THNIS8E UNW
Laboratoum voor Archief
Mek&weg 2,2628 CD Deftt
TeL: 015-780073-Fam 015-781836
A GENERALIZED HYDRODYNA}!IC INTERACTION FORCE MODULE AND ITS USE
IN SIN1JIATÏON OF SHIP INTERACTIONS IN SHALLOW CHANNELS
Paul Kaplan, Virginia Polytechnic Institute and State University, Aerospace and Ocean Engineering Department, Blacksburg, Virginia, USA
ABSTRACT
Hydrodynamic interaction forces between 2 ships in a shallow channel are represented With the effects of differént
relative headings of each vessel,, as well as their respective
motion variables (sway and yaw velocities) included. These forces are then provided in continuous time history form for direct use in simulation studies, which contrasts with prior hydrodynamic force model, obtained from either model test
results or theoretical analysis for constrained conditions such as parallel paths, fixed lateral separation, etc. The
relationship of a computational module (providing such forces) to the simulator equations of motion; the computational
interfaces; and the incorporation of such a module into the total simulator mathematical model are discussed in terms of more realistic modeling and enhancing simulation fidelity and
accuracy.
INTRODUCTION
There have been a number of theoretical studies concerned with the hydrodynamic forces on ships in shallow canals
including interactions between 2 passing ships. The analyses have ranged from simple twodimensional approximations to more complex analysis involving integral, equations, where the ships
have been considered as moving in parallel paths as well as
parallel to the canal walls. While the degree of agreement between theory and model test data varies widely, the form of
the results as well as the effort required for- the computations
precluded utility for simulator studies. This vas a major motivation for carrying out model test studies of forces on passing ships, as was carried out for the recent Panama Canal
simulation studyr at CAORF (see Eda et al. (1986)).
A recent theoretical study by Kaplan and Sankaranarayanan (1987) considered the case of passing ships in an aSyetric canal, with the ships constrained to move parallel to each other
and to the canal walls. The method of analysis used the Lagally theorem for unsteady flow, as Well as a slender body theory lift force model, with the results determined in the form of triple
Moody (1964), also showed fairly good agreement.
All of the. work referred to ¡boye did not consider arbitrary motions of the ships, which would properly reflect their behavior in realistic cases, but only the constraint of parallel motion at fixed lateral separation distances. The most useful tool for. a simulation study would be a mathematical model that would remove that constraint and also include the inter-action effects due to the ship dynamic motions. The present paper describes the development of such a model, and how it is incorporated into a ship simulation model.
HYDRODYNA}IIC ANáLYSIS METhODS
The hydrodynamic analysis methods used potential theory techniques, similar to those used by Kaplan and Saukaranarayanan
(1987). The techniques included basic body modeling by a source distr-ibution along the centerplane; effects of images dùe to the bottom and the channel walls; induced lateral dipole distri-butions; lateral lift forces arising from slender-body theory
analysis; cross-flow drag forces; and the application of the extended unsteady form of the Lagally theorem for determining
the forces. An illustration of the general geometry of the
ships, in the casé of the constrained motion treated by Kaplan and Sankaranarayanau (1987), is given in Figure 1.
In the present case the ships are assumed to have any geometric orientation relative to each other, as well as
relative to the canal walls. The canal can have asynmietric
depth conditions, which allows export-oriented ports to consider deeper dredged regions for outgoing cargo-laden ships. The hydrodynamic analysis calculates the generalized hydrodynamic forces acting on each ship due to the effects of the other ship,
including the inflüence of the dynamic motions of each ship (i.e. sway and yaw), as well as the influence of the channel walls, when the ships operate in shallow water. The case of a single ship is also included, so the bank suction problem is
covered as well.
As an approach to reduce the computation time, the initial source distribution representing each Ship is taken as a
one-dimeflsjonal centerline distribution in place of the
previously used two-dimensional centerplane distribution. The mathematical model also includes dipole distributions for both
WZ
L
'J
F4 :iw
£ "S.Fig. 1. Definition Diagram with Co-ordinate systems.
allows an arbitrary orientation of each ship relative to the canal walls1 and all of the necessary images to account for the bottom1 walls and free surface area also established for each
ship (both sources and dipoles). The correcting dipole
distributions in each ship are then determined, recognizing the effects of varying orientation and changing positions that are possible due to the motions of each ship when they are
maneuvering.
This mathematical model is presently under development, with completion expected by Fall 1987.
Since representative results obtaifled from the theory by
Kaplan and Saukaranarayanan (1987) are shown in Figures 2-5, which illustrate the generally good predictive capability of the theory when applied to the constrained motion cases considered
there. In view of the degree of correlation obtained between theory and experiment in these conètrained motion cases, it can be expected that similarly good results will be found for the present case of generalized motion.
-90
LE.0
o C o z C 200 100 -1200--120C -!00 C 500 12:0Fig. 2. Lateral Force Comparison with SSPA Case 1: Water depth = 1.15T, Separation 1.125B.
-600 0
S?6TI
fF13CLTEC
Fig. 3. Yaw moment ..- Comparison with SSPA Case 1:
250
u -2SD s
LECE3 CLC.TED 5xP
Fig. 4. Lateral Force Comparisoù with NS Case 1:
Water depth = 1.15T, Separation = 30m.
o o n ISO loo. S0 o a LE CE NC -GCO a ECO 1200 5EPETIOti FI) -600 0 600
5ZPEC
FT)- CLCULT!O
Fig. 5. Yaw Moment - Comparison with NS Case 1:
Water depth = 1.15T, Separation = 30m. j
//
of ship horizontal plane dynamic motion effects on the resulting
hydrodynamic forces.
The detailed determination of such effects
coinputationally is only to be considered as involved due to the
great care required to represent geometric changes in a proper
manner.
Other computational aspects applied to this work include
the effects of use of a one-dimensional centerline source
distribution in place of the earlier two-dimensional centerplane
source distribution
inrepresenting the basic hulls, as well as
the method of determining time derivate term!
inthe force
expressions.
The source distribution used has its strength
proportional to the longitudinal rate of change of
cross-sectional area, While the previous source strength vas
propor-tional to the longitudinal rate of change of the lateral section
offset.
The net result of this change is to eliminate one of
the integration operations (in the vertical direction for each
term appearing in final expressions for forces), thereby
reducing the computation time.
The time derivative operation in the earlier effort by
Kaplan and Sankaranarayanan (1987) vas carried out analytically
in terms of the theoretical expressions developed there.
However, for the generalized case there would be too
manyquantities in the geometric representations of distances between
the ships that Would be changing with time; viz, longitudinal
and lateral separations, angular orientations
other motion
variables, etc. to consider an analytic evaluation, from both
the analytic formulation effort point of view as veil
as
computationaily.
The time derivative operations are then
represented by first order changes in the values of desired
quantities, i.e. integral and summation terms, at different time
values, which is a sufficiently accurate method for the
present
application.
In order to check the effects of such changes on calculated
results, some cases seen previously for the constrained
condition were evaluated using the above changes.
The
conditions considered vere 2 Panamax bulk carriers of 257
m.length, passing each other with a speed of 7 kt. for each
vessel.
The calculated results vere within 10% of each Other,
which is mainly due to the change in the form of the
source
distribution, but that is sufficiently good
accuracy for
A major benefit obtained by these procedures is in regard
to computation time.
The computation time corresponding to the
passing case described above, for determining the lateral force
and yaw moment due to interaction from the instant of bow-to-bow
meeting to that of stern-to-stern passing, was 20 sec. for the
original method used by Kaplan and Saukaranarayanan 91987).
in
the present case.the same computation was carried out in about i
sec.
While the real time for this ship passing case was about
70 sec., it must be mentioned that the computation times given
above vere for a large mainframe IBM 3090 computer which is not
the type of maàhine used in ship simulators.
However there are
some relatively fast minicomputers presently in use, or planned
for use, in some simulators that have sufficient computational
speed and capacity that can incorporate the generalized
hydrodynamic interaction force calculation within their
operation when considering fast ti.me simulation.
It is apparent.
that the computation requirements for use in the real time
simulation mode would allow direct utility of this force
calculation.
In establishing the computation of these interaction forces
within a complete ship simulation procedure, it is recognized
that the mathematical model includes all of the motion and
orientation effects of each ship, i.e. it reflects the influence
of the maneuvering of each ship.
The determination of these
interaction forces can be considered to be a separate module
that provides the time histories of the interaction forces
acting on each ship during the time when the ships are
maneuvering (on. the simulator).
This generalized hydrodynamic
interaction force module is structured so that it is linked to
the equations of motion of each ship in order to allow the
simulation model to feed the ship dynamic motions, orientation,
etc. to this force module.
The changing interaction forces will
then also be transmitted to and included in the total dynamic
models of ship motion, so that they will
inturn influence the
subsequent responses.
Thus a coupling will exist which is
considered to be a realistic modeling of ship interaction in a
purely dynamic sense, which is not present in existing simulator
models.
A simplified block diagram representation of this
com-bined computation and simulation procedure is shown in Figure 6.
CONCLUDING R(ARKS
The discussion above describes the basic analysis being
used, as well as the concepts for inclusion of a calculation
procedure for determining the generalized hydrodynainic forces
acting on a ship due to another ship as weIl as physical
boundaries in the waterway region.
These forces are not
included when modeling the conventional hydrodynamic forces
acting on a single ship (alone) when it is in either deep or
shallow water.
The inclusion of these forces, when also
accounting for the effects of the dynamic motLons of the ships
on the forces, allows a full modeling of ship interactions which
Channel. geometry
etc.
Hydrodynamic Force computation - for
each ship
Yaw orientatiOn, position and motion responses Ship i
Fig. 6. Simplified block diagram of computation - simulation
procedure.
studies. Aside from caSes of ship passing an4/or overtaking, the influence of ship orientation, "crabbing" motion, etc. for
the case of 2 ships, as well as for a .single ship near a canal wall, can be directly analyzed and plaçed in a form readily
adaptable for usein a simulator.
The calculation procedure for determining these generalized hydrodynamic interaction forces is structured as a separate
modùle that can be integrated into an overall simulation
procedure for the case of 2 ships. The input requirements will only be ship geometry; speed; canal geometry; etc. without any
need for additional special experimental or empirical data. In
this way the force module will be similar to the major elements that are used in predicting ship motions in waves (e.g see Raf f
(1972) for description of t-he SCORES ship motion and load program). Only similar type input data is necessary to apply
this theoretical model, which is a definite improvement in the
field of maneuverability analysis that has relied largely on
experimental data to establish hydrodynamic force models. The prospect of more realistic force modeling of ship interaction forces, and their use in simulator studies, provides an
opportunity for enhancing the simulation fidelity and accuracy for many problems of interest in simulator applications for restricted waterway design and analysis studies.
Equation of motion Ship 1
--I- Equation motion ofYaw orientation, position and motion responses - Ship 2
RERENCES
Eda, H.; Shizume, P.K.; Case, 1.5.; and Puglisi, J.J. (1986). A Study of Shiphandling Performance in Restricted Water;
Development and Validation of Computer Simulation Model3 Trans. SNANE.
Kaplan, P. and Sankaranarayanan (1987). Hydrodynamic
Interaction of Ships in Shallow Channels, Including Effects of Asymmetry. Proc. RINA mt. Conf. on Ship Maneuverability.
MoOdy, C.G. (1964). The Handling of Ships Through on Widened and Asymmetrically Deepened Section of Gaillard Cut in the Panama Canal. David Taylor Model Basin Rpt. 1705.
Raf f, A.I. (1.972), Program SCORES -- Ship Structural Response in Waves. Ship Structure Comm. Rpt. No. SSC-230.
Remery, G.F.M..(1974). Mooring Forces Induced by Passing Ships. Proc. Offshore Tech. Conf., Paper OTC-2066.