Eigenfunction Matching for a Circular Floating Elastic Plate
Introduction
We show here a solution for a Floating Elastic Plate on Finite Depth water based on Peter, Meylan and Chung 2004. A solution for Shallow Depth was given in Zilman and Miloh 2000 and we will also show this. The solution is an extension of the Eigenfunction Matching for a Circular Dock.
Governing Equations
We begin with the Frequency Domain Problem for a Floating Elastic Plate in the non-dimensional form of Tayler 1986 (Dispersion Relation for a Floating Elastic Plate) We will use a cylindrical coordinate system, [math]\displaystyle{ (r,\theta,z) }[/math], assumed to have its origin at the centre of the circular plate which has radius [math]\displaystyle{ a }[/math]. The water is assumed to have constant finite depth [math]\displaystyle{ H }[/math] and the [math]\displaystyle{ z }[/math]-direction points vertically upward with the water surface at [math]\displaystyle{ z=0 }[/math] and the sea floor at [math]\displaystyle{ z=-H }[/math]. The boundary value problem can therefore be expressed as
[math]\displaystyle{ \Delta\phi=0, \,\, -H\lt z\lt 0, }[/math]
[math]\displaystyle{ \phi_{z}=0, \,\, z=-H, }[/math]
[math]\displaystyle{ (\beta\Delta^{2}+1-\alpha\gamma)\phi_{z}=\alpha\phi, \,\, z=0,\,r\lt a }[/math]
where the constants [math]\displaystyle{ \beta }[/math] and [math]\displaystyle{ \gamma }[/math] are given by
[math]\displaystyle{ \beta=\frac{D}{\rho\,L^{4}g}, \gamma=\frac{\rho_{i}h}{\rho\,L} }[/math]
and [math]\displaystyle{ \rho_{i} }[/math] is the density of the plate. We must also apply the edge conditions for the plate and the Sommerfeld Radiation Condition as [math]\displaystyle{ r\rightarrow\infty }[/math]. The subscript [math]\displaystyle{ z }[/math] denotes the derivative in [math]\displaystyle{ z }[/math]-direction.
Solution Method
Separation of variables
We now separate variables, noting that since the problem has circular symmetry we can write the potential as
[math]\displaystyle{ \phi(r,\theta,z)=\zeta(z)\sum_{n=-\infty}^{\infty}\rho_{n}(r)e^{i n \theta} }[/math]
Applying Laplace's equation we obtain
[math]\displaystyle{ \zeta_{zz}+\mu^{2}\zeta=0 }[/math]
so that:
[math]\displaystyle{ \zeta=\cos\mu(z+H) }[/math]
where the separation constant [math]\displaystyle{ \mu^{2} }[/math] must satisfy the Dispersion Relation for a Free Surface
[math]\displaystyle{ k\tan\left( kH\right) =-\alpha,\quad r\gt a\,\,\,(1) }[/math]
and the Dispersion Relation for a Floating Elastic Plate
[math]\displaystyle{ \kappa\tan(\kappa H)=\frac{-\alpha}{\beta\kappa^{4}+1-\alpha\gamma},\quad r\lt a \,\,\,(2) }[/math]
Note that we have set [math]\displaystyle{ \mu=k }[/math] under the free surface and [math]\displaystyle{ \mu=\kappa }[/math] under the plate. We denote the positive imaginary solution of (1) by [math]\displaystyle{ k_{0} }[/math] and the positive real solutions by [math]\displaystyle{ k_{m} }[/math], [math]\displaystyle{ m\geq1 }[/math]. The solutions of (2) will be denoted by [math]\displaystyle{ \kappa_{m} }[/math], [math]\displaystyle{ m\geq-2 }[/math]. The fully complex solutions with positive imaginary part are [math]\displaystyle{ \kappa_{-2} }[/math] and [math]\displaystyle{ \kappa_{-1} }[/math] (where [math]\displaystyle{ \kappa_{-1}=\overline{\kappa_{-2}} }[/math]), the negative imaginary solution is [math]\displaystyle{ \kappa_{0} }[/math] and the positive real solutions are [math]\displaystyle{ \kappa_{m} }[/math], [math]\displaystyle{ m\geq1 }[/math]. We define
[math]\displaystyle{ \phi_{m}\left( z\right) =\frac{\cos k_{m}(z+H)}{\cos k_{m}H},\quad m\geq0 }[/math]
as the vertical eigenfunction of the potential in the open water region and
[math]\displaystyle{ \psi_{m}\left( z\right) =\frac{\cos\kappa_{m}(z+H)}{\cos\kappa_{m}H},\quad m\geq-2 }[/math]
as the vertical eigenfunction of the potential in the plate covered region. For later reference, we note that:
[math]\displaystyle{ \int\nolimits_{-H}^{0}\phi_{m}(z)\phi_{n}(z) d z=A_{m}\delta_{mn} }[/math]
where
[math]\displaystyle{ A_{m}=\frac{1}{2}\left( \frac{\cos k_{m}H\sin k_{m}H+k_{m}H}{k_{m}\cos ^{2}k_{m}H}\right) }[/math]
and
[math]\displaystyle{ \int\nolimits_{-H}^{0}\phi_{n}(z)\psi_{m}(z) d z=B_{mn} }[/math]
where
We now solve for the function [math]\displaystyle{ \rho_{n}(r) }[/math]. Using Laplace's equation in polar coordinates we obtain
[math]\displaystyle{ \frac{\mathrm{d}^{2}\rho_{n}}{\mathrm{d}r^{2}}+\frac{1}{r} \frac{\mathrm{d}\rho_{n}}{\mathrm{d}r}-\left( \frac{n^{2}}{r^{2}}+\mu^{2}\right) \rho_{n}=0 }[/math]
where [math]\displaystyle{ \mu }[/math] is [math]\displaystyle{ k_{m} }[/math] or [math]\displaystyle{ \kappa_{m}, }[/math] depending on whether [math]\displaystyle{ r }[/math] is greater or less than [math]\displaystyle{ a }[/math]. We can convert this equation to the standard form by substituting [math]\displaystyle{ y=\mu r }[/math] to obtain
[math]\displaystyle{ y^{2}\frac{\mathrm{d}^{2}\rho_{n}}{\mathrm{d}y^{2}}+y\frac{\mathrm{d}\rho_{n} }{\rm{d}y}-(n^{2}+y^{2})\rho_{n}=0 }[/math]
The solution of this equation is a linear combination of the modified Bessel functions of order [math]\displaystyle{ n }[/math], [math]\displaystyle{ I_{n}(y) }[/math] and [math]\displaystyle{ K_{n}(y) }[/math] Abramowitz and Stegun 1964. Since the solution must be bounded we know that under the plate the solution will be a linear combination of [math]\displaystyle{ I_{n}(y) }[/math] while outside the plate the solution will be a linear combination of [math]\displaystyle{ K_{n}(y) }[/math]. Therefore the potential can be expanded as
[math]\displaystyle{ \phi(r,\theta,z)=\sum_{n=-\infty}^{\infty}\sum_{m=0}^{\infty}a_{mn}K_{n} (k_{m}r)e^{i n\theta}\phi_{m}(z), \;\;r\gt a }[/math]
and
[math]\displaystyle{ \phi(r,\theta,z)=\sum_{n=-\infty}^{\infty}\sum_{m=-2}^{\infty}b_{mn} I_{n}(\kappa_{m}r)e^{i n\theta}\psi_{m}(z), \;\;r\lt a }[/math]
where [math]\displaystyle{ a_{mn} }[/math] and [math]\displaystyle{ b_{mn} }[/math] are the coefficients of the potential in the open water and the plate covered region respectively.
Incident potential
The incident potential is a wave of amplitude [math]\displaystyle{ A }[/math] in displacement travelling in the positive [math]\displaystyle{ x }[/math]-direction. The incident potential can therefore be written as
[math]\displaystyle{ \phi^{\mathrm{I}} =\frac{A}{i\sqrt{\alpha}}e^{k_{0}x}\phi_{0}\left( z\right) =\sum\limits_{n=-\infty}^{\infty}e_{n}I_{n}(k_{0}r)\phi_{0}\left(z\right) e^{i n \theta} }[/math]
where [math]\displaystyle{ e_{n}=A/\left(i\sqrt{\alpha}\right) }[/math] (we retain the dependence on [math]\displaystyle{ n }[/math] for situations where the incident potential might take another form).
Boundary conditions
The boundary conditions for the plate also have to be considered. The vertical force and bending moment must vanish, which can be written as
[math]\displaystyle{ \left[\bar{\Delta}-\frac{1-\nu}{r}\left(\frac{\partial}{\partial r} +\frac{1}{r}\frac{\partial^{2}}{\partial\theta^{2}}\right)\right] w=0\,\,\,(3) }[/math]
and
[math]\displaystyle{ \left[ \frac{\partial}{\partial r}\bar{\Delta}-\frac{1-\nu}{r^{2}}\left( -\frac{\partial}{\partial r}+\frac{1}{r}\right) \frac{\partial^{2}} {\partial\theta^{2}}\right] w=0 \,\,\,(4) }[/math]
where [math]\displaystyle{ w }[/math] is the time-independent surface displacement, [math]\displaystyle{ \nu }[/math] is Poisson's ratio, and [math]\displaystyle{ \bar{\Delta} }[/math] is the polar coordinate Laplacian
[math]\displaystyle{ \bar{\Delta}=\frac{\partial^{2}}{\partial r^{2}}+\frac{1}{r}\frac{\partial }{\partial r}+\frac{1}{r^{2}}\frac{\partial^{2}}{\partial\theta^{2}} }[/math]
Displacement of the plate
The surface displacement and the water velocity potential at the water surface are linked through the kinematic boundary condition
[math]\displaystyle{ \phi_{z}=-i\sqrt{\alpha}w,\,\,\,z=0 }[/math]
From the equations the potential and the surface displacement are therefore related by
[math]\displaystyle{ w=i\sqrt{\alpha}\phi,\quad r\gt a }[/math]
and
[math]\displaystyle{ (\beta\bar{\Delta}^{2}+1-\alpha\gamma)w=i\sqrt{\alpha}\phi,\quad r\lt a }[/math]
The surface displacement can also be expanded in eigenfunctions as
[math]\displaystyle{ w(r,\theta)=\sum_{n=-\infty}^{\infty}\sum_{m=0}^{\infty}i\sqrt{\alpha} a_{mn}K_{n}(k_{m}r)e^{i n\theta},\;\;r\gt a }[/math]
and:
[math]\displaystyle{ w(r,\theta)= \sum_{n=-\infty}^{\infty}\sum_{m=-2}^{\infty}i\sqrt{\alpha}(\beta\kappa _{m}^{4}+1-\alpha\gamma)^{-1}b_{mn}I_{n}(\kappa_{m}r)e^{i n\theta},\; r\lt a }[/math]
using the fact that
[math]\displaystyle{ \bar{\Delta}\left( I_{n}(\kappa_{m}r)e^{i n\theta}\right) =\kappa_{m} ^{2}I_{n}(\kappa_{m}r)e^{i n\theta}\,\,\,(5) }[/math]
An infinite dimensional system of equations
The boundary conditions (3) and (4) can be expressed in terms of the potential using (5). Since the angular modes are uncoupled the conditions apply to each mode, giving
[math]\displaystyle{ \sum_{m=-2}^{\infty}(\beta\kappa_{m}^{4}+1-\alpha\gamma)^{-1}b_{mn} \left(\kappa_{m}^{2}I_{n}(\kappa_{m}a)-\frac{1-\nu}{a}\left(\kappa _{m}I_{n}^{\prime}(\kappa_{m}a)-\frac{n^{2}}{a}I_{n}(\kappa_{m}a)\right) \right) =0\,\,\,(6) }[/math]
[math]\displaystyle{ \sum_{m=-2}^{\infty}(\beta\kappa_{m}^{4}+1-\alpha\gamma)^{-1}b_{mn} \left( \kappa_{m}^{3}I_{n}^{\prime}(\kappa_{m}a)+n^{2}\frac{1-\nu}{a^{2} }\left( -\kappa_{m}I_{n}^{\prime}(\kappa_{m}a)+\frac{1}{a}I_{n}(\kappa _{m}a)\right) \right) =0\,\,\,(7) }[/math]
The potential and its derivative must be continuous across the transition from open water to the plate covered region. Therefore, the potentials and their derivatives at [math]\displaystyle{ r=a }[/math] have to be equal. Again we know that this must be true for each angle and we obtain
[math]\displaystyle{ e_{n}I_{n}(k_{0}a)\phi_{0}\left( z\right) + \sum_{m=0}^{\infty} a_{mn} K_{n}(k_{m}a)\phi_{m}\left( z\right) =\sum_{m=-2}^{\infty}b_{mn}I_{n}(\kappa_{m}a)\psi_{m}(z) }[/math]
and
[math]\displaystyle{ e_{n}k_{0}I_{n}^{\prime}(k_{0}a)\phi_{0}\left( z\right) +\sum _{m=0}^{\infty} a_{mn}k_{m}K_{n}^{\prime}(k_{m}a)\phi_{m}\left( z\right) =\sum_{m=-2}^{\infty}b_{mn}\kappa_{m}I_{n}^{\prime}(\kappa_{m}a)\psi _{m}(z) }[/math]
for each [math]\displaystyle{ n }[/math]. We solve these equations by multiplying both equations by [math]\displaystyle{ \phi_{l}(z) }[/math] and integrating from [math]\displaystyle{ -H }[/math] to [math]\displaystyle{ 0 }[/math] to obtain:
[math]\displaystyle{ e_{n}I_{n}(k_{0}a)A_{0}\delta_{0l}+a_{ln}K_{n}(k_{l}a)A_{l} =\sum_{m=-2}^{\infty}b_{mn}I_{n}(\kappa_{m}a)B_{ml} \,\,\,(8) }[/math]
and
[math]\displaystyle{ e_{n}k_{0}I_{n}^{\prime}(k_{0}a)A_{0}\delta_{0l}+a_{ln}k_{l}K_{n}^{\prime }(k_{l}a)A_{l} =\sum_{m=-2}^{\infty}b_{mn}\kappa_{m}I_{n}^{\prime}(\kappa_{m} a)B_{ml} \,\,\,(9) }[/math]
Equation (8) can be solved for the open water coefficients [math]\displaystyle{ a_{mn} }[/math]
[math]\displaystyle{ a_{ln}=-e_{n}\frac{I_{n}(k_{0}a)}{K_{n}(k_{0}a)}\delta_{0l}+\sum _{m=-2}^{\infty}b_{mn}\frac{I_{n}(\kappa_{m}a)B_{ml}}{K_{n}(k_{l}a)A_{l}} }[/math]
which can then be substituted into equation (9) to give us
[math]\displaystyle{ \left( k_{0}I_{n}^{\prime}(k_{0}a)-k_{0}\frac{K_{n}^{\prime}(k_{0} a)}{K_{n}(k_{0}a)}I_{n}(k_{0}a)\right) e_{n}A_{0}\delta_{0l} =\sum_{m=-2}^{\infty}\left( \kappa_{m}I_{n}^{\prime}(\kappa_{m} a)-k_{l}\frac{K_{n}^{\prime}(k_{l}a)}{K_{n}(k_{l}a)}I_{n}(\kappa _{m}a)\right) B_{ml}b_{mn}\,\,\,(10) }[/math]
for each [math]\displaystyle{ n }[/math]. Together with equations (6) and (7) equation (10) gives the required equations to solve for the coefficients of the water velocity potential in the plate covered region.
Numerical Solution
To solve the system of equations (10) together with the boundary conditions (6 and 7) we set the upper limit of [math]\displaystyle{ l }[/math] to be [math]\displaystyle{ M }[/math]. We also set the angular expansion to be from [math]\displaystyle{ n=-N }[/math] to [math]\displaystyle{ N }[/math]. This gives us
[math]\displaystyle{ \phi(r,\theta,z)=\sum_{n=-N}^{N}\sum_{m=0}^{M}a_{mn}K_{n}(k_{m}r)e^{i n\theta }\phi_{m}(z), \;\;r\gt a }[/math]
and
[math]\displaystyle{ \phi(r,\theta,z)=\sum_{n=-N}^{N}\sum_{m=-2}^{M}b_{mn}I_{n}(\kappa _{m}r)e^{i n\theta}\psi_{m}(z), \;\;r\lt a }[/math]
Since [math]\displaystyle{ l }[/math] is an integer with [math]\displaystyle{ 0\leq l\leq M }[/math] this leads to a system of [math]\displaystyle{ M+1 }[/math] equations. The number of unknowns is [math]\displaystyle{ M+3 }[/math] and the two extra equations are obtained from the boundary conditions for the free plate (6) and (7). The equations to be solved for each [math]\displaystyle{ n }[/math] are
[math]\displaystyle{ \left( k_{0}I_{n}^{\prime}(k_{0}a)-k_{0}\frac{K_{n}^{\prime}(k_{0} a)}{K_{n}(k_{0}a)}I_{n}(k_{0}a)\right) e_{n}A_{0}\delta_{0l} =\sum_{m=-2}^{M}\left( \kappa_{m}I_{n}^{\prime}(\kappa_{m}a)-k_{l} \frac{K_{n}^{\prime}(k_{l}a)}{K_{n}(k_{l}a)}I_{n}(\kappa_{m}a)\right) B_{ml}b_{mn} }[/math]
[math]\displaystyle{ \sum_{m=-2}^{M}(\beta\kappa_{m}^{4}+1-\alpha\gamma)^{-1}b_{mn} \left( \kappa_{m}^{2}I_{n}(\kappa_{m}a)-\frac{1-\nu}{a}\left( \kappa _{m}I_{n}^{\prime}(\kappa_{m}a)-\frac{n^{2}}{a}I_{n}(\kappa_{m}a)\right) \right) =0 }[/math]
and
[math]\displaystyle{ \sum_{m=-2}^{M}(\beta\kappa_{m}^{4}+1-\alpha\gamma)^{-1}b_{mn} \left( \kappa_{m}^{3}I_{n}^{\prime}(\kappa_{m}a)+n^{2}\frac{1-\nu}{a^{2} }\left( \kappa_{m}I_{n}^{\prime}(\kappa_{m}a)+\frac{1}{a}I_{n}(\kappa _{m}a)\right) \right) =0 }[/math]
It should be noted that the solutions for positive and negative [math]\displaystyle{ n }[/math] are identical so that they do not both need to be calculated. There are some minor simplifications which are a consequence of this which are discussed in more detail in Zilman and Miloh 2000.
The Shallow Depth Theory of Zilman and Miloh 2000
The shallow water theory of Zilman and Miloh 2000 can be recovered by simply setting the depth shallow enough that the shallow water theory is valid and setting [math]\displaystyle{ M=0 }[/math]. If the shallow water theory is valid then the first three roots of the dispersion equation for the ice will be exactly the same roots found in the shallow water theory by solving the polynomial equation. The system of equations has four unknowns (three under the plate and one in the open water) exactly as for the theory of Zilman and Miloh 2000.
Solving for the case of a more complicated incident wave
Let's consider an incident wave whose potential has the following expression
We can apply the same eigenfunction matching that previously, considering the potential and its normal derivative continuous at [math]\displaystyle{ r=a }[/math]. Thus the potential and its normal derivative expressed at each side of this value of the radius have to be equal. We obtain the relationships
and
for each [math]\displaystyle{ n }[/math]. We solve these equations with the same method that before, by multiplying both equations by [math]\displaystyle{ \phi_{l}(z) }[/math] and integrating from [math]\displaystyle{ -H }[/math] to [math]\displaystyle{ 0 }[/math] to obtain:
and
As a consequence we can find an expression of the coefficients [math]\displaystyle{ a_{ln} }[/math]
which can then be substituted into the equation for the normal derivative of the potential, to give us
for each [math]\displaystyle{ n }[/math].
We notice here that there is still a dependence between the coefficients [math]\displaystyle{ b_{mn} }[/math] and [math]\displaystyle{ D_{ln} }[/math]. We can get rid of this dependence, using a Diffraction Transfer Matrix, which maps the coefficients of the incident wave with the coefficients of the scattered wave within the free water domain.
Matlab Code
A program to calculate the coefficients for circular dock problems can be found here circle_plate_matching_one_n.m Note that this problem solves only for a single n.
Additional code
This program requires dispersion_free_surface.m and dispersion_elastic_surface.m to run
