Chapter 5 – Seakeeping Theory

4.1  Hydrodynamic  Concepts  and  Poten5al  Theory  

4.2  Seakeeping  and  Maneuvering  Kinema5cs  
4.3  The  Classical  Frequency-­‐Domain  Model  
4.4  Time-­‐Domain  Models  including  Fluid  Memory  Effects  

1

Lecture Notes TTK 4190 Guidance and Control of Vehicles (T. I. Fossen)

 Chapter 5 - Seakeeping Theory
Equa5ons  of  Mo5on  
Seakeeping  theory  is  formulated  in  equilibrium  (SEAKEEPING)  axes  {s}  but  it  can  be  
transformed  to  BODY  axes  {b}  by  including  fluid  memory  effects  represented  by  
impulse  response  func5ons.    
The  transforma5on  is  is  done  within  a  linear  framework  such  that  addi5onal  nonlinear  
viscous  damping  must  be  added  in  the  5me-­‐domain  under  the  assump5on  of  linear  
superposi5on.    

!MRB ! MA "!" ! C RB #!$! ! C A #! r $!r

Inertia forces:




Damping forces: !#D p ! D V $! r ! D n #!r $!r ! "
Restoring forces: !g##$ ! g o
Wind and wave forces: # $ wind ! $ wave
Propulsion forces: !$

μ  is  an  addi5onal  term  represen5ng  the  fluid  memory  effects.  

2

Lecture Notes TTK 4190 Guidance and Control of Vehicles (T. I. Fossen)

 5.1 Hydrodynamic Concepts and Potential
Theory
Strip  Theory  (2-­‐D  Poten5al  Theory)  
For  slender  bodies,  the  mo5on  of  the  fluid  can  be  formulated    as  a  2-­‐D  problem.    An  
accurate  es5mate  of  the  hydrodynamic  forces  can  be  obtained  by  applying    strip  theory  
(Newman,  1977;  Fal/nsen,  1990;  Journee  and  Massie,  2001).    
 
 
 
 
 
 




 
 
The  2-­‐D  theory  takes  into  account  that  varia5on  of  the  flow  in  the  cross-­‐direc5onal  plane  
is  much  larger  than  the  varia5on  in  the  longitudinal  direc5on  of  the  ship.  
 
The  principle  of  strip  theory  involves  dividing  the  submerged  part  of  the  craZ  into  a  finite  
number  of  strips.  Hence,  2-­‐D  hydrodynamic  coefficients  for  added  mass  can  be  computed  
for  each  strip  and  then  summed  over  the  length  of  the  body  to  yield  the  3-­‐D  coefficients.  
 
Commercial  Codes:  MARINTEK  (ShipX-­‐Veres)  and  Amarcon  (Octopus  Office)  
3

Lecture Notes TTK 4190 Guidance and Control of Vehicles (T. I. Fossen)

 ShipX (VERES) by MARINTEK
MARINTEK  -­‐  the  Norwegian  Marine  Technology  Research  
Ins5tute  -­‐  does  research  and  development  in  the  
mari5me  sector  for  industry  and  the  public  sector.  The  
Ins5tute  develops  and  verifies  technological  solu5ons  
for  the  shipping  and  mari5me  equipment  industries  and  
for  offshore  petroleum  produc5on.    

VERES  -­‐  VEssel  RESponse  program  is  a  Strip  Theory  Program  which  calculates  wave-­‐induced  




loads  on  and  mo5ons  of  mono-­‐hulls  and  barges  in  deep  to  very  shallow  water.  The  
program  is  based  on  the  famous  paper  by  Salvesen,  Tuck  and  Fal?nsen  (1970).  Ship  
Mo/ons  and  Sea  Loads.  Trans.  SNAME.  

4

Lecture Notes TTK 4190 Guidance and Control of Vehicles (T. I. Fossen)

 ShipX (VERES) by MARINTEK
ShipX  (Veres)

5

Lecture Notes TTK 4190 Guidance and Control of Vehicles (T. I. Fossen)

 OCTOPUS SEAWAY by Amarcon
and  AMARCON  cooperate  in  further  development  of  SEAWAY  


The  Mari5me  Research  Ins5tute  Netherlands  (MARIN)  and  
AMARCON  agree  to  cooperate  in  further  development  of  
SEAWAY.  MARIN  is  an  interna5onally  recognized  authority  on  
hydrodynamics,  involved  in  fron5er  breaking  research  programs  
for  the  mari5me  and  offshore  industries  and  navies.  






SEAWAY  is  developed  by  Professor  J.  M.  J.  Journée  at  the  DelZ  Univiversity  of  Technology  
 
SEAWAY  is  a  Strip  Theory  Program  to  calculate  wave-­‐induced  loads  on  and  mo5ons  of  mono-­‐
hulls  and  barges  in  deep  to  very  shallow  water.  When  not  accoun5ng  for  interac5on  
effects  between  the  hulls,  also  catamarans  can  be  analyzed.  Work  of  very  acknowledged  
hydromechanic  scien5sts  (such  as  Ursell,  Tasai,  Frank,  Keil,  Newman,  Fal5nsen,  Ikeda,  etc.)  
has  been  used,  when  developing  this  code.  
 
SEAWAY  has  extensively  been  verified  and  validated  using  other  computer  
codes  and  experimental  data.

6

Lecture Notes TTK 4190 Guidance and Control of Vehicles (T. I. Fossen)



7

7
Lecture Notes TTK 4190 Guidance and Control of Vehicles (T. I. Fossen)

 5.1 Hydrodynamic Concepts and Potential
Theory
Panel  Methods  (3-­‐D  Poten5al  Theory)  
For  poten5al  flows,  the  integrals  over  the  fluid  domain  can  be  transformed  to  integrals  
over  the  boundaries  of  the  fluid  domain.  This  allows  the  applica5on  of  panel  or  boundary  
3D Visualization of the Wamit file: supply.gdf

element  methods  to  solve  the  3-­‐D  poten5al  theory  problem.    

  3D Panelization of
  a Supply Vessel
4
 
2

0
 
-2
 




Z-axis (m)

-4

-6  
 
-8

-10

 
-12

-40

-30
 
 
-20
3
-10
20

 
0 10
10 0

 
20 -10

30 -20

Panel  methods  divide  the  surface  of  the  ship  and  the  surrounding  water  into  discrete  
40 -30
X-axis (m) Y-axis (m)

elements  (panels).  On  each  of  these  elements,  a  distribu5on  of  sources  and  sinks  is  
defined  which  fulfill  the  Laplace  equa5on.    
 
Commercial  code:    WAMIT  (www.wamit.com)  
8

Lecture Notes TTK 4190 Guidance and Control of Vehicles (T. I. Fossen)

 WAMIT
WAMIT® is the most advanced set of tools available for analyzing wave interactions with
offshore platforms and other structures or vessels.



WAMIT® was developed by Professor Newman and coworkers at MIT in 1987, and it has
gained widespread recognition for its ability to analyze the complex structures with a high
degree of accuracy and efficiency.

Panelization of semi-submersible using WAMIT user supplied tools

Over the past 20 years WAMIT has been licensed to more than
80 industrial and research organizations worldwide.

9


Lecture Notes TTK 4190 Guidance and Control of Vehicles (T. I. Fossen)

 5.1 Hydrodynamic Concepts and Potential
Theory
Poten5al  theory  programs  typically  compute:  
 
•  Frequency-­‐dependent  added  mass,  A(w)  
•  Poten5al  damping  coefficients,  B(w)  
•  Restoring  terms,  C  
•  1st-­‐  and  2nd-­‐order  wave-­‐induced  forces  and  mo5ons    
(amplitudes  and  phases)  for  given  wave  direc5ons  and  frequencies  
•  …  and  much  more  
 



One  special  feature  of  WAMIT  is  that  the  program  solves  a  boundary  value  problem  
for  zero  and  infinite  added  mass.  These  boundary  values  are  par5cular  useful  when  
compu5ng  the  retarda5on  func5ons  describing  the  fluid  memory  effects.  
 
Processing  of  Hydrodynamic  Data  using  MSS  HYDRO  –  www.marinecontrol.org  
The  toolbox  reads  output  data  files  generated  by  the  hydrodynamic  programs:  
 
•       ShipX  (Veres)  by  MARINTEK  AS  
•       WAMIT  by  WAMIT  Inc.  
 
and  processes  the  data  for  use  in  Matlab/Simulink.    
10
 
Lecture Notes TTK 4190 Guidance and Control of Vehicles (T. I. Fossen)

 5.2 Seakeeping and Maneuvering
Kinematics
Seakeeping  Theory  (Perturba5on  Coordinates)  
The  SEAKEEPING  reference  frame  {s}  is  not  fixed  to  
the  craZ;    it  is  fixed  to  the  equilibrium  state:  

! ! !! 1 , ! 2 , ! 3 , ! 4 , ! 5 , ! 6 " T #

! ! !"

!" ! !#" ! !$




-­‐ In  the  absence  of  wave  excita5on,  {s}  coincides  with  {b}.  
-­‐   Under  the  ac5on  of  the  waves,  the  hull  is  disturbed  from  
! sb ! !! 4 , ! 5 , ! 6 " ! ! !"#, "#, "$" ! #its  equilibrium  and  {s}  oscillates,  with  respect  to  its  
equilibrium  posi5on.    

Transforma5on  between  {b}  and  {s}   v nns ! !U cos !, U sin!, 0" ! #

! nns ! !0, 0, 0" ! #
0 0 0 0 0 0
!! ! ! !U!L!" " e 1 " # " "!
" ns ! !0, 0, ! #
0 0 0 0 0 1
!!# ! !# !UL! #
0 0 0 0 !1 0
L :! ! 0 $!
0 0 0 0 0 0
0 0 0 0 0 0 " ! 0 # $" #
e 1 ! !1, 0,0,0, 0, 0" ! # #
" $#
0 0 0 0 0 0

11

Lecture Notes TTK 4190 Guidance and Control of Vehicles (T. I. Fossen)

 5.3 The Classical Frequency-Domain Model
Seakeeping  Analysis  
The  seakeeping  equa5ons  of  mo5on  are  considered  to  be  iner5al:  

Equa5ons  of  Mo5on  

 
  ! ! !" ! !!x, !y, !z, !", !#, !$" ! #



 
  MRB !" ! #hyd " #hs " #exc #
 
   
Cummins   (1962)    showed  that  the  radia5on-­‐induced  hydrodynamic  forces  in  an  ideal  
 fluid  can  be  related  to  frequency-­‐dependent  added  mass  A(ω)  and  poten5al  damping  
 B(ω)  according  to:  
   

!hyd ! !Ā"# ! " K
t
$ !t ! !""%!!"d! #
0

12

Lecture Notes TTK 4190 Guidance and Control of Vehicles (T. I. Fossen)

 5.3 The Classical Frequency-Domain Model

Frequency-­‐dependent  added  mass  A22(ω)  and  poten5al  damping  B22(ω)  in  sway  

Ā ! A!!"

13

Lecture Notes TTK 4190 Guidance and Control of Vehicles (T. I. Fossen)

 5.3 The Classical Frequency-Domain Model
t
!hyd ! !Ā"# ! " K
$ !t ! !""#!!"d! #
0

Matrix  of  retarda5on  func5ons  given  by  

"
2
K!t" ! ! ! 0 B!"" cos!"t"d" #

Cummins  Model  
 
 




If  linear  restoring  forces  τhs =  -Cξ    are  included  in  the  model,  this  results  in  the  
5me-­‐domain  model:  
   
  t
!MRB ! A!!""! ! " K
" # !t # !"!$!!"d! ! C! " %exc #
  0

   

The  fluid  memory  effects  can  be  replaced  by  a  

state-­‐space  model  to  avoid  the  integral

14

Lecture Notes TTK 4190 Guidance and Control of Vehicles (T. I. Fossen)

 5.3 The Classical Frequency-Domain Model

15
Longitudinal  added  mass  coefficients  as  a  func5on  of  frequency.    
Lecture Notes TTK 4190 Guidance and Control of Vehicles (T. I. Fossen)

 5.3 The Classical Frequency-Domain Model

16
Lateral  added  mass  coefficients  as  a  func5on  of  frequency.    
Lecture Notes TTK 4190 Guidance and Control of Vehicles (T. I. Fossen)

 5.3 The Classical Frequency-Domain Model

Longitudinal  poten5al  damping  coefficients  as  a  func5on  of  frequency.    

17
Exponen5al  decaying  viscous  damping  is  included  for  B11.  
Lecture Notes TTK 4190 Guidance and Control of Vehicles (T. I. Fossen)

 5.3 The Classical Frequency-Domain Model

Lateral  poten5al  damping  coefficients  as  a  func5on  of  frequency.  

18
Exponen5al  decaying  viscous  damping  is  included  for  B22  and  B66  while  viscous  IKEDA  damping  is  included  in  B44    
Lecture Notes TTK 4190 Guidance and Control of Vehicles (T. I. Fossen)

 5.3.1 Potential Coefficients and the
Concept of Forced Oscillations
MRB !" ! #hyd " #hs " f cos!!t" #  harmonic  excita5on  

In  an  experimental  setup  with  a  restrained  scale  model,  it  is  possible  to  vary  the  wave  
excita5on  frequency  ω  and  the  amplitudes  fi  of  the  excita5on  force.  Hence,  by  measuring  
the  posi5on  and  aptude  vector  η,  the  response  of  the  2nd-­‐order  order  system  can  be  
fiqed  to  a  linear  model:  

!MRB ! A"!#$!" ! B"!#!# ! C! " f cos"!t# #

for  each  frequency  ω.

 
The  matrices  A(ω),B(ω)  and  C  represents  a  "hydrodynamic  mass-­‐damper-­‐spring  system"  
which  varies  with  the  frequency  of  the  forced  oscilla5on.  
 
This  model  is  rooted  deeply  in  the  literature  of  hydrodynamics  and  the  abuse  of  nota5on  of  
this  false  5me-­‐domain  model  has  been  discussed  eloquently  in  the  literature  (incorrect  
mixture  of  5me  and  frequency  in  an  ODE).      
Consequently,  we  will  use  Cummins    5me-­‐domain  model  and  transform  this  model  to  the  
frequency  domain  –  no  mixture  of  5me  and  frequency!  
19

Lecture Notes TTK 4190 Guidance and Control of Vehicles (T. I. Fossen)

 5.3.2 Frequency-Domain Seakeeping
Models
t
!MRB ! A!!""!" ! " K
# !t # !"!$!!"d! ! C! " %exc #
0

Cummins  equa5on  can  be  transformed  to  the  frequency  domain  

(Newman,  1977;  Fal5nsen  1990)  according  to:  
 
 
!!! 2 "MRB ! A!!#$ ! j!B!!# ! C#!!j!# " "exc !j!# #
 
 



where  the  complex  response  and  excita5on  variables  are  wriqen  as:  

! i !t" ! !" i cos!"t # # i " $ ! i !j"" ! !" i exp!j# i " #

$ exc,i !t" ! $" i cos!"t # % i " $ $ exc,i !j"" ! $" exc,i exp!j% i " #

The  poten5al  coefficients  A(ω)  and  B(ω)  are  usually  computed  using  a  seakeeping  program  
but  the  frequency  response  will  not  be  accurate  unless  viscous  damping  is  included.    
 
The  op5onal  viscous  damping  matrix  BV(ω)    can  be  used  to  model  viscous  damping  such  as  
skin  fric5on,  surge  resistance  and  viscous  roll  damping  (for  instance  IKEDA  roll  damping).    
20

Lecture Notes TTK 4190 Guidance and Control of Vehicles (T. I. Fossen)

 5.3.2 Frequency-Domain Seakeeping
Models
Viscous  frequency-­‐dependent  damping:  
!! 2 !MRB ! A"!#$ ! j!B total"!# ! C !"j!# " "wave "j!# ! "wind "j!# ! """j!# #

B total!!" ! B!!" " B V !!" #

" 1 e !#! "N ITTC !A 1 " 0 0 0 0 0

Viscous  skin  fric5on:

" 2 e !#!
! i e !"#
0

0 0


0 0

B V !!" !
0

0
0

0
0

0
0

" IKEDA!!"
0

0
0

0
# Quadra5c  ITTC  drag:

X ! !X |u|u |u|u
0 0 0 0 0 0
" N ITTC !A1 "u #
0 0 0 0 0 " 6 e !#!

Quadra5c  damping  is  approximated  using  describing  func5ons  

(similar  to  the  equivalent  lineariza5on  method):  
u ! Asin!!t" # 2
y ! N!A"u # N!A" ! c 1 " 8A c 2 " 3A c 3 #
y ! c 1 x " c 2 x|x|"c 3 x 33 # 3! 4
21

Lecture Notes TTK 4190 Guidance and Control of Vehicles (T. I. Fossen)

 5.4 Time-Domain Models including Fluid
Memory Effects
Cummins  equa?on  in  SEAKEEPING  coordinates  
(linear  theory  which  includes  fluid  memory  effects)    

Transform  from  SEAKEEPING  to  BODY  coordinates  

(linearized  kinema?c  transforma?on)  




Linear  seakeeping  equa?ons  in  BODY  coordinates  
(fluid  memory  effects  are  approximated  as  state-­‐space  models)  

Unified  maneuvering  and  seakeeping  model  

 (nonlinear  viscous  damping/maneuvering  coefficients  
 are  added  manually)  
22

Lecture Notes TTK 4190 Guidance and Control of Vehicles (T. I. Fossen)

 5.4.1 Cummins Equation in SEAKEEPING
Coordinates
Cummins  (1962)  Equa5on  

t
!MRB ! Ā"!" ! ! # !t " !"!$!!"d! ! C
K # ! " %wind ! %wave ! "% #
"#

The  Ogilvie  (1964)  Transforma5on  gives

Ā ! A!!" #

"




! !t" ! 2
K ! ! 0 B total!"" cos!"t"d" #

From  a  numerical  point  of  view  is  it  beqer  to  integrate  the  difference:  
 
  K!t" ! 2 ! "#B !"" # B !""$ cos!"t"d" #
  ! 0 total total

 
This  can  be  don  by  rewri5ng  Cummins  equa5on  as:      
t
!MRB ! A"!#$!" ! B total"!#!# ! " K"t # !#!#"!#d! ! C! " $wind ! $wave ! "$ #
0
23

Lecture Notes TTK 4190 Guidance and Control of Vehicles (T. I. Fossen)

 5.4.2 Linear Time-Domain Seakeeping
Equations in BODY Coordinates
t
!MRB ! A"!#$!" ! B total"!#!# ! " K"t # !#!#"!#d! ! C! " $wind ! $wave ! "$ #
0

It  is  possible  to  transform  the  5me-­‐domain  representa5on  of  Cummins  equa5on  from  {s}  to  {b}  
using  the  kinema5c  rela5onships:  
 
  !! ! ! !U!L!" " e 1 " # ! ! !"
  !!# ! !# !UL! # !" ! !#
 
This  gives:  




  t
!MRB   ! A"!#$!!" !UL!$ ! B total"!#!! !U"L!# " e 1 #$ ! # K"t " "#!!""#d" ! C!# " $wind ! $wave ! "$ " $%# #
0
 
The  steady-­‐state  control  force  τ  needed  to  obtain  the  forward  speed  U  when  τwind  =  τwave=  0  
and  δη  =  0  is:  
 
!" ! B total!!"Ue 1 #
 
Hence,  
t
!MRB ! A"!#$!!" !UL!$ ! B total"!#!! ! UL!#$ ! " K"t # "#!!""#d" ! C!# " $wind ! $wave ! $ #
0
24

Lecture Notes TTK 4190 Guidance and Control of Vehicles (T. I. Fossen)

 5.4.2 Linear Time-Domain Seakeeping
Equations in BODY Coordinates
t
!MRB ! A"!#$!!" !UL!$ ! B total"!#!! ! UL!#$ ! " K"t # "#!!""#d" ! C!# " $wind ! $wave ! $ #
0

When  compu5ng  the  damping  and  retarda5on  func5ons,  it  is  common  to  neglect  the  
influence  of  δη  on  the  forward  speed  such  that:  
 
  !! ! v !U!L!" " e 1 " ! v "Ue 1 #
 




Finally,  let  use  replace  ν  by  the  rela5ve  velocity  νr  to  include  ocean  currents  and  define:  
M  =  MRB  +  MA  such  that  :  
  t
  M!" ! C !RB ! ! C !A !r ! D!r ! " K!t # !"#!!!"#Ue 1 $d! ! G# " $wind ! $wave ! $ #
0
 
where  
 
MA ! A!!"
 
"
  C A ! UA!!"L
  C "RB ! UMRB L Linear  Coriolis  and  
D ! B total!!" centripetal  forces  due  
G!C to  a  rota5on  of  {b}  about  {s}  
25

Lecture Notes TTK 4190 Guidance and Control of Vehicles (T. I. Fossen)

 5.4.2 Linear Time-Domain Seakeeping
Equations in BODY Coordinates
Fluid  Memory  Effects  
The  integral  in  the  following  equa5on  represents  the  fluid  memory  effects:  

t
M!" ! C !RB ! ! C !A !r ! D!r ! " K!t # !"#!!!"#Ue 1 $d! ! G# " $wind ! $wave ! $ #
0

! :!
t
! 0 K!t " !"#"!!""Ue 1 $
""
d! #


! ! H!s"#" !Ue 1 $ #
Approximated  by  a  state-­‐
x 10
7 K22(t)
space  model

x! " A rx # B r!"
2.5

2 #
1.5
! " C rx
1

0.5

Impulse  response  func5on

0
"
-0.5 2
K!t" ! ! ! 0 #B!"" # B!""$ cos!"t"d" #
-1
0 5 10 15 20 25
time (s)

26

Lecture Notes TTK 4190 Guidance and Control of Vehicles (T. I. Fossen)

 5.4.3 Nonlinear Unified Seakeeping and
Maneuvering Model with
Fluid Memory Effects
Linear  Seakeeping  Equa5ons  (BODY  coordinates)  

M!" ! C !RB ! ! C !A !r ! D!r ! # ! G$ " %wind ! %wave ! % # Copyright © Bjarne Stenberg/NTNU

Unified  Nonlinear  Seakeeping  and  Maneuvering  Model  


 
•   Use  nonlinear  kinema5cs   Copyright © The US Navy

•   Replace  linear  Coriolis  and  centripetal  forces  with  their  nonlinear  counterparts  
•   Include  maneuvering  coefficients  in  a  nonlinear  damping  matrix  (linear  superposi5on)  
 
!" ! J " !!"# #
M#"r # C RB !#"# # C A !#r "#r # D!#r"#r # $ # G! ! % wind # %wave # % #

27

Lecture Notes TTK 4190 Guidance and Control of Vehicles (T. I. Fossen)

 5.5 Case Study: Identification of Fluid
Memory Effects
The  fluid  memory  effects  can  be  approximated  using  frequency-­‐domain  iden5fica5on.  The  
main  tool  for  this  is  the  MSS  FDI  toolbox  (Perez  and  Fossen  2009)  -­‐  www.marinecontrol.org  
 
When  using  the  frequency-­‐domain  approach,  the  property  that  the  mapping:   !! ! " #
has  rela5ve  degree  one  is  exploited.  Hence,  the  fluid  memory  effects  μ  can  be  
approximated  by  a  matrix  H(s)  containing  rela5ve  degree  one  transfer  func5ons:  

! ! H!s"!" #

h ij !s" !
p r sr "p r!1 sr!1 "..."p 0
sn "q n!1 sn!1 "..."q 0




r ! n ! 1, n " 2

State-­‐space  model:  

H!s" ! C r!sI ! A r" !1 B r #

x! " A r x # B r !"
#
! " C rx
28

Lecture Notes TTK 4190 Guidance and Control of Vehicles (T. I. Fossen)

 5.5.1 Frequency-Domain Identification
using the MSS FDI Toolbox
Consider  the  FPSO  data  set  in  the  MSS  toolbox  (FDI  tool)  and  assumes  that  the  infinite  
added  mass  matrix  is  unknown.  Hence,  we  can  es5mate  the  fluid  transfer  func5on  h33(s)  
by  using  the  following  Matlab  code:  

load fpso
Dof = [3,3]; %Use coupling 3-3 heave-heave
Nf = length(vessel.freqs);
W = vessel.freqs(1:Nf-1)';
Ainf = vessel.A(Dof(1),Dof(2),Nf); % Ainf computed by WAMIT

A = reshape(vessel.A(Dof(1),Dof(2),1:Nf-1),1,length(W))';
B = reshape(vessel.B(Dof(1),Dof(2),1:Nf-1),1,length(W))';




FDIopt.OrdMax = 20;
FDIopt.AinfFlag = 0;
FDIopt.Method = 2;
FDIopt.Iterations = 20;
FDIopt.PlotFlag = 0;
FDIopt.LogLin = 1;
FDIopt.wsFactor = 0.1;
FDIopt.wminFactor = 0.1;
FDIopt.wmaxFactor = 5;

[KradNum,KradDen,Ainf] = FDIRadMod(W,A,0,B,FDIopt,Dof)
29

Lecture Notes TTK 4190 Guidance and Control of Vehicles (T. I. Fossen)

 5.5.1 Frequency-Domain Identification
using the MSS FDI Toolbox

FPSO  iden5fica5on  results  for  h₃₃(s)  without  using  the  infinite  added  mass  A₃₃(∞).  The  leZ-­‐hand-­‐side  plots  show  
30
the  complex  coefficient  and  its  es5mate  while  added  mass  and  damping  are  ploqed  on  the  right-­‐hand-­‐side.  
Lecture Notes TTK 4190 Guidance and Control of Vehicles (T. I. Fossen)

 5.5.1 Frequency-Domain Identification
using the MSS FDI Toolbox

1. 672e007 s 3 " 2. 286e007 s 2 " 2. 06e006 s ! ! H!s"#" !Ue 1 $ #

h 33 !s" ! 4
s " 1. 233 s 3 " 0. 7295 s 2 " 0. 1955 s " 0. 01639

x! " A rx # B r!"
#
!1. 2335 !0. 7295 !0. 1955 !0. 0164 ! " C rx

Ar !
1
0
0
1

00
00

0 0 1 0

1
0
Br !
0
0

Cr ! 1. 672e007 2. 286e007 2. 06e006 0

Dr ! 0
31

Lecture Notes TTK 4190 Guidance and Control of Vehicles (T. I. Fossen)