-00OR22809, US Department of Energy, 2005.
[4] Ibaraki, S., Yamashita, Y., Sumida, K., Ogita, H., and Jinnai, Y., “Development
of the hybrid turbo, an electrically assisted turbocharger”. Mitsubishi Heavy
Industries, Ltd. Technical Review, 43(3), pp. 1–5, 2006.
[5] Ryder, O., Sutter, H., and Jaeger, L., “The Design And Testing Of An Electrically
Assisted Turbocharger For Heavy Duty Diesel Engines”. In Proceedings of the
IMechE 8th International Conference on Turbochargers and Turbocharging, pp.
157–166, 2006.
[6] Panting, J., Pullen, K. R., and Martinez-Botas, R. F., “Turbocharger motor-
generator for improvement of transient performance in an internal combustion
engine”. Proceedings of the IMechE, Part D: Journal of Automotive
Engineering,215(3), pp. 369–383, 2001.
[7] Kolmanovsky, I., Stefanopoulous, A. G., Powell, B. K., “Improving Turbocharged
Diesel Engine Operation With Turbo Power Assist System,” in IEEE International
Conference on Control Applications, vol. 1, pp. 454–459, 1999.
[8] Katrašnik, T., Rodman, S., Trenc, F., Hribernik, A., Medica, V.,“Improvement of
the Dynamic Characteristic of an Automotive Engine by a Turbocharger Assisted by
an Electric Motor,” Journal of Engineering for Gas Turbines and Power, vol. 125, no.
2, p. 590, 2003.
[9] Terdich, N., Martinez-Botas, R. F., Howey, D. A., Copeland, C. D., “Off-Road
Diesel Engine Transient Response Improvement by Electrically Assisted
Turbocharging”, 10th International Conference on Engines & Vehicles, 2011.
[10] Terdich, N., Martinez-Botas, R. F., Romagnoli, A., Pesiridis, A., “Mild
Hybridization via Electrification of the Air System: Electrically Assisted and Variable
Geometry Turbocharging Impact on an Off-Road Diesel Engine,” Proceedings of
ASME Turbo Expo 2013, 2013.
[11] Millo, F., Mallamo, F., Pautasso, E., Mego, G. G., “The Potential of Electric
Exhaust Gas Turbocharging for HD Diesel Engines,” SAE Technical Paper 2006-01-
0437, 2006.
[12] Costall, A. W., Radoslav, I., Langley, T. P. F., “Electric Turbo Assist: Efficient
Rapid Boost for Heavy-Duty Diesel Engines”. THIESEL Conference on Thermo and
Fluid Dynamic Processes in Direct Injection Engines, 2012.
[13] Costall, A. W., Radoslav, I., Langley, T. P. F., “Electric Turbo Assist as An
Enabler for Engine Downspeeding” Proceedings of ASME Turbo Expo, June 11-15,
Copenhagen, Denmark, 2012.
[14] Watkins, S., Winward, E., “Modelling and Control of an Ultra-high-speed
Switched Reluctance Machine”, 8th IET International Conference on Power
Electronics Machines and Drives, 19-21 April, Glasgow, Scotland, 2016.
[15] Terdich, N. and Martinez-Botas, R. F, "Experimental Efficiency Characterization
of an Electrically Assisted Turbocharger", SAE Technical Paper 2013-24-0122, 2013.
[16] Zhao, D., Stobart, R., Dong, G., Winward, E., “Real-Time Energy Management
for Diesel Heavy Duty Hybrid Electric Vehicles”, IEEE Transactions on Control
Systems Technology, Vol. 23, No. 3, May 2015.
[17] Zhao, D., Liu, C., Stobart, R., Deng, J., Winward, E., Dong, G., “An Explicit
Model Predictive Control Framework for Turbocharged Diesel Engines”, IEEE
Transactions on Industrial Electronics, Vol. 61, No. 7, July 2014.
[18] Zhao, D., Liu, C., Stobart, R., Deng, J., Winward, E., “Explicit Model Predictive
Control on the Air Path of Turbocharged Diesel Engines’, American Control
Conference, 2013.
[19] Zhao, D., Stobart, R., "Real-Time Optimal Energy Management of Heavy Duty
Hybrid Electric Vehicles", SAE International Journal of Alternative Power. 2(2):369-
378, 2013.
[20] Zhao, D, Winward, E., Yang, Z., Rutledge, J., Stobart, R., “Control-Oriented
Dynamics Analysis for Electrified Turbocharged Diesel Engines”, SAE Technical
Paper 2016-01-0617, 2016.
[21] Zhao, D., Winward, E., Yang, Z., Stobart, R., “Decoupling Control of Electrified
Turbocharged Diesel Engines”, American Control Conference, 2016.
Acknowledgements
This work was co-funded by Innovate UK (formerly the Technology Strategy Board
UK), under a grant for the Low Carbon Vehicle IDP4 Programme
(TP14/LCV/6/I/BG011L). Innovate UK is an executive body established by the
United Kingdom Government to drive innovation. It promotes and invests in
research, development and the exploitation of science, technology and new ideas
for the benefit of business - increasing sustainable economic growth in the UK and
improving quality of life.
