<?xml version='1.0' encoding='utf-8' ?>
<!DOCTYPE html>
<html xmlns="http://www.w3.org/1999/xhtml">
	<head>
		<meta http-equiv="Content-Type" content="text/html; charset=utf-8"/>
		<link type="text/css" rel="stylesheet" href="doxygen.css"/>
		<link type="text/css" rel="stylesheet" href="tabs.css"/>
	</head>
	<body>
	<div>
<!-- Generated by Doxygen 1.8.5 -->
</div><!-- top -->
<div class="header">
  <div class="headertitle">
<div class="title">Equations of motion </div>  </div>
</div><!--header-->
<div class="contents">
<div class="toc"><h3>Table of Contents</h3>
<ul><li class="level1"><a href="#tm_equations_kinetics">Kinetics for a rigid body moving in water</a></li>
<li class="level1"><a href="#tm_equations_coordinate">Coordinate transformation</a></li>
<li class="level1"><a href="#tm_equations_position">Position vector. Model formulation</a></li>
<li class="level1"><a href="#tm_equations_methods">Methods for numerical integration</a><ul><li class="level2"><a href="#tm_equations_methods_modified">Modified Euler method</a></li>
<li class="level2"><a href="#tm_equations_methods_3rd">3rd-order Runge-Kutta-like method</a></li>
<li class="level2"><a href="#tm_equations_methods_newmark">Newmark-Beta predictor-corrector method</a></li>
</ul>
</li>
</ul>
</div>
<div class="textblock"><h1><a class="anchor" id="tm_equations_kinetics"></a>
Kinetics for a rigid body moving in water</h1>
<p><a class="anchor" id="tm-eq-6-1"></a> </p>
<p class="formulaDsp">
<img class="formulaDsp" alt="\[ \begin{array}{ll} \boldsymbol{\dot P_B} &amp;= \boldsymbol{F} \\ \\ \boldsymbol{\dot L_B} &amp;= \boldsymbol{M} \end{array} \tag{6.1} \]" src="form_820.png"/>
</p>
<p>where:</p>
 
<table width="800">
   <tr> <td align="left" width = "9%">\( \boldsymbol{P_B} \) </td><td align="left"> linear momentum </td></tr>
   <tr> <td align="left">\( \boldsymbol{L_B} \) </td><td align="left">angular momentum about some reference 
    point </td></tr>
   <tr> <td align="left" width = "9%">\( \boldsymbol{F} \) </td><td align="left">external force </td></tr>
   <tr> <td align="left" width = "9%">\( \boldsymbol{M} \) </td><td align="left">external moment about the 
   reference point
</td></tr>
</table>
<p>As reference point for <img class="formulaInl" alt="$ \boldsymbol{L_B} $" src="form_821.png"/> and <img class="formulaInl" alt="$ \boldsymbol{M} $" src="form_822.png"/> , the origin of the body-fixed system of axes will be chosen.</p>
<p>The linear and angular momenta are given by:</p>
<p><a class="anchor" id="tm-eq-6-2"></a> </p>
<p class="formulaDsp">
<img class="formulaDsp" alt="\[ \begin{array}{ll} \boldsymbol{P_B} &amp;= \boldsymbol{m}(\boldsymbol{v} + \boldsymbol{w} \times \boldsymbol{r_c} )\\ \\ \boldsymbol{L_B} &amp;= \boldsymbol{I\omega} + \boldsymbol{mr_c} \times \boldsymbol{v} \end{array} \tag{6.2} \]" src="form_823.png"/>
</p>
<p>where:</p>
 
<table width="800">
    <tr> <td align="left" width = "9%">\( \boldsymbol{m} \) </td><td align="left">body mass</td></tr>
    <tr> <td align="left">\( \boldsymbol{v} \) </td><td align="left">velocity of body origin </td></tr>
    <tr> <td align="left">\( \boldsymbol{\omega} \) </td><td align="left">angular velocity of body </td></tr>
    <tr> <td align="left">\( \boldsymbol{r_c} \) </td><td align="left">position of body centre of mass relative to body origin </td></tr>
    <tr> <td align="left">\( \boldsymbol{I} \) </td><td align="left">body's inertia tensor with respect to body origin </td></tr>
    
</table>
<p>Using ' to denote time derivatives with respect to the body-fixed system of reference, <a class="el" href="a00012.html#tm-eq-6-1">Eq. (6.1)</a> can be written:</p>
<p><a class="anchor" id="tm-eq-6-3"></a> </p>
<p class="formulaDsp">
<img class="formulaDsp" alt="\[ \begin{array}{ll} \boldsymbol{P'_B} + \boldsymbol{\omega} \times \boldsymbol{P_B} &amp;= \boldsymbol{F} \\ \\ \boldsymbol{L'_B} + \boldsymbol{\omega} \times \boldsymbol{L_B} + \boldsymbol{v} \times \boldsymbol{P_B} &amp;= \boldsymbol{M} \end{array} \tag{6.3} \]" src="form_824.png"/>
</p>
<p>Here, the second term in both equations is an effect of the rotating frame of reference, while the third term in the second equation is due to the translation of the body origin. When the centre of mass coincides with the origin, <img class="formulaInl" alt="$ \boldsymbol{r_c} = [0, \, 0, \, 0]^T $" src="form_825.png"/> , this term will vanish since the velocity and the linear momentum then will be collinear.</p>
<p>Let the vectors in <a class="el" href="a00012.html#tm-eq-6-2">Eq. (6.2)</a> be represented by their triples of coordinates with respect to the body coordinate system and let the inertia matrix be represented by a <img class="formulaInl" alt="$ 3 \times 3 $" src="form_826.png"/> matrix with respect to the same axes.</p>
<p>Writing:</p>
<p class="formulaDsp">
<img class="formulaDsp" alt="\[ \boldsymbol{r_c} = \left( x_c, y_c, z_c \right)^T \tag{6.4} \]" src="form_827.png"/>
</p>
<p>the vector cross-multiplication occurring in <a class="el" href="a00012.html#tm-eq-6-2">Eq. (6.2)</a> is equivalent to multiplication from the left by the matrix</p>
<p class="formulaDsp">
<img class="formulaDsp" alt="\[ \boldsymbol{R} = \begin{bmatrix} 0 &amp; -z_c &amp; y_c \\ z_c &amp; 0 &amp; -x_c \\ -y_c &amp; x_c &amp; 0 \end{bmatrix} \tag{6.5} \]" src="form_828.png"/>
</p>
<p>Similarly, the operation is equivalent to left multiplication by <img class="formulaInl" alt="$ -\boldsymbol{R} = \boldsymbol{R}^T $" src="form_829.png"/> .</p>
<p>Using <img class="formulaInl" alt="$ \boldsymbol{\Gamma} $" src="form_830.png"/> to denote the <img class="formulaInl" alt="$ 3 \times 3 $" src="form_826.png"/> identity matrix and defining the matrices</p>
<p class="formulaDsp">
<img class="formulaDsp" alt="\[ \begin{array}{lll} \boldsymbol{B_{11}} &amp;= m \boldsymbol{\Gamma}, \ \boldsymbol{B_{12}} &amp;= m \boldsymbol{R}^T\\ \\ \boldsymbol{B_{21}} &amp;= m\boldsymbol{R}, \ \boldsymbol{B_{22}} &amp;= \boldsymbol{I} \end{array} \tag{6.6} \]" src="form_831.png"/>
</p>
<p><a class="el" href="a00012.html#tm-eq-6-2">Eq. (6.2)</a> can be written:</p>
<p class="formulaDsp">
<img class="formulaDsp" alt="\[ \begin{array}{ll} \boldsymbol{P_B} &amp;= \boldsymbol{B_{11}v} + \boldsymbol{B_{12}\omega} \\ \\ \boldsymbol{L_B} &amp;= \boldsymbol{B_{21}v} + \boldsymbol{B_{22}\omega} \end{array} \tag{6.7} \]" src="form_832.png"/>
</p>
<p>and <a class="el" href="a00012.html#tm-eq-6-3">Eq. (6.3)</a> becomes:</p>
<p><a class="anchor" id="tm-eq-6-8"></a> </p>
<p class="formulaDsp">
<img class="formulaDsp" alt="\[ \begin{array}{ll} \boldsymbol{B_{11}v'} + \boldsymbol{B_{12}\omega'} + \boldsymbol{\omega}\times\boldsymbol{P_B} &amp;= \boldsymbol{F} \\ \\ \boldsymbol{B_{21}v'} + \boldsymbol{B_{22}\omega} + \boldsymbol{\omega}\times\boldsymbol{L_B} + \boldsymbol{v}\times\boldsymbol{P_B} &amp;= \boldsymbol{M} \end{array} \tag{6.8} \]" src="form_833.png"/>
</p>
<p>The 6 x 6 matrix {B} =</p>
<p class="formulaDsp">
<img class="formulaDsp" alt="\[ \boldsymbol{B} = \begin{bmatrix} \boldsymbol{B_{11}} &amp; \boldsymbol{B_{12}} \\ \boldsymbol{{B_{21}}} &amp; \boldsymbol{B_{22}} \\ \end{bmatrix} \tag{6.9} \]" src="form_834.png"/>
</p>
<p>will be referred to as the body-mass matrix.</p>
<p>A detailed derivation of the result <a class="el" href="a00012.html#tm-eq-6-8">Eq. (6.8)</a> presented in a different form is given by <a class="el" href="a00014.html#bibliography">Abkowitz [20]</a>. The form of <a class="el" href="a00012.html#tm-eq-6-8">Eq. (6.8)</a> was chosen to obtain similarity with the expressions for hydrodynamic added-mass force which are presented in the following.</p>
<p>When the body is moving in a fluid, it will experience added inertia due to the acceleration of the surrounding fluid particles. The added inertia force and moment are included in the external <img class="formulaInl" alt="$ \boldsymbol{F} $" src="form_835.png"/> and <img class="formulaInl" alt="$ \boldsymbol{M} $" src="form_822.png"/> in <a class="el" href="a00012.html#tm-eq-6-8">Eq. (6.8)</a>, i.e.:</p>
<p class="formulaDsp">
<img class="formulaDsp" alt="\[ \begin{array}{ll} \boldsymbol{F} &amp;= \boldsymbol{F_1} + \boldsymbol{F_A} \\ \\ \boldsymbol{M} &amp;= \boldsymbol{M_1} + \boldsymbol{M_A} \end{array} \tag{6.10} \]" src="form_836.png"/>
</p>
<p>where subscript <img class="formulaInl" alt="$\boldsymbol{ A} $" src="form_837.png"/> refers to the added inertia.</p>
<p>Assuming the fluid to be ideal, irrotational and unbounded in all directions, <img class="formulaInl" alt="$\boldsymbol{ F_{A}} $" src="form_838.png"/> and <img class="formulaInl" alt="$\boldsymbol{ M_{A}} $" src="form_839.png"/> can be associated with added linear and angular momenta which can be expressed as:</p>
<p><a class="anchor" id="tm-eq-6-11"></a> </p>
<p class="formulaDsp">
<img class="formulaDsp" alt="\[ \begin{array}{ll} \boldsymbol{P_A} &amp;= \boldsymbol{A_{11} v} + \boldsymbol{A_{12}\omega} \\ \\ \boldsymbol{L_A} &amp;= \boldsymbol{A_{21} v} + \boldsymbol{A_{22}\omega} \end{array} \tag{6.11} \]" src="form_840.png"/>
</p>
<p>where <img class="formulaInl" alt="$ \boldsymbol{A_{ij}}, \ i=1,2 \ \text{and} \ j=1,2 $" src="form_841.png"/> are added-mass tensors with respect to the body origin. The tensors will be represented by <img class="formulaInl" alt="$ 3 \times 3 $" src="form_826.png"/> matrices in the body system of coordinates, and the <img class="formulaInl" alt="$ 6 \times 6 $" src="form_842.png"/> matrix</p>
<p class="formulaDsp">
<img class="formulaDsp" alt="\[ \boldsymbol{A} = \begin{bmatrix} \boldsymbol{A_{11}} &amp; \boldsymbol{A_{12}} \\ \boldsymbol{A_{21}} &amp; \boldsymbol{A_{22}} \end{bmatrix} \tag{6.12} \]" src="form_843.png"/>
</p>
<p>will be referred to as the added-mass matrix.</p>
<p>The added inertia force and moment are given by</p>
<p><a class="anchor" id="tm-eq-6-13"></a> </p>
<p class="formulaDsp">
<img class="formulaDsp" alt="\[ \begin{array}{ll} \boldsymbol{A_{11}v'} + \boldsymbol{A_{12} \omega'} + \boldsymbol{\omega} \times \boldsymbol{P_A} &amp;= \boldsymbol{-F_A} \\ \\ \boldsymbol{A_{21}v'} + \boldsymbol{A_{22} \omega'} + \boldsymbol{\omega} \times \boldsymbol{L_A} + \boldsymbol{v} \times \boldsymbol{P_A} &amp;= \boldsymbol{-M_A} \end{array} \tag{6.13} \]" src="form_844.png"/>
</p>
<p>which is completely analogous to <a class="el" href="a00012.html#tm-eq-6-8">Eq. (6.8)</a>. For a derivation of this result, <a class="el" href="a00014.html#bibliography">Newman [18]</a> or <a class="el" href="a00014.html#bibliography">Milne-Thomson [19]</a> can be consulted. <a class="el" href="a00012.html#tm-eq-6-13">Eq. (6.13)</a> is seen to agree with the well-known "paradox'' of D'Alembert, i.e. there is no force on a body of arbitrary shape which moves with constant velocity without rotating in an ideal fluid. However, as is also known, a moment will exist. This moment is given by the term <img class="formulaInl" alt="$ \boldsymbol{v} \times \boldsymbol{P_A} $" src="form_845.png"/> in <a class="el" href="a00012.html#tm-eq-6-13">Eq. (6.13)</a>.</p>
<p>Adding <a class="el" href="a00012.html#tm-eq-6-8">Eq. (6.8)</a> and <a class="el" href="a00012.html#tm-eq-6-13">Eq. (6.13)</a> and defining the virtual-mass matrix</p>
<p class="formulaDsp">
<img class="formulaDsp" alt="\[ \boldsymbol{V} = \begin{bmatrix} \boldsymbol{V_{11}} &amp; \boldsymbol{V_{12}} \\ \boldsymbol{V_{21}} &amp; \boldsymbol{V_{22}} \end{bmatrix} = \begin{bmatrix} \boldsymbol{A_{11}}+ \boldsymbol{B_{11}} &amp; \boldsymbol{A_{12}}+ \boldsymbol{B_{12}} \\ \boldsymbol{A_{21}}+ \boldsymbol{B_{21}} &amp; \boldsymbol{A_{22}}+ \boldsymbol{B_{22}} \\ \end{bmatrix} = \boldsymbol{A}+\boldsymbol{B} \tag{6.14} \]" src="form_846.png"/>
</p>
<p>and the virtual linear and angular momentum</p>
<p class="formulaDsp">
<img class="formulaDsp" alt="\[ \begin{array}{ll} \boldsymbol{P} &amp;= \boldsymbol{P_A} + \boldsymbol{P_B} \\ \\ \boldsymbol{L} &amp;= \boldsymbol{L_A} + \boldsymbol{L_B} \end{array} \tag{6.15} \]" src="form_847.png"/>
</p>
<p>the equation of motion becomes</p>
<p><a class="anchor" id="tm-eq-6-16"></a> </p>
<p class="formulaDsp">
<img class="formulaDsp" alt="\[ \boldsymbol{V} \begin{bmatrix} \boldsymbol{v'} \\ \boldsymbol{\omega'} \end{bmatrix} + \begin{bmatrix} \boldsymbol{\omega} \times \boldsymbol{P} \\ \boldsymbol{\omega} \times \boldsymbol{L} + \boldsymbol{v} \times \boldsymbol{P} \end{bmatrix} = \begin{bmatrix} \boldsymbol{F_1} \\ \boldsymbol{M_1} \end{bmatrix} \tag{6.16} \]" src="form_848.png"/>
</p>
<p>where <img class="formulaInl" alt="$\boldsymbol{ P} $" src="form_849.png"/> and <img class="formulaInl" alt="$\boldsymbol{ L} $" src="form_850.png"/> are obtained from</p>
<p class="formulaDsp">
<img class="formulaDsp" alt="\[ \begin{bmatrix} \boldsymbol{P} \\ \boldsymbol{L} \\ \end{bmatrix} = \boldsymbol{V} \begin{bmatrix} \boldsymbol{v} \\ \boldsymbol{\omega} \end{bmatrix} \tag{6.17} \]" src="form_851.png"/>
</p>
<p>The accelerations occurring in <a class="el" href="a00012.html#tm-eq-6-16">Eq. (6.16)</a> give the rate of change of velocity relative to the rotating body system. It is desirable to refer the accelerations to a fixed frame of reference. The relationship is</p>
<p><a class="anchor" id="tm-eq-6-18"></a> </p>
<p class="formulaDsp">
<img class="formulaDsp" alt="\[ \begin{array}{ll} \boldsymbol{v'} &amp;= \boldsymbol{\dot v} - \boldsymbol{\omega} \times \boldsymbol{v} \\ \\ \boldsymbol{\omega'} &amp;= \boldsymbol{\dot \omega} - \boldsymbol{\omega} \times \boldsymbol{\omega} = \boldsymbol{\dot \omega} \end{array} \tag{6.18} \]" src="form_852.png"/>
</p>
<p>Substituting <a class="el" href="a00012.html#tm-eq-6-18">Eq. (6.18)</a> in <a class="el" href="a00012.html#tm-eq-6-16">Eq. (6.16)</a> and solving for the acceleration gives:</p>
<p><a class="anchor" id="tm-eq-6-19"></a> </p>
<p class="formulaDsp">
<img class="formulaDsp" alt="\[ \begin{bmatrix} \boldsymbol{\dot v} \\ \boldsymbol{\dot \omega} \end{bmatrix} = \boldsymbol{V^{-1}} \begin{bmatrix} -\boldsymbol{\omega} \times \boldsymbol{P} + \boldsymbol{F_1} \\ -\boldsymbol{\omega} \times \boldsymbol{L} - \boldsymbol{v} \times \boldsymbol{P} + \boldsymbol{M_1} \end{bmatrix} + \begin{bmatrix} \boldsymbol{\omega} \times \boldsymbol{v} \\ \boldsymbol{0} \end{bmatrix} \tag{6.19} \]" src="form_853.png"/>
</p>
<p>The added-mass force depends on the body's velocity relative to the fluid. Thus when a homogeneous steady current is present, this relative velocity must be used in the expressions for the added-mass force <a class="el" href="a00012.html#tm-eq-6-11">Eq. (6.11)</a> and <a class="el" href="a00012.html#tm-eq-6-13">Eq. (6.13)</a>. It can be verified that the body-mass force is not influenced by the addition of a constant vector to the body velocity (in spite of the apparent dependency of velocity through the presence of <img class="formulaInl" alt="$\boldsymbol{ v} $" src="form_854.png"/> in <img class="formulaInl" alt="$\boldsymbol{ P_{B}} $" src="form_855.png"/> and <img class="formulaInl" alt="$\boldsymbol{ L_{B}} $" src="form_856.png"/> ). As a consequence, <a class="el" href="a00012.html#tm-eq-6-19">Eq. (6.19)</a> and equations preceding it can be used as they stand, provided <img class="formulaInl" alt="$\boldsymbol{ v} $" src="form_854.png"/> is interpreted as the relative velocity.</p>
<p>The assumptions which led to the expressions for added-mass force does not hold in real life. As regards the assumption of ideal-fluid properties, it is commonly assumed that the effect of viscosity can be treated separately and simply added to the ideal-fluid forces. Hence the viscous forces will be contained in the vectors <img class="formulaInl" alt="$\boldsymbol{ F_{1}} $" src="form_857.png"/> and <img class="formulaInl" alt="$ \boldsymbol{M_{1}} $" src="form_858.png"/> in <a class="el" href="a00012.html#tm-eq-6-19">Eq. (6.19)</a>. Care must be taken, however, if drag-force relationships obtained from experiments are used. It is then likely that <img class="formulaInl" alt="$ \boldsymbol{M_{_{1}}} $" src="form_859.png"/> will include the ideal-fluid effects which are already taken care of by the term <img class="formulaInl" alt="$ \boldsymbol{v} \times \boldsymbol{P_A} $" src="form_845.png"/> in <a class="el" href="a00012.html#tm-eq-6-13">Eq. (6.13)</a> which is included in <img class="formulaInl" alt="$ \boldsymbol{v} \times \boldsymbol{P} $" src="form_860.png"/> in <a class="el" href="a00012.html#tm-eq-6-19">Eq. (6.19)</a>. In this case, therefore, the appropriate components of <img class="formulaInl" alt="$ \boldsymbol{v} \times \boldsymbol{P_A} $" src="form_845.png"/>should be set to zero. Thus, if quadratic current drag coefficients are given for a body, the yaw component of <img class="formulaInl" alt="$ \boldsymbol{v} \times \boldsymbol{P_A} $" src="form_845.png"/> is set to zero.</p>
<p>For surface vessels, the assumption of unboundedness of the fluid is violated, as there will be a free surface. Due to wave-making, memory effects will exist, i.e. the hydrodynamic force will not only be a function of the instantaneous velocity and acceleration. In the frequency domain this is known as frequency dependence of added mass. In this case, the high-frequency limit of the added mass should be used in <a class="el" href="a00012.html#tm-eq-6-19">Eq. (6.19)</a> and the effect of frequency dependence be included in <img class="formulaInl" alt="$ \boldsymbol{F_{1}} $" src="form_861.png"/> and <img class="formulaInl" alt="$ \boldsymbol{M_{1}} $" src="form_858.png"/> as convolution integrals.</p>
<p>When there are rigid boundaries in the vicinity of the body, e.g. sea-bottom, walls or other bodies, the results in this chapter does not apply since the fluid forces will depend on the body's position and orientation and not only on its linear and angular velocities. A general treatment of this problem is complicated. Some special cases in the context of ship steering are treated by <a class="el" href="a00014.html#bibliography">Norrbin [21]</a>.</p>
<p>When body coordinates are used to represent the vectors on the right-hand side of <a class="el" href="a00012.html#tm-eq-6-19">Eq. (6.19)</a>, and the matrix <img class="formulaInl" alt="$ \boldsymbol{V} $" src="form_501.png"/> is defined accordingly, the coordinates of <img class="formulaInl" alt="$ \boldsymbol{\dot v} $" src="form_862.png"/> and <img class="formulaInl" alt="$ \boldsymbol{\dot \omega} $" src="form_863.png"/> also refer to the body system - or rather, the fixed system which at that particular moment coincides with the rotating body system. To express the acceleration vectors in coordinates of the fixed, global system, a coordinate transformation is necessary.</p>
<h1><a class="anchor" id="tm_equations_coordinate"></a>
Coordinate transformation</h1>
<p>The orientation of the body system of axes is thought of as being derived from an original identification with the global system by three rotations in succession.</p>
<ul>
<li>First, the body is rotated by an angle <img class="formulaInl" alt="$ \psi $" src="form_487.png"/> about the common z-axis. This alters the direction of the body x-axis and y-axis.</li>
<li>Second, the body is rotated by an angle <img class="formulaInl" alt="$ \theta $" src="form_864.png"/> about its y-axis. This changes the orientation of the x-axis and the z-axis, but the body y-axis still lies in the global xy-plane.</li>
<li>Finally, the body is rotated by an angle <img class="formulaInl" alt="$ \phi $" src="form_865.png"/> about its x-axis, altering the orientation of the y-a-xis and z-axis.</li>
</ul>
<p>The orientation of the body system is now uniquely determined by the angles, <img class="formulaInl" alt="$ \phi $" src="form_865.png"/> , <img class="formulaInl" alt="$ \theta $" src="form_864.png"/> , <img class="formulaInl" alt="$ \psi $" src="form_487.png"/> and the specified order of rotation.</p>
<p>The matrix for transformation of a vector's coordinates in the body system to the representation in the global system becomes:</p>
<p><a class="anchor" id="tm-eq-6-20"></a> </p>
<p class="formulaDsp">
<img class="formulaDsp" alt="\[ \boldsymbol{\Lambda} = \begin{bmatrix} \cos\psi\cos\theta &amp; -\sin\psi\cos\phi + \cos\psi\sin\theta\sin\phi &amp; \sin\psi\sin\phi + \cos\psi\sin\theta\cos \phi \\ \sin\psi\cos\theta &amp; \cos\psi\cos\phi + \sin\psi\sin\theta\sin\phi &amp; -\cos\psi\sin\phi + \sin\psi\sin\theta\cos \phi \\ -\sin\theta &amp; \cos\theta\sin\phi &amp; \cos\theta\cos\phi \end{bmatrix} \tag{6.20} \]" src="form_866.png"/>
</p>
<p>i.e. if a vector is represented by <img class="formulaInl" alt="$ \boldsymbol{x^{(B)}} $" src="form_718.png"/> in the body system, the representation in the global system is</p>
<p class="formulaDsp">
<img class="formulaDsp" alt="\[ \boldsymbol{x^{(G)}} = \boldsymbol{\Lambda x^{(B)}} \tag{6.21} \]" src="form_867.png"/>
</p>
<p>Conversely,</p>
<p class="formulaDsp">
<img class="formulaDsp" alt="\[ \boldsymbol{x^{(B)}} = \boldsymbol{\Lambda^T x^{(G)}} \tag{6.22} \]" src="form_868.png"/>
</p>
<p>since <img class="formulaInl" alt="$ \boldsymbol{\Lambda^{-1}} = \boldsymbol{\Lambda^T}$" src="form_869.png"/></p>
<h1><a class="anchor" id="tm_equations_position"></a>
Position vector. Model formulation</h1>
<p>The position vector of the body is defined as</p>
<p class="formulaDsp">
<img class="formulaDsp" alt="\[ \boldsymbol{x} = \left( x,y,z,\phi,\theta,\psi \right)^T \tag{6.23} \]" src="form_870.png"/>
</p>
<p>where <img class="formulaInl" alt="$ x,y,z $" src="form_871.png"/> are coordinates of the body origin with respect to the global coordinate system (Strictly, x is not a vector since <img class="formulaInl" alt="$ \phi, \theta $" src="form_872.png"/> and <img class="formulaInl" alt="$ \psi $" src="form_487.png"/> are not coordinates)</p>
<p>It is desirable to express the dynamic model in the form:</p>
<p><a class="anchor" id="tm-eq-6-24"></a> </p>
<p class="formulaDsp">
<img class="formulaDsp" alt="\[ \boldsymbol{\ddot x} = \boldsymbol{f(x,\dot x, \xi)} \tag{6.24} \]" src="form_873.png"/>
</p>
<p>where <img class="formulaInl" alt="$ \boldsymbol{\xi} $" src="form_874.png"/> is a vector of inputs not depending on <img class="formulaInl" alt="$ \boldsymbol{x} $" src="form_795.png"/> or <img class="formulaInl" alt="$ \boldsymbol{\dot x} $" src="form_875.png"/> , e.g. speed and direction of wind and current, undisturbed surface elevation. When the body is coupled to other bodies, <img class="formulaInl" alt="$ x $" src="form_658.png"/> will include the positions or velocities of these bodies as well.</p>
<p>The formulation <a class="el" href="a00012.html#tm-eq-6-24">Eq. (6.24)</a> is preferred to that of the state-space formulation, since a wider class of numerical methods for solution is available.</p>
<p>Consider the vectors:</p>
<p class="formulaDsp">
<img class="formulaDsp" alt="\[ \begin{array}{ll} \boldsymbol{\dot x} &amp;= \left( \dot x, \dot y, \dot\phi, \dot\theta, \dot\psi \right) ^T \\ \\ \boldsymbol{\ddot x} &amp;= \left( \ddot x, \ddot y, \ddot\phi, \ddot\theta, \ddot\psi \right) ^T \end{array} \tag{6.25} \]" src="form_876.png"/>
</p>
<p>The sub-vectors <img class="formulaInl" alt="$ (\dot x, \dot y, \dot z)^T $" src="form_877.png"/> and <img class="formulaInl" alt="$ (\ddot x, \ddot y, \ddot z)^T $" src="form_878.png"/> are the vectors <img class="formulaInl" alt="$ \boldsymbol{v} $" src="form_533.png"/> and <img class="formulaInl" alt="$ \boldsymbol{\dot v} $" src="form_862.png"/> of <a class="el" href="a00012.html#tm-eq-6-19">Eq. (6.19)</a> in global-system representation, i.e.</p>
<p class="formulaDsp">
<img class="formulaDsp" alt="\[ \begin{bmatrix} \dot x \\ \dot y \\ \dot z \\ \end{bmatrix} = \boldsymbol{\Lambda v} , \quad \begin{bmatrix} \ddot x \\ \ddot y \\ \ddot z \\ \end{bmatrix} = \boldsymbol{\Lambda \dot v} \tag{6.26} \]" src="form_879.png"/>
</p>
<p>Let the vector <img class="formulaInl" alt="$ \boldsymbol{\omega} $" src="form_880.png"/> in <a class="el" href="a00012.html#tm-eq-6-19">Eq. (6.19)</a> be expressed as</p>
<p class="formulaDsp">
<img class="formulaDsp" alt="\[ \boldsymbol{\omega} = (p,q,r)^T \tag{6.27} \]" src="form_881.png"/>
</p>
<p>in the body system. The following relationship can be shown to exist, cf. <a class="el" href="a00014.html#bibliography">Norrbin [21]</a>:</p>
<p class="formulaDsp">
<img class="formulaDsp" alt="\[ \begin{array}{ll} \dot \phi &amp;= p +q\sin\phi \tan\theta + r\cos\phi\tan\theta \\ \\ \dot \theta &amp;= q\cos\phi - r\sin\phi \\ \\ \dot \psi &amp;= q\sin\phi\sec\theta + r\cos\phi\sec\theta \end{array} \tag{6.28} \]" src="form_882.png"/>
</p>
<p><a class="anchor" id="tm-eq-6-29"></a> </p>
<p class="formulaDsp">
<img class="formulaDsp" alt="\[ \boldsymbol{\gamma} = (\phi, \theta, \psi)^T \tag{6.29} \]" src="form_883.png"/>
</p>
<p><a class="el" href="a00012.html#tm-eq-6-29">Eq. (6.29)</a> can be expressed in vector form:</p>
<p><a class="anchor" id="tm-eq-6-31"></a> </p>
<p class="formulaDsp">
<img class="formulaDsp" alt="\[ \boldsymbol{\dot \gamma} = \boldsymbol{f}(\boldsymbol{\gamma},\boldsymbol{\omega}) = \boldsymbol{M}(\boldsymbol{\gamma}) \boldsymbol{\omega} \tag{6.30} \]" src="form_884.png"/>
</p>
<p>where:</p>
<p class="formulaDsp">
<img class="formulaDsp" alt="\[ \boldsymbol{M} = \begin{bmatrix} 1 &amp; \sin\phi \tan\theta &amp; \cos\phi \tan\theta \\ 0 &amp; \cos\phi &amp; -\sin\phi \\ 0 &amp; \sin\phi \sec\theta &amp; \cos\phi \sec\theta \end{bmatrix} \tag{6.31} \]" src="form_885.png"/>
</p>
<p>Differentiating <a class="el" href="a00012.html#tm-eq-6-31">Eq. (6.31)</a> with respect to time gives</p>
<p><a class="anchor" id="tm-eq-6-32"></a> </p>
<p class="formulaDsp">
<img class="formulaDsp" alt="\[ \boldsymbol{\ddot\gamma} = \displaystyle \frac{\partial \boldsymbol{f}}{\partial\boldsymbol{\gamma}} \boldsymbol{\dot\gamma}+ \boldsymbol{M\dot\omega} = \displaystyle \frac{\partial \boldsymbol{f}}{\partial\boldsymbol{\gamma}} \boldsymbol{M\omega} + \boldsymbol{M\dot\omega} = \begin{bmatrix} \ddot \phi \\ \ddot \theta \\ \ddot \psi \end{bmatrix} \tag{6.32} \]" src="form_886.png"/>
</p>
<p>where</p>
<p class="formulaDsp">
<img class="formulaDsp" alt="\[ \displaystyle \frac{\partial \boldsymbol{f}}{\partial\boldsymbol{\gamma}} = \begin{bmatrix} \left( q\cos\phi-r\sin\phi \right)\tan\theta &amp; \left( q\sin\phi+r\cos\phi \right) \sec^2\theta &amp; 0 \\ -q\sin\phi - r\cos\phi &amp; 0 &amp; 0 \\ \left( q\cos\phi-r\sin\phi \right)\sec\theta &amp; \left( q\sin\phi+r\cos\phi \right) \tan\theta\sec\theta &amp; 0 \\ \end{bmatrix} \tag{6.33} \]" src="form_887.png"/>
</p>
<p>Assuming, <img class="formulaInl" alt="$ \boldsymbol{x} $" src="form_795.png"/> , <img class="formulaInl" alt="$ \boldsymbol{\xi} $" src="form_874.png"/> and <img class="formulaInl" alt="$ \boldsymbol{\dot x} $" src="form_875.png"/> are known, <a class="el" href="a00012.html#tm-eq-6-24">Eq. (6.24)</a> is realized as follows:</p>
<ol type="1">
<li>The velocity <img class="formulaInl" alt="$ (\dot x, \dot y, \dot z)^T $" src="form_877.png"/> is transformed to body-coordinate representation <img class="formulaInl" alt="$ \boldsymbol{v} $" src="form_533.png"/> using the transpose of the matrix <img class="formulaInl" alt="$ \boldsymbol{\Lambda} $" src="form_888.png"/> in <a class="el" href="a00012.html#tm-eq-6-19">Eq. (6.19)</a>.</li>
<li><img class="formulaInl" alt="$ \boldsymbol{\omega} $" src="form_880.png"/> is calculated from <img class="formulaInl" alt="$ \dot\gamma $" src="form_889.png"/> using the inverse of <a class="el" href="a00012.html#tm-eq-6-31">Eq. (6.31)</a>.</li>
<li>The external force <img class="formulaInl" alt="$ \boldsymbol{F_{1}}$" src="form_890.png"/> and moment <img class="formulaInl" alt="$ \boldsymbol{M _{1}} $" src="form_891.png"/> in <a class="el" href="a00012.html#tm-eq-6-18">Eq. (6.18)</a> are calculated using <img class="formulaInl" alt="$ \boldsymbol{x} $" src="form_795.png"/>, <img class="formulaInl" alt="$ \boldsymbol{v} $" src="form_533.png"/>, <img class="formulaInl" alt="$ \boldsymbol{\omega} $" src="form_880.png"/> and <img class="formulaInl" alt="$ \xi $" src="form_892.png"/> .</li>
<li>The accelerations <img class="formulaInl" alt="$ \boldsymbol{\dot v} $" src="form_862.png"/> and <img class="formulaInl" alt="$ \boldsymbol{\dot\omega} $" src="form_893.png"/> are obtained from <a class="el" href="a00012.html#tm-eq-6-18">Eq. (6.18)</a>.</li>
<li><img class="formulaInl" alt="$ \boldsymbol{\dot v} $" src="form_862.png"/> is transformed to a representation in global system using the matrix <img class="formulaInl" alt="$ \boldsymbol{\Lambda} $" src="form_888.png"/></li>
<li><img class="formulaInl" alt="$ \boldsymbol{\ddot \gamma} $" src="form_894.png"/> is calculated using <a class="el" href="a00012.html#tm-eq-6-32">Eq. (6.32)</a>.</li>
</ol>
<p>Thus <img class="formulaInl" alt="$ \boldsymbol{\ddot x} = \left( \boldsymbol{\dot v^T}, \boldsymbol{\ddot \gamma^T} \right)^T $" src="form_895.png"/> is established.</p>
<h1><a class="anchor" id="tm_equations_methods"></a>
Methods for numerical integration</h1>
<p>Three methods are available. In the following description <img class="formulaInl" alt="$ T $" src="form_435.png"/> is the time step. The following shortened notation is used:</p>
<p class="formulaDsp">
<img class="formulaDsp" alt="\[ \begin{array}{ll} \boldsymbol{x_k} &amp;= \boldsymbol{x}(t_k) \\ \\ \boldsymbol{\dot x_k} &amp;= \boldsymbol{\dot x}(t_k) \\ \\ \boldsymbol{\xi_k} &amp;= \boldsymbol{\xi} (t_k) \\ \\ \boldsymbol{f_k} &amp;= \boldsymbol{f} (\boldsymbol{x_k, \dot x_k, \xi_k}) \end{array} \tag{6.34} \]" src="form_896.png"/>
</p>
<h2><a class="anchor" id="tm_equations_methods_modified"></a>
Modified Euler method</h2>
<p><a class="anchor" id="tm-eq-6-35"></a> </p>
<p class="formulaDsp">
<img class="formulaDsp" alt="\[ \begin{array}{lll} \boldsymbol{\dot x_{k+1}} &amp;= \boldsymbol{\dot x_k} + T\boldsymbol{f_k} &amp; \text{(a)} \\ \\ \boldsymbol{x_{k+1}} &amp;= \boldsymbol{x_k} + T \boldsymbol{\dot x_{k+1}} &amp; \text{(b)} \end{array} \tag{6.35} \]" src="form_897.png"/>
</p>
<p>The method differs from the ordinary method in that <img class="formulaInl" alt="$ \boldsymbol{\dot x_{k+1}} $" src="form_898.png"/> is used instead of <img class="formulaInl" alt="$ \boldsymbol{\dot x_k} $" src="form_899.png"/> in <a class="el" href="a00012.html#tm-eq-6-35">Eq. (6.35 b)</a>. This ensures stability when the method is applied to linear models with no damping.</p>
<h2><a class="anchor" id="tm_equations_methods_3rd"></a>
3rd-order Runge-Kutta-like method</h2>
<p class="formulaDsp">
<img class="formulaDsp" alt="\[ \begin{array}{lll} \Delta \boldsymbol{\dot x_1} &amp;= T\boldsymbol{f_k} &amp; \text{(a)} \\ \\ \Delta \boldsymbol{x_1} &amp;= T\boldsymbol{\dot x_k} + \frac{1}{4} T^2\boldsymbol{f_k} &amp; \text{(b)} \\ \\ \boldsymbol{f'_{k+1}} &amp;= \boldsymbol{f}(\boldsymbol{x_k} +\Delta\boldsymbol{x_1}, \ \boldsymbol{\dot x_k} + \Delta \boldsymbol{\dot x_1}, \ \boldsymbol{\xi_{k+1}}) &amp; \text{(c)} \\\\ \Delta \boldsymbol{\dot x_2} &amp;= T\boldsymbol{f'_{k+1}} &amp;\text{(d)} \\ \\ \Delta\boldsymbol{x_2} &amp;= T \left(\boldsymbol{x_k} + \frac{5}{12}\Delta\boldsymbol{x_1} \right) + \frac{1}{3} T^2\boldsymbol{f'_{k+1}} &amp; \text{(e)} \\ \\ \boldsymbol{\dot x_{k+1}} &amp;= \boldsymbol{\dot x_k} + \frac{1}{2}(\Delta \boldsymbol{\dot x_1} + \Delta \boldsymbol{\dot x_2}) &amp; \text{(f)} \\ \\ \boldsymbol{x_{k+1}} &amp;= \boldsymbol{x_k} + \frac{1}{2} (\Delta \boldsymbol{x_1} + \Delta \boldsymbol{x_2}) &amp; \text{(g)} \end{array} \tag{6.36} \]" src="form_900.png"/>
</p>
<h2><a class="anchor" id="tm_equations_methods_newmark"></a>
Newmark-Beta predictor-corrector method</h2>
<p>Predictor (cf. <a class="el" href="a00012.html#tm-eq-6-35">Eq. (6.35)</a>)</p>
<p class="formulaDsp">
<img class="formulaDsp" alt="\[ \begin{array}{ll} \boldsymbol{\dot x_{k+1}^{(0)}} = \boldsymbol{\dot x_k} + T\boldsymbol{f_k} \\ \\ \boldsymbol{x_{k+1}^{(0)}} = \boldsymbol{x_k} + T\boldsymbol{ \dot x_{k+1}^{(0)}} \end{array} \tag{6.37} \]" src="form_901.png"/>
</p>
<p>Corrector: <a class="anchor" id="tm-eq-6-38"></a> </p>
<p class="formulaDsp">
<img class="formulaDsp" alt="\[ \begin{array}{lll} \boldsymbol{f_{k+1}^{(i)}} &amp;= \boldsymbol{f}\left( \boldsymbol{x_{k+1}^{(i)}}, \ \boldsymbol{\dot x_{k+1}^{(i)}}, \ \boldsymbol{\xi_{k+1}}\right) &amp; \text{(a)} \\ \\ \boldsymbol{\dot x_{k+1}^{(i+1)}} &amp;= \boldsymbol{\dot x_k} + T (1-\gamma)\boldsymbol{f_k} + \gamma \boldsymbol{f_{k+1}^{(i)}} &amp; \text{(b)} \\ \\ \boldsymbol{ x_{k+1}^{(i+1)}} &amp;= \boldsymbol{x_k} + T\boldsymbol{ \dot x_k} + \left( \frac{1}{2}-\beta\right) T^2\boldsymbol{ f_k} + \beta T^2\boldsymbol{ f_{k+1}^{(i)}} &amp; \text{(c)} \end{array} \tag{6.38} \]" src="form_902.png"/>
</p>
<p>Steps (a) - (c) are repeated a predefined number of times, and as a future extension until</p>
<p class="formulaDsp">
<img class="formulaDsp" alt="\[ \begin{array}{ll} \left| \boldsymbol{x_{k+1}^{(i+1)}} - \boldsymbol{x_{k+1}^{(1)}} \right| &amp;< \boldsymbol{\epsilon} \\ \\ \left| \boldsymbol{\dot x_{k+1}^{(i+1)}} - \boldsymbol{\dot x_{k+1}^{(1)}} \right| &amp;< \displaystyle \frac{\boldsymbol{\epsilon}}{T} \end{array} \tag{6.39} \]" src="form_903.png"/>
</p>
<p>where <img class="formulaInl" alt="$ \boldsymbol{\epsilon} $" src="form_904.png"/> is a vector of specified numbers.</p>
<p>The parameter <img class="formulaInl" alt="$ \gamma $" src="form_118.png"/> governs the algorithmic damping in the numerical integration:</p>
<ul>
<li><img class="formulaInl" alt="$ \gamma > \frac{1}{2} $" src="form_905.png"/> gives positive damping</li>
<li><img class="formulaInl" alt="$ \gamma = \frac{1}{2} $" src="form_906.png"/> gives no damping</li>
<li><img class="formulaInl" alt="$ \gamma < \frac{1}{2} $" src="form_907.png"/> gives negative damping</li>
</ul>
<p>The parameter <img class="formulaInl" alt="$ \beta $" src="form_908.png"/> in <a class="el" href="a00012.html#tm-eq-6-38">Eq. (6.38 c)</a> should be chosen in the interval <img class="formulaInl" alt="$ [0, 0.5] $" src="form_909.png"/> . With <img class="formulaInl" alt="$ \gamma = 1/2 $" src="form_910.png"/> , the following known integration methods can be obtained:</p>
<p class="formulaDsp">
<img class="formulaDsp" alt="\[ \begin{array}{lll} \beta &amp; = 0 &amp; \text{: second central difference} \\ \\ \beta &amp; = \frac{1}{12} &amp; \text{: Fox-Goodwins method}\\ \\ \beta &amp; = \frac{1}{6} &amp; \text{: Linear acceleration}\\ \\ \beta &amp; = \frac{1}{4} &amp; \text{: Constant average acceleration (trapezoidal method), unconditionally stable.} \end{array} \]" src="form_911.png"/>
</p>
 </div></div><!-- contents -->
	</body>
</html>