betaFeedback

WorksPhysics and Astronomy

Water Waves: Nonlinear Theory

Abstract

A novel mathematical nonlinear theory of surface gravity waves in deep water is presented, in which analytical analysis of the classical nonlinear equations of fluid dynamics is performed under less restrictive assumptions than those applied by existing theories. In particular, the new theory ensures uniqueness of the solution without the need to employ the so-called radiation condition, and its solutions are such that the liquid always remains at rest at infinity, and the energy supplied to the water by a source of disturbances is finite at all times - all that in contrast with conventional approaches that operate in terms of spatially-infinite harmonic waves. The new theory accounts for the non-linearity of the problem, and yields solutions valid at all times, from zero (the time of setting initial conditions) to infinity. The author describes previously unknown patterns in wave evolution, and confirms those patterns with the experimental results by other researchers.


– Monograph –

Mindlin, Ilia Michailovich. Water Waves: Nonlinear Theory, 2022.

This book presents a novel mathematical nonlinear theory of surface gravity waves in deep water. It performs analytical analysis of the classical nonlinear equations of fluid dynamics under less restrictive assumptions than those applied by existing theories. In particular, the new theory ensures uniqueness of the solution without the need to employ the so-called radiation condition, and its solutions are such that the liquid always remains at rest at infinity, and the energy supplied to the water by a source of disturbances is finite at all times - all that in contrast with conventional approaches that operate in terms of spatially-infinite harmonic waves. The new theory accounts for the non-linearity of the problem, and yields solutions valid at all times, from zero (the time of setting initial conditions) to infinity. The author describes previously unknown patterns in wave evolution, and confirms those patterns with the experimental results by other researchers.

The book is intended for graduate students and researchers specializing in hydrodynamics, from wind-generated waves to tsunamis.

Introduction

This monograph presents a mathematical nonlinear theory of surface gravity waves in deep water, which continues the monograph “Integro-differential Equations in Dynamics of a Heavy Layered Liquid” by the author published in Russia in 1996.

A typical flow geometry considered in this book is that of a heavy liquid occupying a half-space below a lighter liquid separated by the liquid-liquid interface. The two liquids are supposed to be non-viscous, immiscible, and homogeneous with constant densities. The system, initially at rest with a horizontal surface, is set moving either by an initial displacement of its surface, or by initial disturbance to the particle velocities due to, say, a solid body moving below the interface or by variable pressure acting on the free surface of the liquid (the concept of “free surface” is used when upper of the two liquids separated by the interface is a fictitious liquid of zero density). The ensuing motion is irrotational. The position of the interface is not known a priori and is to be found while solving the differential or integro-differential (depending on the formulation of the problem) equations of fluid dynamics. In the book, all problems are formulated in terms of integro-differential equations supplemented by physically reasonable boundary and initial conditions.

The theory presented in this monograph is referred to as nonlinear theory of wave packets (NTWP), to distinguish it from linear wave theory (LWT) and weakly nonlinear wave theory (WNWT) commonly applied to the same class of problems. The three theories stem from the classical equations of fluid dynamics, which are:

  • the Laplace equation for the velocity potential;

  • the nonlinear kinematic condition on the fluid surface which means that a liquid particle in the fluid surface can have no velocity relative to the surface in the direction of the normal;

  • the nonlinear condition of the pressure continuity across the fluid surface;

  • conditions at infinity; and

  • initial conditions.

When the problem includes a solid body moving inside the liquid, the above equation list also includes the kinematic condition on the surface of the solid, as well as the dynamic equation(s) for the solid. The nonlinear conditions on the interface present a barrier for finding the solution to the problem analytically, unless some simplifying assumptions are applied.

The three theories are quite different in technique of simplifying the classical equations and mathematical tools employed for finding solutions. The resulting solutions behave differently in time, and are being applied to different kinds of specific problems. The differences between the Packet Theory and LWT (we concentrate primarily on LWT here, since, as a matter of fact, mathematical formulation of WNWT deviates quite far from the original governing equations of the problem) follow:

  1. The cardinal difference is how conditions at infinity along the interface are prescribed in the theories.

    In the theory of wave packets, the solution of the equations is obtained under boundary conditions which assure that the liquid remains at rest at infinity, and the energy supplied to the water by a source of disturbances is finite at any time.

    In the linear theory, the fluid velocity is required to be bounded at infinity along the interface. Under the condition of boundedness, the problem has a set of steady state solutions in the form of harmonic waves periodic in time and along the free surface. To produce waves periodic along the free surface infinite in extent, infinite energy must be supplied to the liquid. The physics here is somewhat mildly violated for the convenience of expressing the solution with the Fourier integral.

  1. The boundary condition on the free surface imposed in NTWP ensure the uniqueness of the solution of initial-value problem.

    In the LWT, solution of an initial-value problem obtained as a Fourier integral is a linear combination of waves traveling to ±∞\pm\infty in opposite directions. That is why in the LWT, radiation condition is prescribed which means that only waves outgoing from disturbed body of water represent physically reasonable solutions. The waves coming from infinity are ignored.

    The inconsistency of the LWT which represents spatially and temporarily finite motion with a set of infinite harmonics had been noted so long ago that it might had been forgotten by now. Citing Stoker, “the difficulty arises because the problem of determining simple harmonic motions is an unnatural problem in mechanics... The steady state problem is unnatural - in the author’s view, at least - because a hypothesis is made about the motion that holds for all time, while Newtonian mechanics is basically concerned with the prediction - in a unique way, furthermore - of the motion of a mechanical system from given initial condition...

    One should in principle rather formulate and solve an initial value problem by assuming the medium to be originally at rest everywhere outside a sufficiently large sphere, ...” (Stoker, Water Waves, p.176)

  1. In the theory of wave packets, it is assumed that the waves start to propagate away from an initially disturbed body of liquid. It is also assumed that the characteristic horizontal dimension DD of the body is much larger than the vertical displacement dd of the fluid surface in the wave originating area. Equations of the linear theory are obtained from classical equations assuming that amplitude aa of a wave is small compared to the wave length λ\lambda. Small parameters d/D≪1d/D \ll1 and a/λ≪1a/\lambda\ll1, respectively, are used in the theories to simplify the classical equations.

  1. In NTWP, equations are obtained with nonlinear boundary conditions on the evolving interface unknown in advance. The shape of the interface is thought in a parametric form which is equivalent to a nonlinear transformation of the governing equations.

    In LWT, the conditions on the interface are linearized and shifted from the interface to the horizontal plane - the equilibrium position of the interface. Thus the vertical departure of the free surface becomes an explicit function of the horizontal coordinate (and time).

  1. In NTWP, the nonlinear transformation maps the trace of the liquid-liquid interface in a vertical plane onto a half of a circle. The mapping allows us to find a countable (with respect to spacial argument) set of solutions (specific wave packets). Any wave group is a nonlinear combination of the wave packets.

The LWT solution is assembled from a continuum of harmonic waves.

  1. Nonlinear Theory of Wave Packets (NTWP) accounts for the non-linearity and provides the solution valid at all times, from zero (the time of setting initial conditions) to infinity.

The theory of weakly nonlinear waves (WNWT) in deep water is based on the concept that the frequency, amplitude, and wave number of a weakly nonlinear wave train are slowly varying functions of time. On this assumption, the classical equations were reduced to equations governing the evolution of the frequency, amplitude, and wave number. These equations in turn were reduced to nonlinear Schredinger equation for complex-valued wave envelope. Asymptotic behavior of solutions of the Schredinger equation was analyzed by Zakharov and Shabat [13]. Provided that the initial free surface displacement decays sufficiently rapidly with the distance from its peak, it was found, in particular, that an initial wave group with an arbitrary envelope eventually disintegrates into a number of final packets. Each final packet has a specific envelope which travels at a constant velocity, keeps its shape, and retains its shape and velocity after interaction with other final packets. Disintegration of a wave group into final packets traveling at different velocities is also the dispersion. But the final packets don’t disperse.

WNWT is not bridged to the LWT, but rather the two theories represent two unconnected patterns of wave behavior. Moreover, WNWT predictions contradict the observed behavior of water waves: even long waves in mid-ocean (wave length of order of 100-150 km for the ocean depth of 4-5 km) are dispersive, as shown by their records.

Book content overview

The book presents a novel mathematical theory of nonlinear surface gravity waves in deep water. It starts with elaborating the mathematical tools in Chapters 1 and 2.

In Chapters 3 - 5, equations of nonlinear initial-boundary value problem on deep-water gravity waves of finite amplitude are solved approximately (up to small terms of higher order) assuming that the waves on the interface are generated by an initial disturbance to the liquid and the horizontal dimensions of the initially disturbed body of the liquid are much larger than the magnitude of the interface displacement.

A set of dispersive wave packets is found with one-to-one correspondence between the packets and positive integers, say, packet numbers, such that any initial surface displacement gradually disintegrates into a number (limited or unlimited, depending on initial conditions) of the wave packets. The greater the packet number, the shorter the wavelength of the packet’s carrier wave component, the slower the packet travels, the slower the envelope of the packet is widening; evolution of any packet is not influenced by evolution of another packet. The speed of any packet is less than the speed of packet number 1 which, in turn, depends on typical horizontal length of the wave origin.

The analytical solution and original “law of similarity” for the wave packets presented in Chapter 5 are used to test the theory of packets against experiments performed in a water tank (waves on deep water, in terms of linear theory) by Feir (1967) and Yuen, H.C. and Lake, B.M. [12]. The agreement between theoretical predictions based on NTWP and the experimental measurements in a water tank is surprisingly good.

Chapter 6 focuses on applying Packet Theory for estimating wave origin and the traveling waves parameters in a particular scenario of wave generation. The analytical solution and the “law of similarity” for the wave packets are used to estimate an assemblage of dimensional parameters of the wave origin and waves radiated from the origin.

It is assumed that initially the water surface has not yet been displaced from its mean level, but the velocity field has already become different from zero. This means that the motion of a body of water is triggered by a sudden change in the velocity field. The scenario resembles a tsunami generation event. Some characteristics of the latter are in line with the assumptions of the theory of nonlinear wave packets, such as energy supplied to the water by the earthquake being finite, and the horizontal extent of the wave origin (tens to hundreds kilometers) immensely exceeding the departure of the water surface from the mean sea level (order of meters).

Considering the resulting wave as a specific wave packet and applying the “law of similarity”, we develop a procedure for using water surface measurements at a location distant from the wave originating area for estimating duration of the wave origin formation, size of the origin, water elevation in the origin, energy supplied to the water by the motion trigger, distribution of wave heights in the wave packet, speed of the waves. Relation between the speed of the wave of maximum height and its length is obtained.

In section 8 of Chapter 6, we consider a situation when wave records at one or several DART1 buoys are already known, but the records at the next buoys are not obtained yet. Starting from the first record, a line of forecasts of the amplitude and arrival time of the wave of maximum height at the next buoys is produced corresponding to the timeline of the DART records.

We exercise our parameter estimation procedure with DART records of waves triggered by the earthquakes of November 2006 and January 2007 near Kuril Islands and of august 2007 near central Peru.

Though the analytical solution to the problem and the “law of similarity” was obtained for an idealized bottom-less basin, the above estimates agree with observations of long waves in the Pacific as well as with some results of numerical simulations of the Kuril tsunamis.

In Chapter 7, the problem on gravity waves on the free surface of a liquid initially at rest is solved analytically in cases where the external pressure force of limited power is distributed over a large area in the free surface but is otherwise arbitrary (it is also supposed that the pressure acts on the free surface of the water for infinite time interval) According to the solution, the crests of the forced waves move faster than the troughs, the horizontal distance between a crest and the following trough increases, and a chain of waves with “overhanging” develops when the packets run away from the pressure zone. It is proved that at any fixed moment of time, the wave packet in two dimensions dies out like inverse square of the distance from the pressure zone. By passing to the limit in the solution as time goes to infinity, the form of the nonlinear steady-state waves is obtained. At a distance from the pressure zone, the steady-state wave seems as infinite chain of Kelvin-Helmholtz billows. The chain is unstable.

The averaged energy absorbed by the liquid per unit time (energy absorption functional) is obtained from the solution. In general, the waves transport energy toward infinity. But a discrete set of specific frequencies is found at which the energy functional is equal to zero. This means that the steady-state wave does not transport energy toward infinity. Only standing waves have this property.

The frequency spectrum and the form of the forced standing waves are found explicitly. The waves have a finite number of nodes in the free surface infinite in extent. The amplitude of the standing waves dies out like inverse square of the distance from its peak.

In Chapter 8, a problem on gravity waves excited in two layered liquid by oscillating solid cylinder is solved analytically assuming that the ratio of the cylinder radius to the distance to the interface is small. In the available literature, this problem is considered at linearized conditions on the interface, that is, in Cartesian rectangular coordinates. But in the Cartesian coordinates, one can’t linearize a surface of a cylinder, so different simplifications are made. For example, in [10], the cylinder is modeled as a dublet, and solution to the problem on free surface waves generated by a cylinder initially at rest is sought in the form of power series of time; only first three terms of the series are obtained.

On the contrary, in this book, the solution is obtained with no-flow condition on the true surface of cylinder (as well as with nonlinear conditions on the liquid-liquid interface) and is valid at all times, from zero (the time of setting initial conditions) to infinity. According to the solution, a chain of waves with “overhanging” develops when the wave gets away from the oscillating cylinder. Letting the time go to infinity, the steady state waves are found.

Mathematical Theory of a Vortex Sheet in an Ideal Liquid

This chapter introduces the analytical approach to solving the initial value problem being developed in the book. This approach is based on expressing the fully-nonlinear classical equations for flows with a free surface in a particular system of curvilinear coordinates. The resulting solution takes a form which does not involve the Fourier integral nor the so-called radiation condition, but instead expresses the motion in terms of functions vanishing at infinity along the free surface - as different from spatially-infinite sines and cosines.

Generalized velocity potential

Motion of two non-viscous incompressible liquids separated by a liquid-liquid interface is considered below.

The interface is infinite in horizontal directions. The half-space above the interface is occupied by a homogeneous liquid of density γ1\gamma_1, while a homogeneous liquid of density γ2\gamma_2 fills the half-space below the interface. The two-layered liquid is stably stratified, i.e., γ1<γ2\gamma_1 < \gamma_2. Diffusion across the interface and surface tension are neglected.

Let initially (at t=0t = 0) the flow inside each of the liquids is irrotational. When the liquids are acted upon by no external force other than gravity, the flow inside each of the liquids remains irrotational at t>0t > 0. It does not mean that the flow of the two layered liquid is irrotational: the circulation in a closed circuit intersecting the interface is nonzero, and, consequently, the velocity potential does not exist. But it is possible to construct the generalized velocity potential Φ(M)\Phi(M) such that fluid velocity q=grad⁡Φ\mathbf{q} = \operatorname{grad}\Phi inside each of the liquids, and both kinematic and dynamic conditions are satisfied on the interface. With all that, the potential Φ(M)\Phi(M) and grad⁡Φ\operatorname{grad}\Phi jump across the interface SS.

A fluid surface is referred to as a vortex sheet, if there is a jump in tangential velocity at the surface. Two nonlinear conditions are required to be satisfied on a vortex sheet:

i) kinematic condition which means that a liquid particle in the sheet can have no velocity relative to the sheet in the direction of the normal;

ii) dynamic condition of the pressure continuity across the sheet.

The generalized velocity potential constructed in this section ensures that the nonlinear conditions on a vortex sheet evolving with time are satisfied automatically.

Coordinate systems and sketch of the liquid-liquid interface

Figure 1.1. Coordinate systems and sketch of the liquid-liquid interface

From now on we consider the flow in two dimensions. We think of the liquid as being contained between two vertical planes (parallel to the (x,y)(x,y)-plane) at a distance apart. Let the curve Γ\Gamma (in figure 1) be the trace of the interface in the plane (x,y)(x,y), the vertical xx-axis be oriented upward and the yy-axis be horizontal; x=f<0x=f<0, y=0y=0 be the coordinates of the pole O1O_1 of the polar coordinate system in the (x,y)(x,y) plane, θ\theta be the polar angle measured from the positive xx-axis in the counterclockwise direction, tt be the time, q\mathbf{q} denote the velocity of a liquid particle relative to the "Earth-fixed" coordinate system (x,y)(x,y). The equilibrium position of the interface is horizontal plane x=0x=0.

Generalized velocity potential Φ\Phi of the flow induced by the vortex sheet is sought in the form of doublet distribution over the sheet:

Φ(Q)=12π∫Γg(Q1)∂∂n(ln⁡1r) dl,(1)\Phi(Q)=\frac{1}{2\pi}\int_{\Gamma}g(Q_1)\frac{\partial}{\partial n}\left(\ln\frac{1}{r}\right)\,dl, \tag*{(1)}

where dldl is the infinitesimal element of the curve Γ\Gamma. Let Q1Q_1 denote a point on the interface SS, QQ is a point outside SS, r=∣Q1Q∣r=\lvert Q_1Q\rvert is the distance between the points, g(Q1)g(Q_1) is the density of the doublets distributed over the surface SS. In the outside of the interface SS the Laplace’s equation is satisfied provided that the doublet density g(Q1)g(Q_1) is a continuous function. The interface SS is considered to be two-sided with unit normals n+\mathbf{n}_+ and n−=−n+\mathbf{n}_-=-\mathbf{n}_+ directed in the positive and negative sides of the interface, respectively.

For the one-sided limits we introduce the notations

Φ+(Q1)=lim⁡Q→Q1+Φ(Q),Φ−(Q1)=lim⁡Q→Q1−Φ(Q).\Phi_+(Q_1)=\lim_{Q\to Q_1^+}\Phi(Q),\qquad\Phi_-(Q_1)=\lim_{Q\to Q_1^-}\Phi(Q).

where Q→Q1+Q\to Q_1^+ (Q→Q1−Q\to Q_1^-) means that the point QQ approaches the point Q1Q_1 from the positive (negative) side of the interface. The potentials have the following properties (Courant&Hilbert, 1937)

Φ+(Q1)−Φ−(Q1)=g(Q1) in traditional notations,\Phi_+(Q_1)-\Phi_-(Q_1)=g(Q_1)\ \text{in traditional notations,}
Φ+(Q1)−Φ−(Q1)=−g(Q1)(2)\Phi_{+}(Q_1)-\Phi_{-}(Q_1)=-g(Q_1) \tag*{(2)}
2Φ(Q1)=Φ+(Q1)+Φ−(Q1),∂Φ∂n++∂Φ∂n−=0,(3)2\Phi(Q_1)=\Phi_{+}(Q_1)+\Phi_{-}(Q_1),\qquad\frac{\partial\Phi}{\partial n_{+}}+\frac{\partial\Phi}{\partial n_{-}}=0, \tag*{(3)}

where ∂Φ∂n+\frac{\partial\Phi}{\partial n_{+}} (∂Φ∂n−\frac{\partial\Phi}{\partial n_{-}}) denotes the derivative at the point Q1Q_1 in the direction of the positive (negative) normal.

The function Φ(Q)\Phi(Q) is a generalized velocity potential, if the normal component of the fluid velocity and the pressure change continuously across the interface. Equation (3) shows that the normal-velocity condition on the interface is satisfied automatically. All parameters, variables, and equations have already been made non-dimensional by the use of some reference quantities: length L∗L_*, density γ∗\gamma_*, pressure P∗P_*, and dimensional unit of time, T∗T_*, defined by the relation gT∗2=L∗gT_*^2=L_*, where gg is acceleration of free fall. The non-dimensional acceleration of free fall is equal to unity, so the force function for the weight per unit mass is −x-x: F=grad⁡(−x)\mathbf{F}=\operatorname{grad}(-x). Inside each of the two liquids Bernoulli’s equation hold (Milne-Thomson, 1950):

Pγ+12q2+x+∂Φ∂t=c(t).(4)\frac{P}{\gamma}+\frac{1}{2}q^2+x+\frac{\partial\Phi}{\partial t}=c(t). \tag*{(4)}

Vortex sheet in curvilinear coordinates

Assuming that the equation of a vortex sheet is V(x,y,t)=0V(x,y,t)=0, we define curvilinear coordinates uu, vv in the x,yx,y-plane by the relations

v=V(x,y,t),u=U(x,y,t),(5)v=V(x,y,t),\qquad u=U(x,y,t), \tag*{(5)}

where functions V(x,y,t)V(x,y,t), U(x,y,t)U(x,y,t) and their first partial derivatives are assumed to be continuous. The equation of the vortex sheet becomes v=0v=0.

The kinematic condition may be written as

[∂V∂t+∂V∂xqx+∂V∂yqy]V=0=0,(6)\left[\frac{\partial V}{\partial t}+\frac{\partial V}{\partial x}q_x+\frac{\partial V}{\partial y}q_y\right]_{V=0}=0, \tag*{(6)}

where (qx,qy)(q_x,q_y) are the components of the fluid velocity, q\mathbf{q}, in the directions of the Cartesian axes.

The component of the fluid velocity normal to the vortex sheet is

qn=1N[∂V∂xqx+∂V∂yqy]V=0=−1N∂V∂t,q_n=\frac{1}{N}\left[\frac{\partial V}{\partial x}q_x+\frac{\partial V}{\partial y}q_y\right]_{V=0}=-\frac{1}{N}\frac{\partial V}{\partial t},
N2=(∂V∂x)2+(∂V∂y)2;N^2=\left(\frac{\partial V}{\partial x}\right)^2+\left(\frac{\partial V}{\partial y}\right)^2;

the subscript V=0V=0 shows that the expressions in the square brackets are evaluated on the sheet. The normal velocity component is continuous on the sheet.

Notation G^(u,v,t)\hat{G}(u,v,t) is used for any function G(x,y,t)G(x,y,t) expressed in curvilinear coordinates uu, vv, tt, i.e. G(x,y,t)=G^[U(x,y,t),V(x,y,t),t]G(x,y,t)=\hat{G}[U(x,y,t),V(x,y,t),t]. The difference between the notations ∂G^∂t\frac{\partial\hat{G}}{\partial t} and ∂G^∂t\frac{\partial\hat{G}}{\partial t} (as well as between ∂G^∂x\frac{\partial\hat{G}}{\partial x} and ∂G^∂x\frac{\partial\hat{G}}{\partial x}) should be noted: the former means the partial derivative of G^(u,v,t)\hat{G}(u,v,t) with respect to tt, while the latter means that the partial derivative ∂G(x,y,t)∂t\frac{\partial G(x,y,t)}{\partial t} is expressed in curvilinear coordinates u,v,tu,v,t, i.e.

References

  1. [6]Pain, H.J. The Physics of Vibrations and Waves. Wiley, London (1976)
  2. [7]Rabinovich, A. B., Lobkovsky, L. I., Fine, I. V., Thomson, R. E., Ivelskaya, T. N., and Kulikov, E. A.: Near-source observations and modeling of the Kuril Islands tsunamis of 15 November 2006 and 13 January 2007, Adv. Geosci., 14, 105-116, https://doi.org/10.5194/adgeo-14-105-2008, 2008.
  3. [8]Stoker, J.J. Water Waves. Interscience Publishers, Inc., New York ,1953
  4. [9]Suetin, P.K. Classical Orthogonal Polynomials. Nauka, Moscow, 1964 (in Russian)
  5. [10]Tyvand, Peder A. and Touvia Miloh. Axisymmetric interaction between a vortex ring and a free surface. Physics of Fluids, 6, 224 (1994); doi: 10.1063/1.868070DOI
  6. [11]Whittaker, E.T. and G.N. Watson. A Course of Modern Analysis. Cambridge University Press, UK, 1st-5th editions: 1902, 1915, 1920, 1927, 2021.
  7. [12]Yuen, H.C. and Lake, B.M., Nonlinear Dynamics of Deep-Water Gravity Waves, Advances in Applied Mechanics, vol. 22, New York, London: Academic, 1982, p. 67.DOI
  8. [13]Zakharov, V.E. and Shabat, A.B. (1972) Exact Theory of Two-Dimensional Self-Focusing and One-Dimensional Self-Modulation of Waves in Non-Linear Media. Journal of Experimental and Theoretical Physics, 34, 62.

Paper details

vlnc:W_4qjjghw2
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 5

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Physics and Astronomy

Not recorded for this paper

View statement
Looking for related papers…
Contents