S. K. Bhattacharyya
e-mail: skbh@iitm.ac.in
R. Panneer Selvam
Department of Ocean Engineering,
Indian Institute of Technology,
Madras, India
Parameter Identification
of a Large Floating Body
in Random Ocean Waves
by Reverse MISO Method
Dynamics of a large moored floating body in ocean waves involves frequency dependent
added mass and radiation damping as well as the linear and nonlinear mooring line
characteristics. Usually, the added mass and radiation damping matrices can be esti-
mated either by potential theory-based calculations or by experiments. The nonlinear
mooring line properties are usually quantified by experimental methods. In this paper, we
attempt to use a nonlinear system identification approach, specifically the Reverse Mul-
tiple Inputs-Single Output (R-MISO) method, to a single-degree-of-freedom system with
linear and cubic nonlinear stiffnesses. The system mass is split into a frequency indepen-
dent and a frequency dependent component and its damping is frequency dependent. This
can serve as a model of a moored floating system with a dominant motion associated with
the nonlinear stiffness. The wave diffraction force, the excitation to the system, is assumed
known. This can either be calculated or obtained from experiments. For numerical illus-
tration, the case of floating semi-ellipsoid is adopted with dominant sway motion. The
motion as well as the loading are simulated with and without noise assuming PM spec-
trum and these results have been analyzed by the R-MISO method, yielding the frequency
dependent added mass and radiation damping, linear as well as the nonlinear stiffness
coefficients quite satisfactorily. @DOI: 10.1115/1.1493201#
Keywords: Added Mass, Damping, Nonlinear, Noise, System Identification, Random
Wave, R-MISO, Wave Spectrum
Introduction
The response of a system to an excitation depends on the pa-
rameters embedded in the equation of motion. Methods to esti-
mate these parameters form the major concern of System Identi-
fication ~SI!. Estimation of the parameters of the equation of
motion is also referred to as parameter identification. SI can be
based upon time or frequency domain approaches. In this paper, a
relatively new frequency domain nonlinear SI method, specifi-
cally, the Reverse Multiple Input-Single Output ~R-MISO!
method @1,2#, has been adopted for the problem of a large moored
floating body under random wave loading based upon a linear
spectral description. The SI problem consists of predicting the
frequency dependent added mass and radiation damping of the
floating body as a Single-Degree-Of-Freedom ~SDOF! system as
well as the linear and nonlinear stiffnesses of the system provided
by the mooring lines. Cubic nonlinearity is assumed for modeling
the mooring line stiffness. The linear stiffness can be partially due
to the hydrostatic restoring force of a floating body and partially
due to the mooring line. However, in the SDOF system consid-
ered, the hydrostatic stiffness is absent, the degree of freedom
being sway. The definition of degrees of freedom of a floating
body is shown in Fig. 1 and the schematic diagram of a moored
floating body with direction of motion considered for parameter
identification in this paper is shown in Fig. 2. The excitation and
the response time series are utilized in the SI analysis, and these
data have been simulated with and without noise in the numerical
example and the R-MISO method applied to these data.
The R-MISO method is a generalized non-iterative frequency
domain method in which conditioned spectral density functions of
the input and output are used and the roles of the input and output
are reversed to form a MISO model. It is robust, computationally
light, and requires no starting estimates. It has found application
in a variety of nonlinear systems, namely, Duffing, Van der Pol,
Mathieu, and dead-band systems @3#. It has been used for identi-
fication of parameters in nonlinear integro-differential equations
of motion @4#. R-MISO SI algorithm based on the conditioned
spectral density functions is well established @5#.
Ocean engineering applications of SI are relatively few. Kalman
filtering algorithms have been used in the identification of hydro-
dynamic coefficients associated with the Morison’s equation @6#.
Extended Kalman filtering algorithms have been developed and
applied to identify linear and nonlinear hydrodynamic coefficients
in the mathematical models of ship maneuvering from ship trial
data @7#. SI technique based on NARMAX routines has been used
to establish the possibility of a simple extension to the Morison’s
equation, which can improve wave force prediction @8,9#. Similar
investigations were also carried out to determine wave force
mechanisms on a vertical cylinder in random waves using nonlin-
ear frequency domain analysis procedure, which is similar to
MISO methods @10#.
R-MISO method in SDOF systems under random ocean waves
has been focus of some recent investigations @11–13#, wherein the
determination of the drag and inertia coefficients embedded in the
Morison’s equation formed the objective.
System Identification
System Identification involves deducing the parameters of a
system, whose governing equation of motion is known, from the
excitation and response time history. In this paper, we are con-
cerned with the nonlinear equation of motion of a SDOF system
given by
@ m1a ~
v !# x ¨ 1c ~
v ! x ˙ 1kx 1Kx
3
5 f (1)
Contributed by the OOAE Division of THE AMERICAN SOCIETY OF MECHANI-
CAL ENGINEERS for publication in the ASME JOURNAL OF OFFSHORE MECHANICS
AND ARCTIC ENGINEERING. Manuscript received by the ASME OOAE Division,
June 2001; final revision, February 2002. Associate Editor: S. Calisal.
Journal of Offshore Mechanics and Arctic Engineering MAY 2003, Vol. 125 Õ 81
Copyright © 2003 by ASME
Downloaded From: http://asmedigitalcollection.asme.org/ on 04/23/2015 Terms of Use: http://asme.org/terms