Paper Number

24.4

Evaluating a Pump Controlled Open Circuit Solution

Kim Heybroek, Jan-Ove Palmberg
Fluid and Mechanical Engineering Systems, Linköping University, Sweden

Johan Lillemets, Martin Lugnberg, Martin Ousbäck

Volvo Construction Equipment

ABSTRACT
In this article the authors have studied a hydraulic
system configuration where each actuator/supply system
comprises an electrically controlled variable
displacement pump/motor working in an open circuit
together with four separate electrically controlled valves.
Performance, operability and energy consumption are
evaluated in a wheel loader, first with its original load
sensing hydraulic system and then modified with a
completely new pump controlled hydraulic system.
Measurements published in this article demonstrate the
advantages and drawbacks of pump control in an open
circuit compared both to a pump controlled, closed
circuit solution and to a valve controlled, load sensing
system. Performance is evaluated by looking at the fuel
consumption, productivity, and operability of a wheel
loader. Theoretical calculations of energy efficiency
including the dominant losses in system components are
evaluated and validated. Fuel consumption is measured
and put side by side with measurements of the same
machine equipped with a load sensing hydraulic system.
INTRODUCTION
There are three energy related key benefits using an
open circuit pump controlled system instead of a
traditional valve controlled system in a wheel loader.
art. In this study the test results and experience of the
hydraulic system implemented in a medium-sized wheel
loader are presented. The emphasis in this study is to
compare the performance in open circuit pump control to
other pump controlled solutions as well as to
conventional load sensing technologies. The comparison
primarily covers energy efficiency and system
complexity level.
THE OPEN CIRCUIT SOLUTION CONCEPT
If the pump does not have a pre-defined high and low
pressure compartment, it is said to work in a closed
circuit. In the other case, when the pump only operates
against high pressure on one side, it is working in an
open circuit. The open circuit solution, in this article
referred to as OCS, comprises four separate valves
combined with a number of sensors on top of an
electronically controlled pump, used for either pressure
or flow control. Figure 1 illustrates the simplified
schematics of an implementation of the OCS on two
drives. The two functions are mechanically connected in
parallel via the pump shaft. There are three pressure
sensors for each drive, which is a prerequisite to achieve
all the modes of operation studied in this article. The
required controllers for the pumps and valves and a
supervisory control are also shown in the figure. The
OCS was first presented by K. Heybroek, J-O. Palmberg
in 2006(1).

1) No metering losses arise because of unequal

drive pressure levels, due to separate pumps for
each drive.

2) The pressure drop over valves is maintained at

Supervisory controller

a low level as to valves are not used for flow anti-cavitation

throttling at any great extent. This includes check valves

energy recuperation. p valve

contr.
p p valve
contr.
p
on/off and pressure
proportional transducers
3) The differential pressure resulting in a meter-in valves
open-circuit,
pressure drop, caused by the load sensing p pump
contr. cross-center
p pump
contr.

architecture is eliminated. pump

further
drives

Obvious drawbacks in a pump controlled system are the

greater number of components and the increased
Figure 1 - Simplified circuit diagram of the open
complexity in control. By studying the results from a
circuit solution implemented on two drives.
demonstration machine the advantages and drawbacks
of the open circuit solution is compared to state-of-the-
STATE-OF-THE-ART
Today’s mobile machines most often contain valve
controlled drives arranged in an open loop circuit. For
the purpose of saving energy the constant pressure
systems have often been replaced by load sensing
pumps and load sensing valves.
Regarding pump-controlled solutions, some concepts
have previously been developed by, for instance; Habibi,
S. and Singh, G.(2), Wendel, G.(3), Rahmfeld, R. and
Ivantysynova, M.(4),. For various reasons, not all of these
concepts are suitable for mobile applications.
LOAD SENSING SYSTEMS
In mobile applications the load sensing solutions have,
in the past, significantly reduced the energy
consumption. The load sensing concept renders a cost
effective and compact solution, as there is usually only a
need for one pump to power several cylinder drives. The
load sensing systems use variable pressure controlled
pumps, where the pressure reference is set by the
cylinder load having the greatest pressure magnitude.
However, in applications with several drives operating at
unequal drive pressure levels, the load sensing systems
still result in energy losses, referred to as metering
losses. A pressure difference over the valve is required
to create flow which yields a load velocity. This is
achieved by the pump controller, often
hydromechanically setting the supply pressure to the
load pressure plus an additional differential pressure,
here referred to as LS- p. The LS- p yields a pressure
drop over the valve and is thereby yet another source of
losses. Another inherent problem with the load sensing
solution is the throttling losses associated with meter-out
flow control instead of recuperating energy. The
schematic of a conventional load sensing regulator is
illustrated in Figure 2. In this study the load sensing
system is abbreviated LS-S.
Figure 2 - A load sensing pump with a conventional
hydromechanical regulator(5).
PUMP CONTROLLED SYSTEMS
The most competitive pump-controlled controlled
hydraulic concept comparable to the OCS is the closed
circuit solution, here referred to as CCS, presented by
Rahmfeld and Ivantysynova. In the circuit, the differential
volume of the asymmetric cylinder is balanced on its low
pressure side by a charge pump and an accumulator,
see Figure 3. Additionally, shut-off valves for load
holding in emergency situations, as well as high
pressure relief valves are optional. If several drives are
needed, these can be coupled via the low pressure side,
sharing the charge pump and the accumulator. The total
number of components can consequently be kept low. In
a similar way to a hydrostatic transmission this circuit
handles four-quadrant actuation of the asymmetric
cylinder in a simple hydromechanical manner, making
the solution robust and reliable. The hydraulic machine
works as pump or motor depending on the load
quadrant, automatically recuperating energy through the
pump shaft when possible. In construction machinery
the desired lowering speed can be comparatively high,
often twice as high as the lifting speed. This means that
the pump in a closed circuit must be dimensioned to
handle the lowering flow. The closed circuit was
implemented in an O&K wheel loader and evaluated and
compared to a standard valve controlled machine(6). The
measurements from the evaluation demonstrated a
reduction in fuel consumption of 15% in a truck loading
scenario. The energy recuperation capabilities are
claimed to be the main reason for that reduction.
control command
Q B, p B
5 61 91
31
EDC
2
LP
1 4
32 92
7
62
Q A, p A
Figure 3 - Closed circuit solution schematics(7).
TARGET APPLICATION
THE WHEEL LOADER
The wheel loader is a true multipurpose machine used in
all sorts of environments, Figure 4. However, the main
purposes of the machine are often simple; lifting,
lowering and transporting loads. To achieve this, both
working hydraulics and a propulsion system are
essential. These two systems dominate the useful
energy consumption to a large extent the non-
useful energy consumption in the wheel loader. One
reason why the wheel loader was chosen as the target
application of this study is the good possibility to save
energy in the working hydraulics by energy recuperation
from lowering motions, energy which today is dissipated
as heat.
 control mode. In the OCS-prototype the pumps are
mounted in a tandem configuration shown in Figure 6.

Figure 4 - The target application is a medium sized

wheel loader.

Figure 6 - Electronically controllable pumps, 100%

THE OPEN CIRCUIT SOLUTION PROTOYPE
displacable in both directions, mounted in tandem.
The OCS is customized to fit onto a medium sized retail Valves – A valve package for the lift and the tilt drive
wheel loader. The old load sensing system is still was manufactured, consisting of eight custom made
present, but is now only used for steering. The original seat valves, illustrated in Figure 7. The valves used for
85 cc pump was replaced with two 28 cc pumps, which meter-out flow control are proportionally controlled and
is more than enough for steering. For the working electrically pressure compensated Valvistor® valves.
hydraulics, two 75 cc electrically controlled open circuit
pumps were added to the power take-out (PTO). A
software priority function resembles the system behavior
of the reference machine in prioritizing the tilting drive.
Also, a torque limitation is implemented to justify the
comparison with a system with less installed hydraulic
power. The working hydraulic LS-valve is replaced with
a new valve block, with separated sections for the lift
and the tilt functions. The powertrain configuration and
auxiliary hydraulics left unmodified. An overview of the
hydraulic system is illustrated in Figure 5.

Figure 7 - Custom made valve block containing eight

proportionally controlled poppet valves.

Tank pressure enhancement – When oil is taken from

the tank the pressure must be enhanced by about 5 bar
to avoid cavitiation. In the OCS demonstrator this is
achieved by the use of a pressure relief valve, set to a
fixed cracking pressure, alternative a) in Figure 8. In
alternative b) the check valve is replaced with an
electrically controlled on/off valve. Alternative c) is a
third solution, where the tank is a closed system in which
the pressure has been enhanced by other means.

Alt. a) Alt. b) Alt. c)

Plift Tcommon Ptilt Plift Tcommon Ptilt Plift Tcommon Ptilt

Figure 5 - Schematics of the hydraulic system of the

OCS prototype wheel loader. pT,enhanced

Pumps –Two open circuit pumps power the lifting and

tilting drives respectively. These pumps are electrically Figure 8 – Optional solutions to tank pressure
controlled and are operated either in flow or pressure enhancement.
Sensors – Three pressure sensors are used for each pressure sensor failure and the auxiliary control loop
drive, where one is integrated in the pump. In addition to handles load holding functionality. In the actual OCS
pressure, shaft speed and swashplate angle are also prototype, steps 2b, 4 and 5 are at this stage not yet
measured. implemented.

Control system – The software used for control and data

acquisition is implemented in a real-time controller unit Mode
Selection
Power
Management
Mode
Transition
operating at a sample rate of 100 Hz.
1 4 6
Load quadrant Power-need Define method for
CONTROL STRATEGIES definition evaluation pressure matching

This section describes the desired features of a

2 5 7
Mode efficiency Revise selected Supply pressure
controller for the open circuit solution. The basic evaluation control mode build-up

controller workflow is shown in Figure 9, where the steps 3 8

are: Preliminary mode

selection
Deploy valve
setting

1) Identify the present load quadrant by looking at

load pressure levels and operator command Pump control Safety functions control Auxilliary control

signals.
Figure 9 - Schematics of a proposed controller for
2) The present load pressure and the desired
actuator velocity are used in an algorithm to the OCS when used in a wheel loader.
select the best mode with regards to energy
THE IMPACT OF MODE SELECTION
efficiency. The main components in such
evaluations for recuperative motions are:
The four valves render a solution versatile in control, as
the cylinder chambers can be connected to pump and/or
a. Minimize losses in valves comparing
tank as well as be closed at any time. In previous
different control modes.
studies the authors has investigated what influence
alternative control modes have on energy efficiency(8).
b. Minimize losses in components such as
What is here referred to as “control mode” is the way of
pumps, cylinders and hydraulic lines, have
controlling the load at a requested velocity, i.e.
comparing different control modes.
lifting/lowering a hanging load is achieved either in non-
differential or differential mode, either with or without
3) A preliminary mode is selected for further
meter-out flow control. The objective is to recuperate the
evaluation.
most energy possible for a commanded specified motion
and still fulfill the lowering speed requirements. Ideally,
4) Investigate if there is any immediate need for
the recuperated energy equals the energy required to lift
energy to be recuperated and if the workload on
the load; in reality, this ratio is much lower due to losses.
the engine can be reduced.
The losses are mainly related to the efficiency of
cylinders, valves and pumps(7).
5) Find an alternative control mode that matches
the power need, go back to 4 and iterate over
Measurement results from the OCS prototype show the
the mode selection until a match is found.
differences in potentially recuperable energy in different
modes of operation. In the non-differential and
6) Given the selected mode, define rules for how to differential control mode the maximum lowering speed is
match pump and load pressure prior to opening restricted by the maximum pump flow. If meter-out
valves. control is used the lowering speed is restricted by the
flow taken through the meter-out valve. In tests where
7) Command a pump pressure defined by 6. meter-out is used the mean lowering time is specified as
3 seconds. All tests are performed with the diesel
8) Execute valve settings defined by the selected operating at idle speed, 700 rpm. The efficiency in
control mode. recuperation is here referred to as the recuperation ratio,
defined in Equation 1:
Depending on the chosen mode of operation the pump
must be controlled differently. This is handled by the
pump controller which has the mode and operator
E recup
η recup = (1)
joystick signal as input. The control signals are sent to Ein
the pump, which has its own distributed electrical control
system, regulating the swashplate angle. The control
where Erecup is the recuperated energy output on the
signal for the pump is either relative displacement or
pump shaft and Ein is the energy input on the pump shaft
pressure. A safety control loop handles for example
powering the hydraulics.
Non-differential lowering Differential lowering

Figure 10 illustrates the valve setting and operating Figure 13 illustrates the valve setting and operating
region of the control mode “non-differential lowering”. region for the “differential lowering”.
v v
F F F F

Differential
Normal working region
working region

open open open

qp pp vn* v qp pp vd* v

Figure 10 - Left: valve configuration for non- Figure 13 - Left: valve configuration for differential
differential control. Right: working region for mode. control. Right: working region for mode.
As the entire load flow is controlled by the pump in non- Due to the differential coupling the limited maximum
differential lowering, no throttling losses related to meter- allowable load force in this mode is decreased.
out control are present (the left efficiency plot in Figure Exceeding this limit requires pressure control in order to
11). The pump/motor efficiency combined with the valve stay below maximum system pressure(10). The entire
losses in non-differential mode is illustrated to the right load flow is still controlled by the pump; the left plot in
in Figure 11. Figure 14 shows that no throttling losses related to
5 System efficiency 5 System efficiency
meter-out control are present, although higher lowering
speed is achieved.
x 10 x 10
0.8

0.6 0.5 0.7

0.6

0.3
0.9
0.8

0.7
0.6 0.5
0.2 0.1

η η
0.7
0.6

0.10.2
0.5
0.4

6 6
0.3

0.2
1

0.4 0.3
0.7
0.40.3

0.4
0.5

5 5 5
x 10 System efficiency x 10
5 System efficiency
0.2

0.4

0.4
0.8 0.7

0.8
η

0.1
4 4 0.

0.2
0.3
6

0.7
0.5
0.6
6 2

0.2

0.2
0.8

0.3

0.3
0.7

0.6
0.9
0.8

0.3
0.7
0.6 0.5
0.2 0.1
F [N]

F [N]
0.7
0.6

0.10.2

0.3
0.5
0.4

0.9
1

0.6
0.3

0.2

0.5
0.3
0.5

0.7

3 3
0.8

5 5
0.5
0.40.3

0.4
0.5
0.4
0.6

0.4
0.6

0.1
0.9

0. 0.2
2 2 4 0.3 4
4
0.2

0.7 0.5 0.3

0.
1

0.40.2
0.8

F [N]

0.80.7
F [N]

4
0.1
0. 0.8 0. 0 .

0.3
0.5
0.6
7 6

0.3 2
7 0.5 0.
0.8

1 1
0.9
0.8

0.6

0.3
0.7
0.6

0.
0.2

3 0.9 0.9 3 6
0.7
0.6

0.2 0.1
1

0.4
0.5
0.4

0.5 0.7
0.3

0.2

0.5 1 0.4 0.2 0.1

0.

0.4
0.3 0.1
3

0.

0.5
0 0

0.
0.7 0.6
2 2

0.4
0 0.1 0.2 0.3 0.4 0.5 0 0.1 0.2 0.3 0.4 0.5

3
0.4
0.5
0.1
0.9
V [m/s] V [m/s]
0.8

0.8
1

0.6

1 1
0.7

00..320.4
0.5

0.8

0.1

0.7
0.7
0.6 5

0.06.5
0.

0.2

0.6
0.9

0.4 0.3
0.2 0.3 0.1
Figure 11 - Left: efficiency including valve losses.
0.50.1 0.4
0 0 0.1 0.4 0.2 0.3
0 0.1 0.2 0.3 0.4 0.5 0 0.1 0.2 0.3 0.4 0.5
V [m/s] V [m/s]
Right: efficiency including valve and pump losses.

Lowering loads in non-differential mode is associated Figure 14 - Left: efficiency including valve losses.
with unacceptable lowering times, almost 15 seconds. In Right: efficiency including valve and pump losses.
Figure 12 the power and energy are illustrated for the Lowering the load at the same engine speed as in the
mode. The bar to the right illustrates the recuperation non-differential control mode the operation now takes 6
ratio calculated from measured pump pressure, relative seconds instead of 15. However, the recuperation ratio
pump displacement and calculated pump/motor is lower than for the non-differential control mode,
efficiency. The average recuperation ratio is 36%. averaging 30%.

Power and energy ηrecup

Power and energy ηrecup 1.0 1.0 1
1.0 1.0 1
0.9
0.75 0.75
0.9
0.75 0.75
0.8
0.8 0.5 0.5
Normalized Energy [-]
Normalized Power [-]

0.5 0.5 0.7

Normalized Energy [-]
Normalized Power [-]

0.7
0.25 6s 0.25
0.6
0.25 15 s 0.25
0.6
0 0 0.5
0 0 0.5
0.4
-0.25 -0.25
0.4
-0.25 -0.25 0.3
-0.5 -0.5
0.3
Power 0.2
-0.5 -0.5
0.2 -0.75 -0.75
Power Energy 0.1
-0.75 -0.75
Energy 0.1
-1.0 -1.0 0
0 5 10 15 20 25 30
-1.0
0 5 10 15 20 25 30
-1.0 0 Time [s]
Time [s]

Figure 15 - Power, energy and recuperation ratio.

Figure 12 - Power, energy and recuperation ratio.
Non-differential lowering + meter-out Differential lowering + meter-out

Figure 16 illustrates how increased lowering speed is Figure 19 illustrates how increased lowering speed is
achieved by meter-out flow control. achieved by meter-out flow control.
v v
F F F F
Fn*
q-contr. Normal + meter-out q-contr.
working region Differential + meter-out
working region

open open open

qp pp qp pp vd* * v
vmax
vn* v

Figure 16 - Left: valve setting for non-differential + Figure 19 - Left: valve configuration for differential +
meter-out control. Right: working region for mode. meter-out control. Right: working region for mode.

In order to achieve the required lowering time of 3 The differential mode is not sufficient to fulfill the
seconds, flow must be throttled over the meter-out requirement of 3 seconds in lowering time. However, in
orifice, whereas the loss related to control mode this mode half of the flow is throttled to tank. Again, the
increases dramatically. Only a fifth of the lowering flow pump losses are considered in combination with the
goes through the motor, the rest is throttled to tank. The losses associated with the control mode. The resulting
efficiency for this strategy is illustrated in Figure 17. efficiency is presented in Figure 20. It should be noted
that the operating conditions for the motor have changed
5
x 10 System efficiency x 10
5 System efficiency due to the transformation in pressure and flow that come
with this mode.
0.3

0.7
0.8

0.4

η η
0.7
0.3
0.8
0.6

0.3
0.7

0.6

0.5
0.4
0.2

6 6
0.5
0.4

0.3

0.2
0.9

0.1 0.2
0.7
0.6
0.5

0.6
0.1
1

0.4 0.5

0.2

5 5

5
x 10 System efficiency x 10
5 System efficiency
4 4 0.

0.8
0. η 3
η
1 0.70.9

0.6
0.1
0.3

2
0.8

0.4

0.7

0.5
6

0.1
F [N]

F [N]

0.7
0.3

0.4
6
0.8

0.3

0.20.3
0.3
0.6

0.7

0.
0.7

0.6

0.5
0.4
0.2 0.1
0.5

2
0.4

0.3

0.2
0.9

0.2 0.1 0.4

0.7
0.6
0.5

0.8 0.6

3 3
0.6

0.5
1

0.2
0.3
5 5
0.5

0.2
0.9

0.8

0.5

2 2 0.6 0. 0.
0.3 0.2

0.4
4 3

0.4
4 4
0.7 0.5

0.9
0.8
F [N]
F [N]

0.7

0.6
1

1 1

0.1
0 . 0.

0.8
0.3
0.4

0.5
0.8

0.1
0.6

0.4
0.20.3
0.3

7 8

0.7
7 5 4
3

0.0.5 0.2 0 . 0.
8

0.6 0.5
0.6

0.9 3
0.2

0.

0.2 3
0.

0.1 0.9
0.2
0.4

0.3

0.4 0.3
0.9
0.6
0.5
0.7

0.7

0.6
0.1

0.5
1

0.6
0.1 0.1

0.3
0.2

0.4
0 0
0 0.1 0.2 0.3 0.4 0.5 0 0.1 0.2 0.3 0.4 0.5
2 2
V [m/s] V [m/s]

0.5
0.4
0.6
0.5
0.7

0.7
1 1 0.8

0.8

0.7
1

0.1
00.0.5.6
0.7
0.8

0.5
0.9

0.1
0.6

Figure 17 - Left: efficiency including valve losses.

0.2
0.4

4
0.3
0.4

0.3 0.3
0.2 0.4 0.6 0.2
0.5 0.3 0.1 0.2
0 0 0.1
0 0.1 0.2 0.3 0.4 0.5 0 0.1 0.2 0.3 0.4 0.5

Right: efficiency including valve and pump losses. V [m/s] V [m/s]

The limit where meter-out control becomes necessary Figure 20 - Left: efficiency including valve losses.
naturally depends on pump size, engine speed and Right: efficiency including valve and pump losses.
piston area. In Figure 18 the bar to the right shows the
recuperation ratio for the lowering motion. The average The calculated recuperation ratio is roughly double for
ratio over three cycles is 8%. The achievable lowering the differential + meter mode compared to the non-
speed in this mode is a trade-off in recuperated energy differential + meter-out mode. The average ratio
versus lowering time. calculated over three cycles is 16% with a lowering time
averaging the targeted 3 seconds.
Power and energy ηrecup
1.0 1.0 1
Power and energy ηrecup
0.9 1.0 1.0 1
0.75 0.75
0.9
0.8 0.75 0.75
0.5 0.5
Normalized Energy [-]

0.8
Normalized Power [-]

0.7
0.5 0.5
Normalized Energy [-]
Normalized Power [-]

0.25 0.25 0.7

2s 0.6
0.25 0.25
2s 0.6
0 0 0.5

0 0 0.5
0.4
-0.25 -0.25
0.4
0.3 -0.25 -0.25
-0.5 -0.5
0.3
0.2
Power -0.5 -0.5
-0.75 -0.75 0.2
0.1 Power
Energy
-0.75 -0.75
0.1
-1.0 -1.0 0 Energy
0 1 2 3 4 5 6 7 8 9 10
Time [s] -1.0
0 1 2 3 4 5 6 7 8 9
-1.0
10
0

Time [s]

Figure 18 - Power, energy and recuperation ratio.

Figure 21 - Power, energy and recuperation ratio.
Conclusions from the mode selection evaluation: Heavy material is loaded and lifted up slowly to a truck,
followed by rapid lowering in differential mode.
It is important to note that all calculations are based on
measurements from a lifting and lowering operation with
the wheel loader standing still at idle engine speed. At
low speed the hydraulic machine has a comparatively
high internal leakage, resulting in poor efficiency both in
the lifting and the lowering phase.

The fact that the recuperation ratio is higher in the non-

differential mode than in the differential mode might
appear surprising. This is a result of the additional
pressure losses apparent in the differential mode due to
its extended flow path via several sharp bends and
edges, illustrated in Figure 22. If a dedicated valve,
connecting the cylinder chambers, is installed closer to
Figure 23 - Mode selection in different fields of
the load, the losses in differential mode will be reduced.
application.
Moreover, the increased pressure level in differential
mode affects the cylinder efficiency, but to what extent
has not been further investigated within the framework of
this study. SYSTEM PERFORMANCE

v v The goal of this study is to evaluate the OCS with a

F F degree of emphasis on energy efficiency and compare
Normal Differential the system to a conventional load sensing system. In
Flowpath Flowpath order to do so several tests were conducted. A first
simple test is to perform a single function efficiency test,
then proceed to a more complex loading cycle efficiency
open open open test, finally looking at a long distance efficiency test.

qp pp qp pp The performance of the new hydraulic system is

Figure 22 - Flow path in non-differential and
differential mode.
Important in this comparison is how great a recuperation
ratio can be achieved for a given speed and force couple
• If v < vd* a higher recuperation ratio is achieved
in differential mode than non-differential mode +
meter-out for the same speed
evaluated by comparing various measurement results of
two wheel loaders following the same assignment.
Besides the modification, the machines must be
equivalent to ensure a credible comparison between the
two. The machine chosen as reference machine is of the
same brand and model as the new prototype with
identical specifications as to the working hydraulics.
Figure 24 shows the simplified schematics of the
reference machine’s hydraulic system.

• If v < vn* operation in non-differential mode

yields a higher recuperation ratio than
differential mode

The differential meter-out mode is more energy efficient

than the corresponding non-differential mode, which is
probably a result of the higher pump efficiency in that
case. At load forces exceeding Fd* the differential mode
is not an option unless pressure control is applied.
However, the limits in speed and force are determined
by the maximum pump size and cylinder area properties
that can be optimized given a specific field of
application. An example of how different modes can be
used to achieve a broader working region with
maximized energy recuperation is illustrated in Figure
23. In case 1 the wheel loader is used for material
handling. High elevation is reached rapidly and loaded
material lowered carefully in non-differential mode. In Figure 24 - Simplified hydraulic system layout of the
case 2 the wheel loader is used in a truck loading cycle. reference machine.
EFFICIENCY TEST - SINGLE FUNCTION
It is interesting to compare the theoretical expectations
as regards hydraulic energy reduction from energy
recuperation to the reduction in fuel consumption in
practice. The test is performed by simple lifting and
lowering a defined load a specified number of times,
measuring the fuel tank level just before and after the
operation.
In order to compare the results between the two
machines it is necessary to mention a few things about
the energy distribution in the wheel loader during this
specific test. Only one function is used at a time
throughout the test. The significant difference between
the OCS-prototype and the reference machine is that
energy is recuperated by the OCS machine. Assuming
the hydraulic system to be the dominant consumer of
useful diesel energy and the losses in the lifting phase
are equal for the two machines; the reduction in fuel
consumption depends only on how much energy is
recuperated. The percentage reduction is thereby also
directly comparable to the recuperation ratio described
in the previous section. The energy flow through the
reference wheel loader is illustrated to the left in Figure
25. First, the fuel energy is transformed into heat and
mechanical energy in the diesel, which is then
transformed into hydraulic power which is used to
actuate the load. When lowering the load again the
potential energy is transformed into heat in the hydraulic
valves. No work has been achieved. To the right is the
corresponding energy flow through the OCS prototype
loader.
LS-S ELS-S,fuel OCS »ELS-S,fuel (1-erecup)
Diesel
Ed,loss Ed,loss
(» 0.6 ELS-S,fuel) PTO (» 0.6 EOSC,fuel)
Eh
hyd.
Eh,other
Eh,loss
Figure 25 - Illustration of energy saving from
recuperation in the single function efficiency test.
Each cycle of lifting and lowering lasts for 20 seconds
and is repeated 15 times for each machine. The test is
performed with two different loads: an empty bucket and
a weight bucket. The lowering time is the same for both
cases and takes about 6 seconds. The measurement
results are presented in Figure 26.
1,2
Fuel consumption / max fuel consumption
LS-S
Normalized fuel consumption [-]
OCS
1 19.2%
decrease
0,8
23.3%
0,6
decrease
0,4
0,2
0
Heavy w height bucket Empty bucket
Figure 26 - Single function efficiency test results.
The result with weight bucket is a 19.2% improvement
compared to the reference machine. The result with an
empty bucket is a 23.3% improvement. The weight
bucket test is performed in non-differential + meter-out
control mode and the empty-bucket test in differential
mode + meter-out. Relating to the theories presented in
the previous section the test with weight bucket should
be compared to the scenario where non-differential
lowering with meter-out is used. However, in this case
the test is conceived with a greater load and a lower
speed, resulting in lower throttling losses, and thereby a
higher recuperation ratio. The empty bucket test should
be compared to the differential + meter out control
mode. Also during this test, less flow is throttled to tank
compared to the scenario described in the previous
section, resulting in a recuperation ratio closer to “pure”
differential lowering.
EFFICIENCY TEST – TRUCK LOADING CYCLE
To evaluate the working hydraulics efficiency in a typical
field of operation for the wheel loader, a truck loading
Diesel cycle is considered. Operation in the truck loading cycle
is often the main purpose of these machines; therefore,
the cost effectiveness of the cycle has economic
»Eh
significance for the entrepreneur. The cycle illustrated in
PTO
Eh,recup= Figure 27 lasts for about 30 seconds and includes the
Eh
Eh erecup machine going into a pile of granular material, lifting a
hyd. full bucket, reversing out from the pile, forwarding to
Eh,other
empty the bucket onto a truck, going back out while
Eh,loss lowering the empty bucket and then starting the next
cycle. The energy distribution in the truck loading cycle
differs considerably from the single function test. In this
cycle the working hydraulics, transmission, brakes and
steering interact continuously with each other. What
remains the same for both systems is the actual work
being done during the loading cycle. The powertrain,
including the propulsion system, is also the same for
both systems. In the reference wheel loader about 50%
of the diesel engine’s output power is used by the
propulsion system; the rest is used by the hydraulics.
 Pother = Pidle,loss + Pcooling + Pcyl . fric.,loss + Pline,loss (5)

The power losses integrated over the loading cycle yield

the total energy loss for the load sensing hydraulic
system.

cycle t .

E loss .tot − ref = (P p ,loss + Pthrottle + PLS − ∆p + Pother )dt (6)

t =0

The energy distribution in the reference wheel loader is

presented in Figure 28. The distribution is based on
Equations 2-6, using measurement data from a truck
loading cycle.

LS-S
ELS-S,fuel

Ein,LS-S

Diesel hydraulic
Figure 27 - The short loading cycle is a typical field system
of application for the wheel loader(9).
Ethrottle+metering losses
Losses in the reference wheel loader: Ed,loss Epump losses
PTO ELS-Dp losses
(» 0.6 ELS-S,fuel)
What really differ between the two systems are the 0.5 0.5 Ework Eother losses
losses in the hydraulic system. The predominant losses prop. hyd.

in the hydraulic system of the reference machine are Ep,loss Eh,loss

listed below: Ewheel Ework

The pump losses are calculated from measured Figure 28 - Energy distribution in the wheel loader
pressure, shaft speed, relative displacement and during the truck loading cycle, using the load
mapped pump efficiency. sensing system.

1 Losses in the OCS-prototype wheel loader:

Ppump ,loss = ∆p P ,i ⋅ ε p , i ⋅ D p ,i ⋅ n p ⋅ −1 (2)
i = drive η p ,tot
For the OCS prototype losses in the hydraulic machine
working as a motor, occur during the recuperative
The throttling losses are calculated from the energy motions, calculated from measured pressure, shaft
dissipated over meter-out orifices during regenerative speed, and relative displacement and mapped
motions measuring actuator speed and pressure drop efficiency.
from load to tank side.
Pmotor ,loss = ∆p m ,i ⋅ ε m , i ⋅D p ,i ⋅ n p ⋅ (1 − η m ,tot ) (7)
Pthrottle = Ai ⋅ x p ,i ⋅ ∆p ( L −T ),i (3) i = drive
i = drive

However the PLS- p term is no longer valid and the Pthrottle

The LS- p losses are calculated from the sum of term is significantly reduced due to the possibility in
pressure losses in the load sensing valves using recuperation. The Pother element has increased due to
measured actuator speed and pressure drop from pump increased idling losses even though the cooling term
to load side. decreases with the decrease in total losses. The total
hydraulic loss for the OCS prototype is given by Eq. 8.
PLS − ∆p = Ai ⋅ x pi ⋅ ∆p ( P − L ),i (4)
i = drive cycle t .

Eloss.tot −OCS = (P p ,loss + Pm ,loss + Pthrottle + Pother )dt (8)

All other hydraulic losses are collected into the term t =0
Pother in this text listed in order of importance; energy
required for cooling, idle losses in pumps, cylinder The energy distribution in the OCS-prototype is
friction losses and line losses. presented in Figure 29.
 OCS 5 System efficiency ηrecup
x 10
EOCS,fuel 0.3 1

0.2
0.6
η

0.3
0.4
6 0.9

0.3 0.5
Ein,OCS » 0.8 Ein,LS-S

0.7
0.8

0.1
0.8

0.2
0.
5 3
0.4

0.9
0.7

0.1

0.5 0.4
Diesel hydraulic Erecup 4 0.6
system 0.5

0.2
F [N]
0.4

0.6 0.7

0.3
0.6

0.4
0.5

0.3
3

0.8
0.7

0.1
0.5
0.4

0.1 0.2

0.6
0.7
Emotor losses

0.5
2 0.9 0.3

0.4
0.6
Epump losses
Ed,loss 0.2
PTO Ethrottle losses

0.5

19%
0.6
(» 0.6 ELS-S,fuel) 1

0.2
0.3
Eother losses 0.7 0.1

00. 0. .6
Ework 0.8 0.7

0.7
34
0.5 0.6 0.5 0.5
0 0.1 0.2
0.1 0.4 0.2 0.3 0.4 0.2 0.3
0.1
0.1 0
prop. hyd. -0.1 0 0.1 0.2 0.3 0.4

Ep,loss V [m/s]
Eh,loss
Ewheel Ework
Figure 31 - The energy recuperation ratio in the truck
Figure 29 - Energy distribution in the wheel loader loading cycle.
during the truck loading cycle, using the open circuit
solution. Lowering an empty bucket in the differential mode is not
a problem, but regarding differential operation of the
Switching to the differential control mode in the tilting drive, a full bucket of gravel unfortunately often
recuperative phase yields a different point of operation in results in too high a pressure level. However, choosing
pressure as well as displacement. This means that the the cylinder area ratio differently a higher recuperation
pump efficiency at a given speed and force couple ratio could also be achieved in this drive. In Figure 32
depends on which control mode is applied. In Figure 30 the power, energy and recuperation ratio throughout the
the pump efficiency during the recuperative phase of the loading cycle are illustrated, including both drives.
loading cycle is presented, operating in the two modes ηrecup
Power and energy
respectively. The efficiency is found through 1

interpolation in measurement data delivered by the 0.9

pump manufacturer where pressure level, relative 0.8

displacement and shaft speed have been considered in

Normalized Energy [-]

Normalized Power [-]

0.7
the calculation. As seen in the figure the pump efficiency 0.6
is 5-10% higher in the differential control mode. 0.5
Lift
Tilt 0.4

1
0.3
Normal state
Differential state 0.2
0.8 Power
0.1
Energy
0
0.6 0 2 4 6 8 10 12 14 16 18
ηm,tot

Time [s]

0.4

Figure 32 – Power, energy and recuperation ratio.

0.2

The energy consumption during a truck loading cycle is

0
62.5 63 63.5 64 64.5
Time [s]
Figure 30 - Pump efficiency in the recuperative
phase of the loading cycle depending on chosen
mode of operation.
The force and speed characteristics of the lifting drive in
one typical loading cycle is illustrated in Figure 31. The
speed and force data is derived from pressure- and
position signals, acquired during operation. The contour
plot in the background shows the system efficiency. The
efficiency calculation encompasses pump losses as well
as valve losses given the best possible control mode.
The bar to the right shows the recuperation ratio for the
truck loading cycle, in this case 19% of the hydraulic
energy input for the lifting drive throughout the cycle.
About the same percentage applies for the tilting drive,
resulting in an average of about 20% of the energy input
being recuperated throughout one cycle.
65 65.5 66 66.5 67 measured according to Volvo standard fuel
measurement procedures. The two machines were
driven by an experienced driver. The tests were
performed during the same day, loading gravel from the
same spot. For every six cycles the truck was fully
loaded and weighed. Each machine was driven until four
truckloads were filled. Each cycle was clocked in order
to measure productivity. The measurement results are
presented in Figure 33. The left staple pair shows a 9%
reduction in consumed fuel volume per loaded mass of
gravel, representing how the efficiency in pure
Newtonian work is affected. The middle staple pair
shows how the system saves 13% in fuel volume over
time. The right staple pair shows a 2% longer average
cycle time. The error bars on top of the staples show the
maximum and minimum measurement value throughout
the tests. The reduced size of the error bars in
measurements from the OCS indicates that the
operability of the machine is good.
 Perform ance LS-S EFFICIENCY TEST - TRANSPORTATION DRIVING
1,2 2% OCS
9% 13% longer
1 less less In both wheel loaders the hydraulic pumps are directly
powered from the power take-out unit which is
Normalized individually [-]

0,8 mechanically coupled to the diesel engine.

0,6
Consequently the pumps are in motion as long as the
engine is on. In the reference machine there is only one
0,4 pump powering the working hydraulics, steering and
auxiliary functions. In the OCS two pumps are used for
0,2
the working hydraulics and for the steering and auxiliary
0
functions, two small pumps had to be installed instead of
fuel volume/loaded mass fuel volume/time cycle time one big pump, due to space limitations. The result is four
pumps instead of one. To measure how the increased
number of pumps affects the energy consumption, a
Figure 33 - Performance measures from the truck
transportation driving test was arranged. The distance
loading cycle.
driving test was performed on an oval test track at the
Volvo Construction Equipment testing ground. The
engine speed was set to 1600 rpm and the gearbox is
The comparison elucidates some important energy
locked in the fourth gear for both test objects. The test
related benefits and drawbacks of the two systems,
was carried out during 15 minutes of driving. The
which are summarized in Table 1.
measured difference in fuel consumption is presented in
Figure 34.
Table 1 - Apparent energy related aspects by
introduction of the OCS and their importance.
LS-S
Effected energy aspects Impact on fuel 1,2 Fuel consum ption in distance
5.4% more fuel OCS
by introducing the OCS consumption in a
1
hydraulic system truck loading cycle
Fuel consumption [-]

0,8

Elimination of p-margin Medium impact – the

0,6
related to LS-control pumps are mostly able
to keep this margin 0,4
except at maximum flow
0,2

Elimination of metering Low impact – because 0

losses related to unequal the drives are actuated LS-S OCS

drive pressure levels, simultaneously only to a

introducing separate very small extent
pumps for each drive
Reduction in energy Low/medium impact –
required to cool the oil reducing the total
being throttled to tank in energy consumption
lowering motions as well implies less energy
as cooling of engine required for cooling of
losses (in case of saving) diesel losses
Introduction of increased Low/medium impact –
losses related to the increased viscous
presence of extra pumps friction losses according
to next section
Introduction of energy High – if the best control
recuperation in lowering mode is chosen
motions according to previous
section
Figure 34 - Fuel consumption in a transportation
driving test comparing the OCS prototype machine
to the reference machine.
OPEN VERSUS CLOSED CIRCUIT SOLUTION
In measurements from 2004, the CCS demonstrated a
fuel consumption reduction of 15% in a truck loading
scenario similar to the one described in this article. Why
the energy reduction presented in this study is lower
naturally depends on many factors. One probable
reason is the conceptual differences in the propulsion
system. The CCS prototype uses a hydrostatic
transmission whereas the OCS prototype uses a
hydrodynamic transmission. In a system where the
propulsion requires less energy, the impact of an
improvement in the hydraulics is greater. A better way of
comparing the performance of the two solutions is to
look at the capabilities of the individual systems side by
side without a specific application involved. In Table 2
the overall capabilities and thereby similarities, strengths
and weaknesses of the systems are listed.
Table 2 - Quality measures.
Quality measures OCS CCS
Four-quadrant operation Yes Yes
Energy recuperation Yes Yes
Recuperation in differential mode Yes No
Pressure sensors needed Yes No
Charge pump needed No Yes
Accumulator needed No Yes
Based on closed loop control No Yes
Increased tank pressure needed Yes No
Regarding the differences in hardware requirement, the
aspects of component integration must be considered.
Both concepts require pressure relief functionality, anti-
cavitation and load holding functionality. In the CCS
almost all the components can be integrated to the
pump housing, including charge pump, pressure relief
valves, check valves and load-holding valves. For the
OCS it is suggested that a distributed valve configuration
be used, mounted as close as possible to the cylinder
drives. However, in the valve configuration, pressure
relief functionality anti cavitation and load holding
functionality are easily incorporated(10).
Regarding four-quadrant actuation, the CCS intrinsically
operates over all four quadrants; whereas the OCS
operates over only two quadrants. To make the open
circuit solution controllable over all four quadrants, the
four valves must be controlled actively. This of course
makes control more complex, but it also results in
greater flexibility when it comes to the selection of
modes targeting the highest possible efficiency.
In lifting operations both concepts are more or less
equally energy efficient. The dominating loss component
in this quadrant lies in the pump efficiency. In lowering
operations the OCS is capable of operation at higher
efficiency over a larger region of operation, due to the
possibility of operating in the differential control mode.
The CCS cannot recuperate energy in the differential
mode as a hydraulic coupling of the cylinder chambers
results in equal pressure on both sides of the hydraulic
machine, whereas no torque is generated.
For the wheel loader in this study the CCS would require
much larger pumps to achieve the specified lowering
time of 3 seconds. The flow leaving the lift cylinder is
about 4*10-3 m3 which in terms of displacement
corresponds to a 350 cc pump at the given engine
speed. Another almost as big pump is required for the
tilting motion. Having such big pumps when only one
third of that size is sufficient to achieve the specified
lifting times, does not render the solution cost effective.
FUTURE CAPABILITIES
The open circuit solution is not restricted to use only with
the diesel engine as primary mover, but would actually
benefit from the presence of an electro-hydraulic hybrid
power system. In such a system the diesel engine is
replaced by electrical motors to power each drive
independently. Moreover, a cost effective solution would
be to replace the variable pumps with smaller, high
speed, fixed displacement pumps, independently speed-
controlled by the electrical motors. This would increase
the recuperation ratio due to greater maximum flow in
the differential mode and consequently reduce the
throttling losses. Moreover, a prerequisite to save
energy in the OCS prototype is that the recuperated
energy is consumed instantly. In an electro-hydraulic
hybrid system with energy storage this would not be an
issue. An alternative to the meter-out flow control would
be to use the excess flow to boost the hydraulic fan
motor. In future development of this type of mode-
switching systems one should always consider the
differential mode when selecting the cylinder area ratio
in order to fully benefit from using the potential load
energy in recuperative quadrants on all cylinder drives.
CONCLUSION
The calculated results from the single function efficiency
test showed that the non-differential mode is more
energy efficient than the differential mode, but the
differential mode soon becomes a better alternative if
meter-out flow control is needed in order to achieve the
required lowering speed. The test showed how the open
circuit solution, within reasonable limits in lowering
speed, saves about 20% fuel compared to the reference
machine, equipped with conventional load sensing
hydraulics. In the truck loading experiment fuel
consumption was reduced by about 10% compared to
the reference machine. According to the theoretical
calculations this result corresponds to a 20% reduction
of losses in the OCS hydraulic system. In a final test the
OCS-prototype was driven a longer distance without
using the hydraulics. The result showed how the
increased number of installed pumps increases the fuel
consumption by about 5%. Comparing the open circuit
solution to the closed circuit solution, the OCS has a
greater potential for energy recuperation over a broader
working region while fewer sensors and more active
control are required for the closed circuit. Concerning
operability, where smoothness in control and system
response is important for a wheel loader application,
many of these parameters are easily adjusted in
software which makes the hydraulic system usable in
many other applications as well.
ACKNOWLEDGMENTS
We wish to thank the staff at the department of
Management and Engineering as well as staff at Volvo
Construction Equipment for helping out with much of the
practical work presented in this article.
REFERENCES E-mail: kim.heybroek@liu.se, Tel: +46 (0) 13 28 11 63

1) K. Heybroek, J-O. Palmberg, and J. Larsson. Open LIST OF NOTATIONS

circuit solution for pump controlled actuators.
Proceedings of the 4th FPNI-PhD Symposium, Index Explanation Unit
2006. Sarasota,Florida, USA., pp 27-40.
LS- p Differential pressure determined by Pa
2) S. Habibi and G. Singh. “Derivation of design the load sensing controller
requirements of optimization of a high performance
hydrostatic actuation system”, International Journal Dp Pump displacement m3
of Fluid Power, 1(2), 2000, pp 11–23. Erecup Total energy output on the pump J
shaft when pump operates as a
3) G. Wendel. “Hydraulic system configurations for motor, measured throughout a
improved efficiency”, Proceedings of the 49th specified operation
National Conference on Fluid Power, March 2002,
pp 567–573. Ein Total energy input to the pump shaft J
when pumping, measured
4) R. Rahmfeld and M. Ivantysynova, “Displacement throughout a specified operation
controlled linear actuator with differential cylinder - a F Load force acting on the cylinder N
way to save primary energy in mobile machines”, 5th rod
International Conference on Fluid Power
Transmission and Control (ICFP’2001), 2001, pp F n* Maximum load force in the non- N
296–301. differential control mode, system
pressure at maximum
5) Krus, P. “On load sensing fluid power systems – with F d* Maximum load force in the N
special reference to dynamic properties and control differential control mode, system
aspects”, Ph.D. thesis, 1988, Linköping, pp 73-80. pressure at maximum
p/m Relative pump/motor displacement -
6) R.Rahmfeld, M. Ivantysynova and J. Weber,
“Displacement controlled wheel loader – a simple recup The relation between recuperated -
and clever solution”, 4th IFK (International Fluid energy and energy input to the
Power Conference), 2004, Dresden, Germany. hydraulic pumps throughout a
specified operation
7) C. Williamson and M. Ivantysynova, “The effect of
pump efficiency on displacement controlled p/m Pump/motor efficiency -
actuators”, Proceedings of the Tenth Scandinavian v Actuator velocity m/s
International Conference on Fluid Power, 2007, *
vn Maximum actuator velocity in the m/s
Tampere, Finland, 2(3), pp 301-326.
non-differential control mode, pump
flow at maximum
8) Heybroek, K. and J-O. Palmberg, “Applied control
strategies for a pump controlled open circuit vd* Maximum actuator velocity in m/s
solution”, 6th IFK (International Fluid Power differential control mode, pump flow
Conference), 2008, Dresden. at maximum
qp Pump flow m3/s
9) Filla, R. “An event driven operator model for
dynamic simulation of construction machinery”, The pp Pump pressure Pa
Ninth Scandinavian International Conference on
Fluid Power, June 2005, Linköping, Sweden. Px Power losses, x=defines the source W
of the power loss
10) K. Heybroek, J-O. Palmberg. ”Mode switching and np Pump shaft speed rad/s
energy recuperation in open circuit pump control”,
pp/m Pressure difference over Pa
Proceedings of The 10th Scandinavian International
pump/motor
Conference on Fluid Power, SICFP'07, pp 297-209.
p(L-T) Load pressure minus tank pressure Pa
CONTACT p(P-L) Pump pressure minus load pressure Pa

Kim Heybroek, M.Sc., Ph.D.-student at the division of xp Piston displacement m

Fluid and Mechanical Engineering Systems at the
Department of Management and Engineering (IEI),
Linköping University, Sweden.