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

net/publication/290326213

A mathematical model of the propeller pitch change mechanism for the

marine propulsion control design

Article · January 2012

DOI: 10.1201/b11810-98

CITATIONS READS

11 5,894

3 authors:

Marco Altosole Michele Martelli

Università degli Studi di Genova Università degli Studi di Genova
67 PUBLICATIONS 947 CITATIONS 74 PUBLICATIONS 733 CITATIONS

SEE PROFILE SEE PROFILE

Stefano Vignolo
Università degli Studi di Genova
11 PUBLICATIONS 191 CITATIONS

SEE PROFILE

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

Smart control of ship propulsion systems View project

Autonomous Ship View project

All content following this page was uploaded by Michele Martelli on 26 February 2019.

The user has requested enhancement of the downloaded file.

 Proceedings of the Institution of Mechanical
Engineers, Part M: Journal of Engineering for
the Maritime Environment
http://pim.sagepub.com/

Controllable pitch propeller actuating mechanism, modelling and simulation

Michele Martelli, Massimo Figari, Marco Altosole and Stefano Vignolo
Proceedings of the Institution of Mechanical Engineers, Part M: Journal of Engineering for the Maritime Environment
published online 11 January 2013
DOI: 10.1177/1475090212468254

The online version of this article can be found at:

http://pim.sagepub.com/content/early/2013/01/10/1475090212468254

Published by:

http://www.sagepublications.com

On behalf of:

Institution of Mechanical Engineers

Additional services and information for Proceedings of the Institution of Mechanical Engineers, Part M: Journal of Engineering for the
Maritime Environment can be found at:

Email Alerts: http://pim.sagepub.com/cgi/alerts

Subscriptions: http://pim.sagepub.com/subscriptions

Reprints: http://www.sagepub.com/journalsReprints.nav

Permissions: http://www.sagepub.com/journalsPermissions.nav

>> OnlineFirst Version of Record - Jan 11, 2013

What is This?

Downloaded from pim.sagepub.com at Universita di Genova on February 7, 2013

Original Article

Proc IMechE Part M:

J Engineering for the Maritime Environment
0(0) 1–15
Controllable pitch propeller actuating Ó IMechE 2013
Reprints and permissions:
mechanism, modelling and simulation sagepub.co.uk/journalsPermissions.nav
DOI: 10.1177/1475090212468254
pim.sagepub.com

Michele Martelli1, Massimo Figari1, Marco Altosole1

and Stefano Vignolo2

Abstract
This article focuses on the mathematical model of the pitch control mechanism for a marine controllable pitch propeller,
with the aim of describing the dynamic behaviour of this kind of system and its influence on ship performance. Too great
a load on the blades can result in high pressures in the actuating system, response delays and control system problems,
which are ultimately responsible for most mechanism failures. The behaviour of the controllable pitch propeller actuating
mechanism is considered in terms of blade position, oil pressures inside the controllable pitch propeller hub and magni-
tudes of the forces acting on the blades. In the proposed mathematical model, the forces acting on the propeller blade
are evaluated taking into account the yaw motion of the ship, the propeller speed (including shaft accelerations and
decelerations) and the turning of the blade during the pitch change. On the basis of the introduced procedure, a control-
lable pitch propeller numerical model as part of an overall propulsion and manoeuvrability simulator representing the
dynamic behaviour of a twin-screw fast vessel is developed. The aim of this work is to represent the ship propulsion
dynamics through time-domain simulation, based on which the designers can develop and test several design options, in
order to avoid possible machinery overloads with their consequent failures and to obtain the best possible ship perfor-
mances. In this aspect, the controllable pitch propeller model is an essential design tool.

Keywords
Ship propulsion, controllable pitch propeller, pitch mechanism, simulation

Date received: 14 February 2012; accepted: 18 October 2012

Introduction propulsion control options.1,2 Hence, good design and

optimisation of the whole control system based on
For several years, the University of Genoa has been simulation require a detailed representation of the
involved in the development of propulsion control sys- dynamics of the pitch control mechanism of the CPP
tems for naval vessels, mostly propelled by gas turbines by means of a reliable time-domain numerical model.3,4
driving controllable pitch propellers (CPPs). The Although CPPs have been in use for many years, a
increasing complexity of these propulsion systems well-known mathematical procedure able to take into
requires control system functions able to manage high account variable phenomena does not yet exist. The
power in several propulsion configurations and during principal difficulties are due to an uncertain under-
critical ship manoeuvres (slam start, crash stop, severe standing of the loads acting on the propeller blade and
turning circles, etc.). on the moving parts of this mechanism. A contribution
Optimising the whole propulsion control system of on this topic is proposed hereinafter where a theoretical
the vessel, especially in critical conditions, means finding treatment of the pitch control mechanism and the cor-
an optimal compromise between performance and responding simulation model are described. The model
safety. In particular, each propulsive component, includ-
ing the pitch actuating mechanism, has to be protected 1
Department of Electric, Electronic, Telecommunication Engineering and
from possible dangerous overloads. At the same time, Naval Architecture, Università Genova, Genoa, Italy
important elements such as prompt response of the pro- 2
Department of mechanical, energy, management, transportation and
pulsion system, fuel saving and reduced cavitation phe- mathematical engineering, Università Genova, Genoa, Italy
nomena need to be pursued by the ship designer.
Corresponding author:
Simulation techniques can be a very useful tool in Massimo Figari, Department of Marine, Electrical and Electronic
representing marine propulsion dynamics, based on Engineering, Università Genova, Via Montallegro 1, Genoa 16145, Italy.
which automation designers can develop and test Email: [email protected]

Downloaded from pim.sagepub.com at Universita di Genova on February 7, 2013

2 Proc IMechE Part M: J Engineering for the Maritime Environment 0(0)

considers as state variables the propeller blade position 14

and the oil pressures inside the CPP hub. In fact, the Mechanical
blade position is the key factor for the ship perfor- Hydraulic
12
Electrical
mance, while the oil pressure is the key parameter of
the mechanism; too high pressures can be responsible 10

Total failures
for mechanism failure, and hence, they should always
be monitored by the ship automation systems. 8
In order to develop a simulation model suitable for
6
CPP control system design and testing, current knowl-
edge of the loads acting on a blade has to be assessed.
4
The total spindle torque is the torque acting around the
spindle axis of the propeller blade, and it needs either to 2
be balanced by the hub mechanism in order to hold the
blades in the required pitch setting or, alternatively, to 0
be overcome when a pitch change is required. The spin- 0 1 2 3 4 5 6 7 8 9 10
dle torque can be subdivided into three components as Years in service
follows5
Figure 1. Failure of CPPs versus years of service.
QS (J, u) = QSH (J, u) + QSI (n, u) + QSF (J, u) ð1Þ

where QS (Nm) is the total spindle torque at a given The intensive use of the CPP can occur if new con-
value of non-dimensional propeller advance coefficient trol strategies are used.12 An example of such applica-
J and at a given value of pitch angle u; QSH (Nm) is the tions includes the use of the CPP to protect the engine
hydrodynamic component of the spindle torque due to from load fluctuation due to waves or when the pitch is
the water pressure field acting on the blade surfaces; moved to reduce the propeller noise.
QSI (Nm) is the inertial component (QSC in Carlton Taking into account the above-mentioned literature,
nomenclature) resulting from the blade mass distribu- the authors proposed a model13 to determine the
tion and QSF (Nm) is the frictional component of the dynamic effects (mainly due to inertia and hydraulic oil
spindle torque resulting from the relative motion of the pressure) during the pitch change. The model was then
surfaces within the blade hub. revised with the collaboration of a major CPP manu-
The spindle torques are usually expressed in the non- facturer. The aim of this article is to describe in detail a
dimensional form of KQSH and KQSI. These coeffi- realistic model of the pitch control mechanism. The
cients are similar to the conventional propeller torque developed CPP model is incorporated into an overall
coefficient, and they are related to the respective spindle ship propulsion and manoeuvre simulation model able
torque components as follows to evaluate the whole ship performance, as well as
detailed parameters such as the oil pressure inside the
QSH two chambers of the hydraulic mechanism. Through
KQSH = ð2Þ
rn2 D5 this simulation approach, it is possible to describe the
QSI dynamic performance of the CPP mechanism during
KQSI = ð3Þ ship manoeuvres. The use of this kind of simulation
r b n2 D 5
model is planned for the design and the setting of the
where r and rb (kg/m3) are the water and the blade propulsion control system.
mass densities, respectively; n (rps) is the propeller
speed and D (m) is the propeller diameter.
Little data about the spindle torque have been
Pitch mechanism overview
published. Some of the most exhaustive work is the CPPs are generally actuated by hydraulic oil power sys-
experimental data of the 3-Bladed JP-CPP Series.6,7 tems due to their high power–volume ratio.14 Usually,
Unfortunately, these data are not sufficient to accu- the simplest hydraulic–mechanical actuator is used: the
rately represent the behaviour of the CPP actuating piston in a cylinder.
mechanism during ship manoeuvres. The main elements of a CPP hydraulic system are
Some recent publications by Bakker et al.,8 the following: a tank, pumps, filters, valves, pipelines,
Wesselink et al.9 and Godjevac et al.10 deal with the oil distribution (OD) box, double-effect cylinder, cooler
problem of the evaluation of the spindle torque acting and sensors. The oil flows from the tank to the OD box
on a blade. The aim of these works is to identify a pos- in a normal pipe inside the engine room. The OD box
sible procedure for the correct evaluation of the fretting is a directional valve located on the shaft from which
motion in order to prevent possible failures due to an the oil flows through a twin pipe inside the shafting to
over intensive use of the pitch actuating mechanism. In the propeller hub piston.
fact, the failure of a CPP mechanism is not a remote Inside the CPP hub, a double-effect hydraulic cylin-
possibility as shown in Figure 1.11 der is actuated by the oil pressure in a longitudinal

Downloaded from pim.sagepub.com at Universita di Genova on February 7, 2013

Martelli et al. 3

where Qhyd (Nm) is the hydraulic torque, Fhyd (N) is the

hydraulic force, eyp (m) is the eccentricity of the yoke
pin and u (rad) is the pitch angle.
The relationship between the linear movement (pis-
ton stroke) and the pitch angle (blade angular position)
is expressed as

x = xMAX AH  tgðuMAX AH  uÞ eyp ð5Þ

where x (m) is the instantaneous linear position

(stroke), xMAX_AH (m) is the maximum stroke ahead
and uMAX_AH (rad) is the maximum pitch ahead. All
the variables appearing on the right-hand side of equa-
tion (5) are derived from the corresponding mechanism
drawing represented in Figure 4.

Figure 2. Main elements inside a CPP hub.

Reference frames
A convenient frame of reference to describe the propel-
direction (stroke) (see Figure 2). Two actuating cham- ler blade motion is the hub fixed frame. The latter,
bers are needed in order to move the blade in both pos- together with all other reference frames used in this
itive and negative pitch angles. article, are defined in the following:
The piston is connected to the blades (generally four
or five) by a pin. Different types of connecting mechan-
Inertial n-frame ðOn , n1 , n2 , n3 Þ. It is a local geogra-
isms exist to convert the piston stroke into a blade
phical frame fixed to the Earth. The positive unit vector
angular rotation. In the present work, the pin–slot n1 points towards the north, n2 points towards the east
mechanism is considered (see Figure 3). On the left of and n3 points towards the centre of the Earth
the figure, the rotation of the blade carrier driven by (Figure 5). The origin On is located on the mean water
the piston inside the CPP hub is shown, while on the free surface at an appropriate location.
right the pin–slot kinematic is shown.15
For the pin–slot mechanism, the relationship Ship fixed b-frame ðOb , b1 , b2 , b3 Þ. This frame is fixed
between the hydraulic force generated inside the hub to the hull. The positive unit vector b1 points towards
(piston force) and the spindle torque is the following the bow, b2 points towards starboard and b3 points
downwards. Often, for marine vehicles, the axes of this
Fhyd  eyp frame are chosen to coincide with the principal axes of
Qhyd = ð4Þ
cos2 (u) inertia; this determines the position of the origin Ob of

Figure 3. Pin–slot mechanism.

Downloaded from pim.sagepub.com at Universita di Genova on February 7, 2013

4 Proc IMechE Part M: J Engineering for the Maritime Environment 0(0)

Figure 7. Hub fixed frame.

Figure 4. Relationship between linear and angular positions.

Figure 8. Blade fixed frame.

blade and e2 = e3 ^ e1 . The origin Oe of the frame is

located in the centre of the shaft line. The transforma-
tion between the bases fbi g and fei g are given by
8
> e 1 = b1
>
<
e2 = cos(u)b2 + sin(u)b3
>
>
Figure 5. Inertial frame. : e3 = sin(u)b2 + cos(u)b3
8
< b1 = e 1
>
b2 = cos(u)e2  sin(u)e3
>
:
b3 = sin(u)e2 + cos(u)e3

where u (rad) denotes the blade angular position as

shown in Figure 7.
Blade fixed f-frame ðO, f1 , f2 , f3 Þ. This frame is fixed to
any single blade. The unit vector f3 coincides with the
spindle axis of the blade, while the unit vectors f1 and
f2 describe the rotation of the blade around its spindle
Figure 6. Ship fixed frame. axis as indicated in Figure 8. The origin O coincides
with Oe. The transformation between the bases fei g
the frame. The transformation between the bases fni g and ffi g are expressed as
and fbi g are described by 8
> f = cos(u)e1 + sin(u)e2
>
< 1
8 f2 = sin(u)e1 + cos(u)e2
> n1 = cos(c)b1  sin(c)b2
>
< >
n2 = sin(c)b1 + cos(c)b2 >
: f 3 = e3
>
> 8
: n3 = b3
< e1 = cos(u)f1  sin(u)f2
>
8
e2 = sin(u)f1 + cos(u)f2
< b1 = cos(c)n1 + sin(c)n2
> >
:e =f
b2 = sin(c)n1 + cos(c)n2 3 3
>
:
b3 = n3 where u (rad) denotes the pitch angle as shown in
Figure 8.
where c (rad) denotes the yaw angle as in Figure 6.
Using the angular velocity composition theorem, it is
Hub fixed e-frame ðOe , e1 , e2 , e3 Þ. This frame is fixed possible to represent the angular velocities (with respect
to the hub. The positive unit vector e1 points towards to the inertial n-frame) of a single blade vB and the hub
the bow, e3 coincides with the spindle axis of a given vH , respectively, as

Downloaded from pim.sagepub.com at Universita di Genova on February 7, 2013

Martelli et al. 5

_ 3 + ue
vB = vc + vu + vu = cb _ 1 + uf
_ 3 ð6Þ

_ 3 + ue
vH = vc + vu = cb _ 1 ð7Þ

where c,_ u_ and u_ (rad/s) denote the time derivatives of

the yaw angle, the blade angular position and the pitch
angle, respectively.

Motion equations
The proposed mathematical model is based on two
equations: the motion of the blade and the motion of
the piston. The first one is the equation of motion of a
blade around its f3 axis, evaluated in the e-frame
1
€=
u (Qhyd + QS + QF ) ð8Þ
I33
where u € (rad/s2) is the blade angular acceleration, QF
(Nm) is the torque due to the interaction forces between
propeller blade and blade bearing, Qhyd (Nm) is the
hydraulic torque, QS (Nm) is the total spindle torque
acting on the blade and I33 (kg m2) is the moment of Figure 9. Hub and blade reference frames.
inertia of the blade about the spindle axis f3 .
The second differential equation describes the
motion of the cylinder and the weight force is given by
X
Z ð
m x€ = A1 p1  A2 p2  Bp x_ + Fi ð9Þ W
F = rb gdt = mb g ð12Þ
i=1
b
where m = my + mp + mo (kg) is the sum of the yoke
mass (my), propeller mass (mp) and oil mass (mo), x (m) where a0 (m/s2) is the linear acceleration of the origin O
is the cylinder position, A1 and A2 (m2) are the yoke with respect to the inertial frame, rb (kg/m3) is the mass
areas of the astern chamber and the ahead chamber, density of the propeller blade, (P  O) (m) is the posi-
respectively, p1 and p2 (Pa) are the pressures inside the tion vector of a generic point P of the blade with respect
two to the origin O, mb (kg) is the blade mass, g (m/s2) is the
PZ chambers, Bp (kg/s) is the damping coefficient and gravity acceleration, b is the whole set of the points
i = 1 Fi (N) is resultant of the reaction forces due to
the interaction between each blade and the piston (Z constituting the blade and vrP (m/s) is the linear velocity
denoting the blades number). vector of a generic point P of the blade evaluated in the
In order to properly implement the differential equa- e-frame as shown in Figure 9.
tions (8) and (9), all the forces and moments acting on a More explicitly
single blade have to be evaluated. This will be shown in vrP = vr0 + vu ^ (P  O) = uf
_ 3 ^ (P  O) ð13Þ
the next sections.
Recalling the definition of centre of gravity for a
Inertial and gravity forces blade
ð
The yaw motion of the ship and the rotation of the pro- 1
(G  O) = r (P  O)dt ð14Þ
peller give rise to corresponding Coriolis and transpor- mb b
b
tation inertial forces acting on each blade in the e-frame.
Moreover, gravity yields a sinusoidal varying force. it is possible to express the inertial forces (10) and (11)
More in detail, the Coriolis force is defined by in a simpler form
ð
FC = 2rb vH ^ vrP dt ð10Þ FC = 2mb vH ^ vrG ð15Þ
b FS = mb ao  mb vH ^ ½vH ^ (G  O)
the transportation force is expressed as mb v_ H ^ (G  O) ð16Þ
ð
The total contribution of inertial and weight forces act-
FS = rb ½aO + vH ^ ðvH ^ (P  O)Þ + v_ H ^ (P  O)dt
ing on a blade is then given by the sum
b

ð11Þ FI = FW + FS + FC ð17Þ

Downloaded from pim.sagepub.com at Universita di Genova on February 7, 2013

6 Proc IMechE Part M: J Engineering for the Maritime Environment 0(0)

In addition to this, the moments with respect to the ori- forces in b2 direction gives rise to a transverse force Q/r;
gin O of the above forces have to be evaluated. To this the forces along b3 have been neglected. In particular,
end, the moment of a generic (density of) force acting propeller thrust T (N) and propeller torque Q (Nm) are
on a blade is given by given by
ð
MO = (P  O) ^ Fdt ð18Þ T = rKt n2 D4 ð24Þ
b Q = rKq n2 D5 ð25Þ

Combining equation (18) and the definition of the iner- where Kt is the propeller thrust coefficient, Kq is the
tia tensor with respect to the origin O for each blade propeller torque coefficient and r = 0.35D is the
ð assumed propeller radius.
IO (v) = rb (P  O) ^ ½v ^ (P  O)dt ð19Þ The non-dimensional coefficients Kt and Kq are a
function of the non-dimensional advance coefficient J
b
and pitch angle u, derived from propeller open water
it is possible to demonstrate (see Appendix 2) that the tests. The resultant hydrodynamic force is assumed act-
moments of the inertial forces can be expressed as ing on the hydrodynamic centre of the blade CH; the
components of the vector ðCH  OÞ in the b-basis are
MC
O = vu ^ IO (vH )  vH ^ IO (vu ) + IO (vu ^ vH ) denoted by (xCH, yCH, zCH). Then
ð20Þ
FHD, b1 = T ð26Þ
MSO = mb (G  O) ^ aO  vH ^ IO (vH )  IO (v_ H ) 
Q
ð21Þ FHD, b2 = ð27Þ
r
where vu = uf _ 3 denotes the angular velocity of the where T* (N) is the single blade thrust and Q* (Nm) is
blade evaluated in the e-frame. The moment due to the the required torque by a single blade
weight force is
T
MW T = ð28Þ
O = mb (G  O) ^ g ð22Þ Z
Q
The total moment with respect to the origin O due to Q = ð29Þ
inertial and weight forces acting on a blade is then given Z
by the sum T and Q are evaluated using expressions (24) and (25).
The resultant moment QSH of hydrodynamic forces
MIO = MW S C
O + MO + MO ð23Þ with respect to the origin O is evaluated in the b-basis;
All moments are evaluated in Newton metre. The com- its components QSH, b1 , QSH, b2 and QSH, b3 (Nm) are
ponent of the vector MIO along the unit vector f3 pro- given by
vides the scalar quantity QSI. Q
QSH, b1 = FHD, b2  zCH =   zCH ð30Þ
r
QSH, b2 = FHD, b1  zCH = T  zCH ð31Þ
Hydrodynamic forces
QSH, b3 = FHD, b1  yCH + FHD, b2  xCH
In the literature, different numerical methods based, for
instance, on the potential approach16 or on Reynolds Q
= T  yCH +  xCH ð32Þ
averaged Navier–Stokes equations solver (RANSE)17 r
have been proposed in order to predict propeller hydro- The above resultant moment QSH needs to be decom-
dynamic loads. Unfortunately, due to their high com- posed into the f-basis in order to be used in the motion
putational time, these methods are difficult to apply in equation (8).
the present simulation approach. Therefore, in the pro-
posed work, the hydrodynamic forces have been evalu-
ated through a quasi-steady methodology based on
Frictional forces
regressions of experimental data.6,18
The latter evaluates the variation of the forces acting The blade seat on the hub is made by a bearing that
on a blade by using the propeller open water character- supports forces in both axial and radial directions
istics and the position of the centre of pressure in a (Figure 10). These friction forces have been derived for
steady-state regime. This procedure is efficient from the each direction separately.
point of view of the computation time; however, the The frictional radial force and moment have been
precision is difficult to assess. evaluated by the procedure proposed by Godjevac
In evaluating the total hydrodynamic forces, it has et al.,10 and it is not further discussed here. The fric-
been assumed that the sum of all forces in b1 direction tional axial force FAX is the vectorial sum of the two
equals the total propeller thrust T, while the sum of all components in the blade fixed frame f, as shown in

Downloaded from pim.sagepub.com at Universita di Genova on February 7, 2013

Martelli et al. 7

friction seat of the propeller blade root and kf is the

non-dimensional frictional coefficient of the blade and
hub materials. The sum of the axial and radial friction
torques provides the total friction spindle torque QSF.

Hydraulic forces
The oil pressure needed to turn the blade or, alterna-
tively, to hold it in the desired position is supplied by a
hydraulic power pack consisting of pumps and valves.
The actuating system consists of a double-effect piston
(yoke) with single rod and circular section. The devel-
oped force is proportional to the yoke area and the oil
Figure 10. Blade seat. pressure; a difference exists between the developed force
in forward and backward directions, with the following
relationship

p D2p p
FThrust = (N) ð37Þ
4
Fax,f1 p(D2p  dr2 )p
f1 FTraction = (N) ð38Þ
4
Fax,f2
where Dp (m) is the piston diameter, dr (m) is the rod
Fax diameter and p (Pa) is the pressure inside the hub
chamber.
To evaluate the dynamic change of pressure, the fol-
f2 lowing relationships are used

Figure 11. Axial contact.  State equation

ri, oil
roil = ri, oil + p ð39Þ
B
Figure 11 and equation (33). The third component has
been neglected for simplicity  Continuity equation
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi X X  
FAX = F2AX, f + F2AX, f (N) ð33Þ d(roil V) droil dV
m_ i  m_ o = = V + roil
1 2
dt dt dt
The interaction between the blade carrier and the hub ð40Þ
is considered as a pointwise contact. By using the static
equilibrium of forces, the axial components along f1 where V (m3) is the piston chamber volume, roil (kg/m3)
and f2 can be assessed is the oil mass density, B (Pa) is the oil bulk modulus,
X m_ i (kg/s) is the mass flow in and m_ o (kg/s) is the mass
FAX, f = 0 ) FHD, f + FI, f + FAX, f = 0 flow out.
1 1 1 1
By neglecting the dependence of the density on the
) FAX, f = FHD, f  FI, f ð34Þ
X 1 1 1 temperature and combining equations (39) and (40)
FAX, f = 0 ) FHD, f + FI, f + FAX, f = 0
2 2 2 2 V dp dV
qi  qo =  + ð41Þ
) FAX, f = FHD, f  FI, f ð35Þ B dt dt
2 2 2

where FHD, f and FHD, f (N) are the components of where qi (m3/s) is the volumetric flow in and qo (m3/s) is
1 2
the hydrodynamic forces along f1 and f2 , respectively, the volumetric flow out.
obtained by the projection of equations (26) and (27) in The leakage of the hydraulic actuator between the
the blade fixed frame f; and FI, f and FI, f (N) are the two chambers is defined through the following
1 2
components of the inertial forces along f1 and f2 , coefficient
respectively, obtained by equation (17).
Dp e3 1
Then the torque due to the axial force is given by Cip = p    ð42Þ
2 Lp 6m
QFR, AX = kf  FAX  da (Nm) ð36Þ
where e (m) is the orifice thickness, Lp (m) is the thick-
where da (m) is the distance between the point where ness of the piston head and m (Pa s) is the oil dynamic
the axial force is supposed to be applied and the hub viscosity.

Downloaded from pim.sagepub.com at Universita di Genova on February 7, 2013

8 Proc IMechE Part M: J Engineering for the Maritime Environment 0(0)

PUMPS NUMBER
PITCH OIL FLOW
SETPOINT

CONTROLLER P1

DIFFERENTIAL
EQUATIONS P2
Qs SOLVER
FORCES
& ACTUAL
MOMENTS PITCH

Figure 12. Calculation of actual pitch.

Finally, the pressure differential equation becomes manoeuvre is taken into account by means of the intro-
! duction of asymmetrical variations of both wake and
p D2p B thrust deduction factors.
p_ i = qi  Cip pi  x_ p D2 ð43Þ
4 p All the models are managed by the overall controller
4 x+V 0
through its main inputs. The overall controller func-
tions regard both the propulsion and the electric power
management system (PMS), while the simulator is able
Simulation approach to represent machinery dynamics and ship manoeuvr-
ability. Every numerical sub-model is developed in
The calculation process for the propeller pitch
MATLAB-SimulinkÒ, a widely used platform for
dynamics is illustrated in Figure 12. The pitch set point,
dynamic systems simulation. The high-level scheme of
depending on the bridge lever position, is transformed
the whole simulation model is illustrated in Figure 13.
by the CPP controller into a proper oil flow acting on
the piston. By solving the pressure and motion equa-
tions (43), (8) and (9), it is possible to achieve the
Simulation results
desired value of the propeller pitch. The model of the
CPP mechanism, previously described, has been tested Hereinafter some results obtained by including the CPP
by simulation. model into the overall propulsion and manoeuvrability
The CPP model of Figure 12 has been incorporated simulation model of a naval vessel are shown. The mod-
into a ship performance model, where the ship beha- elled ship has a propulsion system with two shaft lines
viour has been simulated by means of a mathematical and two CPPs, as shown in Table 1.
model that includes the interactions between the The simulations have been used to assess the design
dynamics of the propulsion plant, the propulsion con- of the propulsion control system; different kinds of ship
trol system and the ship motions.2,3,4,13 The model is operations, both in transient as well as in steady-state
time dependent, and it consists of a set of differential conditions, have been simulated. Typically, transient
equations, algebraic equations and numerical tables manoeuvres were slam start, crash stop, acceleration
that represent the various elements of the propulsion and deceleration. Steady-state conditions were sailing
system. All the main elements of the propulsion plant at design speed and sailing at off design speeds. For
such as the main engines and their governors, the hull, each simulated operation, simulated results pertaining
the two propellers, the two rudders and the shaft lines to ship manoeuvring characteristics, machinery perfor-
are modelled as separate subsystems linked to each mance and control system behaviour were available.
other. Also, the manoeuvrability of the ship with 3 The following results highlight the capacity of the CPP
degrees of freedom (surge, sway and yaw) is modelled. sub-model to catch the pertinent phenomena in each
Hull forces and moments are evaluated from regression scenario. Figure 14 plots the spindle torque versus ship
series, which were tested in detail in the study of speed. The simulated propulsion mode is ‘navigation’ in
Viviani et al.19 In particular, regression formulae dedi- both ahead and astern runs. The figure shows the ‘total
cated to twin-screw vessels were obtained starting from spindle torque’ (black line) as well as its components
Clarke and Inoue models with minor tuning in order to (dots) as defined in equation (1); values are reported in
consider model test results. Regarding rudder forces, non-dimensional form, with respect to the spindle tor-
the model described in the study of Viviani et al.20 is que at full ahead speed (100%). Values in Figure 14
adopted. Asymmetrical behaviour of shaft lines during represent steady-state results, that is, the spindle torque

Downloaded from pim.sagepub.com at Universita di Genova on February 7, 2013

Martelli et al. 9

Figure 13. Simulation scheme.

PMS: power management system.

Table 1 Simulated ship characteristics. predominance of the frictional component (cross) at the
highest speed. In astern condition, the hydrodynamic
Ship mass 6000 ton component (star) is the highest one, while the other two
Length 140 m
Beam 20 m
cancel each other due to the negative sign of the inertial
Draft 5m component (circle).
Max speed 27 knots Figure 14 also shows that in the ship speed range
Propulsion power 36 MW 220% to +30%, the total spindle torque is positive but
Shaft torque 1000 kNm very small, and thus, the CPP actuating device would
have the capability of a very quick response. However,
the main engine is usually not able to tolerate rapid
load increase at minimum revolutions per minute, so
Spindle torque
110 the analysis of transient manoeuvres allows the designer
90 a deep insight in the overall propulsion plant behaviour
Spindle torque components [%]

70
in order to properly set and eventually optimise the
controller parameters.
50
In the following figures 15 to 24, some results of two
30
simulated manoeuvres of the ship are shown: a turning
10
circle and the crash stop. The turning circle manoeuvre
-10 has been chosen to emphasise the capability of the
-30 simulator to take into account the interaction between
-50 Total hydrodynamic phenomena (asymmetrical behaviour of
-70 Inertial the two shaft lines) and propulsion control. The ship
Hydrodynamic
-90 has been simulated sailing at maximum power, first on
Frictional
-110
a straight course than during a turn to port side.
-60 -40 -20 0 20 40 60 80 100 Figure 15 shows the command lever position versus
Ship Speed [%] time and the rudder angle versus time.
Simulation results are presented in Figures 16 to 19
Figure 14. Trends of spindle torque components as function of
in non-dimensional form, scaled to the maximum value
ship speed.
in ahead condition. Figure 16 shows the shaft speed ver-
sus time. When the rudder was turned port side, the
shaft torque increased over the maximum allowed and
obtained when the ship speed is stabilised on a fixed the shaft speed decreased. Figure 17 shows the pitch
value. angle behaviour. The propulsion controller commanded
From Figure 14, it is possible to assess the impor- a voluntary pitch reduction in order to prevent over tor-
tance of the different components of the spindle torque. que of the shaft line.4,13 Figure 18 shows the total spin-
In ahead condition, the total spindle torque is approxi- dle torque on the two propellers, port (internal shaft
mately divided into three equal parts, with a slightly line) and starboard (external shaft line), versus time.

Downloaded from pim.sagepub.com at Universita di Genova on February 7, 2013

10 Proc IMechE Part M: J Engineering for the Maritime Environment 0(0)

TELEGRAPH Spindle torque

Lever Position [%]

110 140
90
70
50
30 120
10
-10

Spindle torque [%]

-30
-50 100
110 140 170 200 230 260 290 320 350
Time [s]
80
RUDDER Starboard
Rudder angle [%]

100
75 Port
50 60
25
0
-25
-50 40
-75
-100
110 140 170 200 230 260 290 320 350 20
Time [s] 110 140 170 200 230 260 290 320 350
Time [s]
Figure 15. Turning circle manoeuvre.
Figure 18. Spindle torque versus time in turning circle.

Shatfline Speed Oil pressure

110 210
190
100 170
150
Oil pressure [%]

90
130
Speed [%]

110
80
Setpoint 90
Feedback 70
70
50

60 30
Ahead Starboard Chamber
10
Ahead Port Chamber
50 -10
110 140 170 200 230 260 290 320 350 110 140 170 200 230 260 290 320 350
Time [s] Time [s]

Figure 16. Revolutions per minute versus time in turning Figure 19. Oil pressure versus time in turning circle.
circle.

The large difference between the two torques is a conse-

quence of the asymmetrical flow on the two propel-
Pitch Angle
110 lers.21 In Figure 19, the oil pressure inside the piston
chamber, port and starboard mechanisms is shown.
105
The most valuable result, from the CPP mechanism
point of view, is the 20% increase of the spindle torque
100
during turning with respect to full ahead, straight
course value (Figure 18, t = 200 s). This phenomenon
Pitch [%]

95
would produce a similar increase in the piston actuat-
ing pressure, as shown in Figure 19, approximately at
90
the same time instant. The last proposed manoeuvre is
Starboard Setpoint
the crash stop in order to show the capability of the
Starboard Feedback CPP model and its control system to simulate the acti-
85
Port Setpoint vation of the two cylinder chambers as a function of
Port Feedback
the acting loads.
80
110 140 170 200 230 260 290 320 350 Figure 20 shows the command lever position and the
Time [s] rudder angle versus time of the crash stop manoeuvre.
The simulation started with the ship in full ahead condi-
Figure 17. Pitch angle versus time in turning circle. tion, and after 150 s, the full astern command was

Downloaded from pim.sagepub.com at Universita di Genova on February 7, 2013

Martelli et al. 11

TELEGRAPH Pitch Angle

Lever Position [%]
110
90 100
70
50
30 80
10 Setpoint
-10 Feedback
-30 60
-50
110 140 170 200 230 260 290 320 350

Pitch [%]
40
Time [s]
20
RUDDER
Rudder angle [%]

100 0
75
50
25 -20
0
-25 -40
-50
-75
-100 -60
110 140 170 200 230 260 290 320 350 110 140 170 200 230 260 290 320 350
Time [s]
Time [s]

Figure 22. Pitch angle versus time in crash stop.

Figure 20. Crash stop manoeuvre.

Shatfline Speed Spindle torque

110 140
120
100
100
80 Starboard
60 Port
Spindle torque [%]

90 40
20
Speed [%]

0
80
-20
Setpoint
-40
Feedback
70 -60
-80
-100
60 -120
-140
50 -160
110 140 170 200 230 260 290 320 350 110 140 170 200 230 260 290 320 350
Time [s] Time [s]

Figure 21. Revolutions per minute torque versus time in crash Figure 23. Spindle torque versus time in crash stop.
stop.

simulated. The propulsion controller, at each step,

reacted by commanding a required shaft speed and a Oil pressure
210
required pitch. Figures 21 to 24 show the simulation Ahead Chamber
190
results. In particular, Figure 21 shows the shaft speed Astern Chamber
170
during the manoeuvre where two values are reported:
the speed required by the controller and the instanta- 150
Oil pressure [%]

neous shaft speed. At certain points, due to the high 130

level of propeller torque required (mainly when the pro- 110
peller begins to provide astern thrust), the shaft speed 90
falls well below the required value. Such shaft speed 70
behaviour, although not the desired effect, was the result 50
of hundreds of simulations performed with the aim of 30
optimising the overall ship and propulsion behaviours.
10
Figure 22 represents the propeller pitch behaviour
-10
during the manoeuvre. For such a variable, the propul- 110 140 170 200 230 260 290 320 350
sion controller optimisation, performed by simulation, Time [s]
led to a smooth response even for this severe man-
oeuvre. In Figure 23, the time history of the total Figure 24. Oil pressure versus time in crash stop.

Downloaded from pim.sagepub.com at Universita di Genova on February 7, 2013

12 Proc IMechE Part M: J Engineering for the Maritime Environment 0(0)

spindle torque acting on the propeller blade is shown. 2. Altosole M, Benvenuto G, Campora U, et al. Real-time
For such a manoeuvre, the developed CPP model pre- simulation of a COGAG naval ship propulsion system.
dicted a maximum spindle torque 60% higher than the Proc IMechE, Part M: J Engineering for the Maritime
maximum ahead value. In Figure 24, the oil pressures Environment 2009; 223(1): 47–62.
inside the two activation chambers versus time are 3. Altosole M, Figari M, Michetti S, et al. Simulation of
the dynamic behavior of a CODLAG propulsion plant.
shown. For such a variable, the model predicted a
In: Proceedings of international symposium warship 2010
dynamic pressure in the actuating chamber double that conference, London, UK, 9–10 June 2010, pp.109–115.
of the maximum ahead condition, with a foreseeable London: The Royal Institution of Naval Architect.
high stress in the hydraulic system. 4. Altosole M, Benvenuto G, Martelli M, et al. Advances in
automation design for fast vessels propulsion. In: Pro-
ceedings of the 9th symposium on high speed marine vehi-
Conclusion cles, HSMV 2011, vol. I, Naples, Italy, 25–27 May 2011.
5. Carlton J. Marine propellers and propulsion. UK: Elsevier
In this article, a mathematical and numerical model Science & Technology, 2007.
able to represent the performance of the CPP mechan- 6. Chu C, Chan ZL, She YS, et al. The 3-bladed JD-CPP
ism has been illustrated. The work is mainly focussed series. In: 4th Lips Propeller symposium, Drunen, The
on the evaluation of the several loads acting on the pis- Netherlands, 4–5 October 1979, pp.53–79.
ton and on the blades of the propeller, taking into 7. Sheng CP, Chu C and Chan ZL. The estimation of blade
account the influence of the ship and CPP blade spindle torque of controllable pitch propellers. In: 4th
motions (in terms of rotation, velocity and accelera- Lips Propeller symposium, Drunen, The Netherlands, 4–5
October, 1979, pp.83–88.
tion). To this end, different reference frames have been
8. Bakker JC, Grimmelius HT and Wessenlink AF. The use
introduced in order to suitably consider each contribu-
of nonlinear models in the analysis of CPP actuator
tion of forces and moments. behavior. In: ICMES conference, London, UK, 6–10
Initial results seem positive but a scientific assessment March 2006. London: IMarEST.
of the model fidelity has not been performed yet, and it 9. Wessenlink AF, Stapersma D, Van de Bosch D, et al.
is believed very difficult to do so both in model as well Non linear aspect of propeller pitch control. In: ICMES
as in full scale. A verification of the oil pressure values, conference (eds Ales MR, Bailey A, Newell JM et al.),
for some manoeuvres, is scheduled on a naval vessel. London, UK, 6–10 March 2006, pp. 105–120. London:
The main result of the proposed procedure is the IMarEST.
assessment of the velocity of the pitch change and the 10. Godjevac M, Van Beek T, Grimmelius HT, et al. Predic-
oil pressures produced inside the two hub chambers. tion of fretting motion in a controllable pitch propeller
during service. Proc IMechE, Part M: J Engineering for
These parameters are important for the design of the
the Maritime Environment 2009; 223: 541–560.
propulsion control. In fact, the rate of change of the 11. Godjevac M. Wear and friction in a controllable pitch pro-
propeller pitch becomes a crucial aspect especially when peller, PhD Thesis, Uitgeverij BOXpress, Oisterwijk, The
the automation designer has to set the proper logic to Nether lands, 2009.
manage critical or emergency manoeuvres, such as 12. Figari M and Guedes Soares C. Fuel consumption and
severe turning circles, slam start or crash stop of the exhaust emissions reduction by dynamic pitch control. In:
vessel. In these conditions, the whole propulsion plant Analysis and design of marine structures (Guedes Soares C
simulation model has been demonstrated to be a useful & Das PK eds). (Guedes Soares C & Das PK eds) Lon-
tool in assessing the numerical value of some important don: Taylor & Francis, 2009, pp.543–550.
variables in the early stages of the design process. 13. Altosole M, Martelli M and Vignolo S. A mathematical
model of the propeller pitch change mechanism for the
marine propulsion control design. In: Proceedings of the
14th symposium on sustainable maritime transportation
Funding
and exploitation of sea resources, IMAM 2011 (eds
This research received no specific grant from any fund- Guedes Soares C & Rizzuto E), vol. II, Genoa, Italy,
ing agency in the public, commercial or not-for-profit 12–16 September 2011, pp.649–656.
sectors. 14. Geway RP, De Mulder EPHM and Wessenlink AF. The
behavior of the controllable pitch propeller mechanism.
In: 4th Lips Propeller symposium, Drunen, The Nether-
Acknowledgement lands, 4–5 October 1979, pp.101–115.
The authors wish to thank Fincantieri S.p.A., Marine 15. Wind J. Principles of mechanism used in controllable
Systems and Components Business Unit, Genoa, Italy, pitch propellers. In: 3th Lips Propeller symposium,
and Seastema S.p.A., Genoa, Italy, for the support Drunen, The Netherlands, 1978.
16. Gaggero S and Brizzolara S. A panel method for trans-
received during the research activity.
cavitating marine propellers. In: 7th International sympo-
sium on cavitation (CAV2009), Ann Arbor, MI, 17–22
August 2009.
References 17. Gaggero S, Villa D and Brizzolara S. RANS and panel
1. Fossen TI. Marine Control System. Trondheim, Norway: method for unsteady flow propeller analysis. J Hydrodyn
Marine Cybernetics, 2002. 2010; 22(5, Suppl. 1): 564–569.

Downloaded from pim.sagepub.com at Universita di Genova on February 7, 2013

Martelli et al. 13

18. Kuiper G. The Wageningen propeller series. Netherlands: yCH position of the hydrodynamic centre of
MARIN Publication 92-001, 1992. the blade along b2 (m)
19. Viviani M, Dubbioso G, Soave M, et al. Hydrodynamic zCH position of the hydrodynamic centre of
coefficients regressions analysis and application to twin the blade along b3 (m)
screw vessels. In: 13th Congress of international maritime A1 yoke area of the astern chamber (m2)
association of Mediterranean (IMAM 2009) (Gören Ö, A2 yoke area of the ahead chamber (m2)
Okan B & Karakas SC eds), Istanbul,_ Turkey, 12–15 B oil bulk modulus (Pa)
October 2009.
20. Viviani M, Altosole M, Cerruti M, et al. Marine propul- Bp damping coefficient (kg/s)
sion system dynamics during ship manoeuvres. In: 6th Cip oil leakage coefficient
International conference on high-performance marine vehi- D propeller diameter (m)
cles (Hiper 2008)(Bertorello C ed), Naples, Italy, 18–19 Dp piston diameter (m)
September 2008, pp.81–93. F generic force acting on the propeller
21. Viviani M, Podenzana Bonvino C, Mauro S, et al. Analy- blade (N)
sis of asymmetrical shaft power increase during tight FAX frictional axial force (N)
manoeuvres. In: 10th International symposium on practi- FC Coriolis force (N)
cal design of ships and other floating structures (PRADS), FI inertial force acting on the propeller
Houston, TX, 1–5 October 2007.
blade (N)
FS transportation force (N)
Appendix 1 FW weight force (N)
Notation Fhyd hydraulic force acting on the pin–slot (N)
FHD hydrodynamic force acting on the
ao acceleration of the origin O (m/s2) blade (N)
b1 , b2 , b3 b-basis FThrust forward yoke force (N)
da distance between the application point of FTraction backward yoke force (N)
the axial frictional force and the hub seat I33 inertia moment of the propeller blade
(m) around the spindle axis (kg m2)
dr rod diameter (m) Io inertia tensor (kg m2)
E orifice thickness (m) J propeller advance coefficient
e1 , e2 , e3 e-basis Kq propeller torque coefficient
eyp eccentricity of the yoke pin (m) Kt thrust propeller coefficient
f1 , f2 , f3 f-basis KQSH hydrodynamic spindle torque coefficient
g gravity acceleration (m/s2) KQSI inertial spindle torque coefficient
kf friction coefficient Lp thickness of the piston head (m)
m total mass, including propeller mass, yoke MO moment acting on the blade with respect
mass and oil mass (kg) to the origin O (Nm)
mb propeller blade mass (kg) MIO moment of the inertial forces acting on the
mo oil mass (kg) blade with respect to the origin O (Nm)
mp propeller mass (kg) MSO moment of the transportation forces
my yoke mass (kg) acting on the blade with respect to the
m_ i oil mass flow in (kg/s) origin O (Nm)
m_ o oil mass flow out (kg/s) MW
O moment of the weight forces acting on the
n propeller speed (rps) blade with respect to the origin O (Nm)
n1 , n2 , n3 n-basis (P 2 G) position vector of the generic point P of
p oil pressure inside the hub chamber (Pa) the blade with respect to the gravity centre
p1 oil pressure inside the astern chamber (Pa) G (m)
p2 oil pressure inside the ahead chamber (Pa) (P 2 O) position vector of the generic point P of
qi oil volumetric flow in (m3/s) the blade with respect to the origin O (m)
qo oil volumetric flow out (m3/s) Q total torque required by the propeller
r propeller radius (m) (Nm)
v linear velocity vector (m/s) Q* torque required by the single blade of the
vrG velocity vector of the gravity centre G (m/ propeller (Nm)
s) QFR,AX torque due to the axial frictional force
vrO velocity vector of the origin O (m/s) (Nm)
vrP velocity vector of the generic point P (m/s) Qhyd hydraulic torque acting on the pin–slot
x piston stroke (m) (Nm)
xMAX_AH maximum stroke ahead (m) QS total spindle torque (Nm)
xCH position of the hydrodynamic centre of QSF frictional component of spindle torque
the blade along b1 (m) (Nm)

Downloaded from pim.sagepub.com at Universita di Genova on February 7, 2013

14 Proc IMechE Part M: J Engineering for the Maritime Environment 0(0)

QSH hydrodynamic component of spindle Moreover, it is easily seen that the following identities
torque (Nm) hold too
QSI inertial component of spindle torque
(Nm)
Q2F torque due to the interaction forces vu ^ (P  O)½(P  O)  vH 
between propeller blade and blade bearing = (P  O) ^ ½ðvu ^ vH Þ ^ (P  O)
(Nm) + vH ^ (P  O)½vu  (P  O) ð47Þ
T propeller thrust (N)
T* thrust of the single blade of the propeller and
(N)
V actual chamber volume (m3) vH ^ ½(P  O) ^ ½vu ^ (P  O) = vH ^ vu jP  Oj2
V0 initial chamber volume (m3)
Z propeller blades number  vH ^ (P  O)½vu  (P  O) ð48Þ
b whole set of the points constituting the
blade
Making use of equations (45) to (48), we have then
u propeller shaft angular position (rad)
m oil dynamic viscosity (pa s)
r water density (kg/m3) 2(P  O) ^ ½vH ^ ½vu ^ (P  O)
rb mass density of the propeller blade (kg/
= vu ^ ½(P  O) ^ ½vH ^ (P  O)
m3)
roil oil density (kg/m3) + vu ^ vH jP  Oj2
u pitch angle (rad)
uMAX_AH maximum pitch angle ahead (rad) + (P  O) ^ ½ðvu ^ vH Þ ^ (P  O)
c ship yaw angle (rad)
vB blade inertial angular velocity (rad/s) + vH ^ (P  O)½vu  (P  O)
vH hub inertial angular velocity (rad/s) = vu ^ ½(P  O) ^ ½vH ^ (P  O)
vu shaft angular velocity (rad/s)
vu blade angular velocity (rad/s) + (P  O) ^ ½ðvu ^ vH Þ ^ (P  O)
vc ship yaw velocity (rad/s)
F reaction force due to the interaction vH ^ ½(P  O) ^ ½vu ^ (P  O)
between propeller blade and piston (N) ð49Þ

Appendix 2
Inserting the content of equation (49) into equation
For convenience of the reader, a sketch of the proofs of (44), recalling the general definition of the action of the
the formulae (20) and (21) is given. To start with, let inertia tensor on a vector a
ð ð
MC O = 2 rðP  OÞ ^ ½vH ^ ½vu ^ ðP  OÞdt IO (a) =  r(P  O) ^ ½a ^ (P  O)dt ð50Þ
b b
ð44Þ
and taking into account that the vectors vP and vu are
be the moment (with respect to the point O) of the independent of the integration variables, we end up
Coriolis force acting on a single blade. Using the skew- with the final expression
symmetric property of the vector (wedge) product
between vectors (a ^ b=b ^ a), as well as the well-    
MC
O =  vu ^ IO ðvH Þ  vH ^ IO vu + IO vu ^ vH
known formula a ^ ðb ^ cÞ = bða  cÞ  cða  bÞ (the dot
denoting scalar product between vectors), we can work ð51Þ
out the vector expression under the integral (44) in dif- where IO denotes the inertia tensor of the single blade
ferent ways. In particular, we get the identities with respect to the point O. Concerning the moment of
the transportation force with respect to the point O, we
(P  O) ^ ½vH ^ ½vu ^ (P  O) =vu
have
^ ½(P  O) ^ ½vH ^ (P  O)  vu ^ vH jP  Oj2 ð
ð45Þ MSO = r(P  O) ^ ½aO + vH ^ ½vH ^ (P  O)
b
and also
+ v_ H ^ (P  O)dt ð52Þ
(P  O) ^ ½vH ^ ½vu ^ (P  O)
On the one hand, it is immediately seen that
= vu ^ (P  O)½(P  O)  vH  ð46Þ

Downloaded from pim.sagepub.com at Universita di Genova on February 7, 2013

 Martelli et al. 15
ð
 r(P  O) ^ aO dt = M(G  O) ^ aO ð53Þ From equation (55), it follows that
ð
b
 r(P  O) ^ ½vH ^ ½vH ^ (P  O)dt =  vH ^ IO ðvH Þ
And b

ð ð56Þ
 r(P  O) ^ ½v_ H ^ (P  O)dt = IO ðv_ H Þ ð54Þ Taking equations (53), (54) and (55) into account, we
b can rewrite equation (52) in the final form
On the other hand, we have the identity MSO = M(G  O) ^ aO  vH ^ IO ðvH Þ  IO ðv_ H Þ ð57Þ

(P  O) ^ ½vH ^ ½vH ^ (P  O)

= (P  O) ^ ½vH ðvH  (P  O)Þ
= (P  O) ^ ½vH ðvH  (P  O)Þ + vH ^ vH jP  Oj2
= vH ^ ½(P  O)ðvH  (P  O) + vH jP  Oj2 Þ
= vH ^ ½(P  O) ^ ½vH ^ (P  O) ð55Þ

Downloaded from pim.sagepub.com at Universita di Genova on February 7, 2013

View publication stats