CPP propeller - Towards high performance pitch control - Wärtsilä

WÄRTSILÄ TECHNICAL JOURNAL 01.2008
Towards high performance pitch control
AUTHOR: Milinko Godjevac, Research Engineer, Wärtsilä in the Netherlands
This article discusses a novel control strategy for a controllable pitch propeller (CPP) driven by a diesel engine.
This new control strategy should contribute to more effective use of the complete power train.
Fig. 1 – Dredger and its control system.
Load
control
Setpoint
load
prop. shaft torque
Pitch
control
Setpoint
pitch
Mechanical load
Setpoint
speed
Electrical load
Electronical load sharing
communication when
running parallel
Pitch
feedback
hydraulics
Speed
governor
VIKING 25
Propulsion gearbox
OD
box
6.6 kV
10MW
Propulsion
clutch
Flywheel
Wärtsilä
12V46
Main
generator
Fuel preservation is important for the future of mankind.
Any increase in efficiency offers not only fuel savings,
but also reduced engine emissions and more ecological
solutions. For this reason, control strategies for developing
highly efficient propulsive thrust to a ship are important.
The traditional ship with a CPP, driven by a diesel
engine, is controlled via the “combinator”, which sets
the optimum combination of pitch and shaft rotational
speed for stationary ship conditions at no waves and other
added ship resistance conditions (“ideal” condition). The
combinator consists of running up/down and look-up
tables for the (pre-set) optimal pitch and rotational speed
combination. The “optimum” is not only influenced
by fuel consumption considerations, but also by the
propeller noise and cavitation issues. In contrast to ideal
conditions, at sea or maneuvering conditions necessitate
the adjustment of shaft speed and/or pitch. In references
[1] and [2] it was also shown that engine speed
disturbances, caused by the seaway, could be suppressed
better by pitch control than by controlling the fuel rack.
The time needed to get the rotational shafting (diesel
engine, gearbox and propeller) from one shaft-speed level
to another is measured in the magnitude of a few seconds.
For ships with large power take-offs using generators, the
CPP pitch is used to compensate for fast and high frequent
load changes of the generator, as well as for the seaway. For
dredgers (see Figure 1), the generator is used to drive the
dredge pump, which can jump in load from minimum to
maximum and vice versa, rather quickly. The diesel engine
load is kept constant by continuous and fast pitch changes.
The existing situation
Figure 2 shows the simplified CPP actuating system and
pitch control. The oil volumes in the pitch actuating
cylinder and the oil pipelines are controlled by counter
balancing valves. When no pressure exists at the ahead
and astern ports of the oil-distribution box, these counter
balancing valves (CBV) are closing. When these are closed,
the propeller pitch stays as it is. These counter balance
valves, together with the proportional valves with load
sensing, make the task of the pitch controller rather easy.
But the cost is a rather slow actuating pitch, which can
have different dead times from one pitch change to another.
This has to be overcome in order to have fast and accurate
pitch control.
p
in detail 49

[ MARINE / IN DETAIL ]
[ MARINE / IN DETAIL ]
Fig. 2 – Simplified CPP control system.
p S
, p C1
, p C2
(bar)
Fig. 4 – Block diagram of the new pressure controller.
50 in detail
feedback sensor
header tank
or pump
xy
U pv
Valve
drive
Pitch
Controller
0 set
Pitch setpoint
OD-box
Q C1
, p C1
p S
, Q S
CBV
Q C2
, p C2
i 1
i 2
proportional
valve
p S
, p C1
, p C2
, p H1
(bar)
p R
, Q R
pitch ahead
p H
hub
load
sensing
setp.
valve
pitch
Fig. 3 – Modeled and measured results of CPP actuating model and pitch control.
load-sensing control
pressure control
pitch position control
P v,max
S
(C|D)&E
α act
+
C+
S 2
–
S 1
+
–(C|D)
N
A|C
4 + –
u
B|D S 5
P ahead
P astern
ΔP is
α ref
C
+
–
C ΔP_yoke
C pitch
D C+δ D+δ
C load sensing
i prv
p S
p C1
p C2
p H1
E
(C|D)&E
S 6
+
C main valve
X m, ref –
i valve
–(C+δ/D+δ)
X m, act
main spool
position control
The current situation for in-stationary
maneuvering conditions and for increased
vessel resistance in seaway, calls for a “load
control” to be used. The traditional load
control performs rather coarsely and is not
able to precisely control the diesel engine
load as a function of the actual shaft speed.
This load control overrides the pitch set by
the combinator, and reduces it in order to
keep the diesel engine within its operating
envelope. To avoid unnecessary pitch
changes in adverse weather, the passive
load control is provided with an automatic
adjusting dead zone. When pitch
fluctuations, caused by the passive load
control action, are sensed and when these
are periodic, the dead zone will be
increased step by step until the periodic
fluctuations have been removed. In heavy
seaway, this results in an increased service
margin and reduced average diesel engine
loads of up to 60 or 70% of the CSR point.
This passive load control automatically
takes care of “good seamanship” needs,
but it results in a lack of usage of the
propulsion plant.
As opposed to passive load control,
active load control implies an active pitch
change in order to keep the load on the
engine under control. This was investigated
in [3] where the measurements were
carried out on board a 9000 cargo tonner.
These measurements indicate active load
control as being ineffective and causing
additional wear to the pitch adjusting
system. However, looking at the spectrum
analysis presented in the report [3], it can
be clearly seen that the pitch adjusting
system was three to four times slower
than necessary to suppress torque
oscillations. Thus these measurements and
conclusions needed further investigation.
In order to investigate if the load
control could be used in the seaway, the
CPP actuation and control system has
been mathematically modelled [4].
The model contains the following main
elements:
Pitch actuating mechanism.
External pitch actuating force model
(hydro-dynamic forces and moments).
Hub mounted pitch-actuating cylinder.
Oil supply lines and pitch
feedback pipes.
Counter balance valve.
Oil-distribution box.
Hydraulic proportional valves,
load sensing and pumps.
Feedback sensors and pitch
closed loop controller.

WÄRTSILÄ TECHNICAL JOURNAL 01.2008
The study indicated that the response
of the CPP system is slow, mainly because
of the fact that oil has to be compressed
prior to each pitch change. The oil must
be compressed to the point that the hub
pressure reaches the threshold level of the
Coulomb friction. The dead time before
pitch change can easily vary from 0.5
seconds to 3.0 seconds. Figure 3
shows modeled and measured results
for the pitch change. The results show
good correlation between the model
and the actual response as measured.
brake cylinder
frame
brake
leakage valve
mass actuator
Schematical drawing of the test bench.
load
Pressure control
In the controllable pitch propeller, a
counter balance valve is installed in the
shaft line in order to maintain the pitch
should the hydraulics fail. In the new
controller, the counter balance valve is
kept open whilst there is no change in
pitch. This required the control strategy to
be changed. Whilst there was no change
in pitch position, the hydraulic pressure
had to ensure that the pitch could not
change, but as soon as a change of pitch
becomes necessary, the position control
actuates the yoke so that the desired pitch
is achieved. Then it is switched back to
differential pressure control to maintain
the pitch. Figure 4 shows the new control
strategy [5], which will be described below.
When the pitch setpoint α ref
changes,
the pitch error increases so that condition
C or D becomes true as soon as the pitch
error exceeds a certain value. Condition C
indicates that the pitch error is negative,
meaning that the pitch α act
has to be
decreased, while condition D indicates
the opposite. Assuming that the pitch
has to be increased (condition D is true),
the pressure controller will increase the
differential pressure in the hub-cylinder to
the upper limit of the friction band. This
is ensured by condition D switching S4 to
set the upper part of the curve. In this way,
the Coulomb friction band in the system is
quickly crossed as pressure buildup is more
efficient when carried out by a pressure
controller than by a position controller.
When the required differential pressure
is set by the pressure controller CΔ P_yoke
with certain accuracy, condition E
becomes true, so that S6 switches from
pressure control to pitch position control.
At the same time, S2 activates the
electronic load-sensing to ensure a loadindependent
yoke speed when the pitch
is changed, which is done by the pitch
Fig. 5 – Test bench set-up.
position controller C pitch
. The reference
pitch is set δ more accurately than an
error in the pitch is detected to prevent
chattering and unwanted switching
between the pressure and position control.
When the required pitch is achieved,
S6 switches back to pressure control and
load sensing is turned off by S2. Now the
pressure controller sets the differential
pressure needed to hold the pitch, given by
the mean between the upper and the lower
part of the exemplified curve of Figure 2.
When the pitch set point changes, or
when the pressure difference set to hold
the pitch is not adequate due to changes in
the seaway or weather leading to a change
in pitch, the same procedure starts again to
match the actual pitch with its set point.
Note that the validity of Figure 2 can
be extended by taking shaft speed N and
ship speed through water u into account,
as pitch, shaft speed, and ship speed
define the thrust and the hydrodynamic
forces acting on the blades, thus the pitch
actuating force. This extension would
make the pressure controller more reliable.
In house tests
In order to verify the model shown in
Figure 4, a special test set-up has been
constructed in Wärtsilä in the Netherlands.
Figure 5 shows the test bench set-up.
Results obtained from the workshop test
bench indicate a better performance with
the new control strategy and a reduction in
dead time, Figure 6 shows the results from
the new and old controllers. With the
new controller, it appears to be possible to
follow seaway, because a sine as a pitch set
point having a frequency of 0.1 Hz (wave
encounter frequency) and an amplitude
of 10% of the stroke, can be tracked.
During the test bench measurements,
it became clear that the non-linear
differential pressure controller is
substantially faster than a linear differential
pressure controller. Furthermore, with
the new control strategy switching
between pressure and position control,
the pitch can be set in a system with
high Coulomb friction and a load
force while no instability occurs. p
in detail 51

[ MARINE / IN DETAIL ]
[ MARINE / IN DETAIL ]
position [%]
position [%]
Fig. 6 – Seaway simulation for a pressure (up) and position (bottom) controller.
power (%)
10
0
–10
10
0
–10
5 10 15 20 25 30 35 40
5 10 15 20 25 30 35 40
100
80
60
40
20
load limit
load control curve
f 1
f 4
f 5
f 2
f 3
0 20 40 60 80 100
rpm (%)
a1
a2
a3
a4
a5
old
new
V
0.93V
0.86V
0.71V
lines of specific
fuel consumption
(kg kW 1 h –1 )
respect to the reliability of the mechanical
system. The dynamics of the oil pipes
and the response of the hydraulic system
are shifting towards a resonant frequency.
Friction in the system introduces
non-linear effects and makes tuning of
the controller very difficult. Last but not
least, tribology and reliability aspects
of the bearings have to be considered.
More active pitch control means more
movement and would therefore result
in higher wear. In [6] the big amplitude
sliding has been reported as not being a
problem for conventional use of a CPP
under designed conditions. However,
this high control means more small
amplitude and (relatively) high frequency
sliding, which should be done in such
a manner that fretting will not occur.
On the other hand, the operation of
a CPP in seaways is easily carried out
in an improper manner causing control
movement due to a mismatch between
the seaway and the actuator. From several
references (vessels which are currently
in operation) where a more active pitch
control has been used, the indication is
that this has never led to increased wear
problems. Also, the misuse of the CPP will
lead to increased dynamic loads. Without
improvements to the hydraulic system,
the use of the CPP in seaways would be
too slow to follow the wave and would
lead to increased dynamic loads. For new
ship systems where such a solution needs
consideration, the consequence of the
increased use should be studied as well.
When done properly, it is expected that the
new solution will contribute to fuel savings
by a factor of several percentage points.
Fig. 7 – Operating envelope of a propulsion plant.
CONCLUSION
There are two main applications that can
benefit from the present work, but both
have the same starting point. The work
presented above is an attempt to make the
pitch control faster and more dynamic,
thereby giving a reduced service margin.
The first application is for dredgers, where
the reduced service margin results in
more power for the dredging equipment,
and quicker suppression of disturbances.
The second application is for cargo ships
52 in detail
sailing in rough seas. Figure 7 shows the
operating envelope of a propulsion plant,
with the old control system propulsion
plant operating in the red point. With the
new control system, the service margin
would be smaller and the operating point
will go up. These effects result in higher
speed and lower fuel consumption, a
rough estimate of which is between 2 and
5% of the specific fuel consumption.
However, besides the benefits, there
are several issues to be considered with
REFERENCES:
1. H T Grimmelius, D Stapersma, ‘Control
optimisation and load prediction for
marine diesel engines using a mean
value simulation model’, ENSUS 2000,
Newcastle-upon-Tyne (September 2000).
2. H T Grimmelius, D Stapersma, ‘The impact
of propulsion plant control on diesel engine
thermal loading’, CIMAC Conference,
Hamburg (May 2001).
3. S. van der Steenhoven, ‘Monitoring 9000
tonner’, Wärtsilä internal report,
(September 2003).
4. J. Bakker, A. Wesselink, ‘The use of nonlinear
models in the analysis of CPP
actuator behavior’, WMTC Conference,
London (March 2006).
5. J.H.H. Huijbers, ‘Non-linear propeller
pitch control’, MSc thesis, TU Eindhoven,
(December 2007).
6. M. Godjevac, T. van Beek, H. Grimmelius,
D. Stapersma, ‘Wear mechanism of a CPP’,
WMTC Conference, London (March 2006).