betaFeedback

WorksPhysics and Astronomy

Hamiltonian variational formulation of three-dimensional, rotational free-surface flows, with a moving seabed, in the Eulerian description

Abstract

Hamiltonian variational principles provided, since 60s, the means of developing very successful wave theories for nonlinear free-surface flows, under the assumption of irrotationality. This success, in conjunction with the recognition that almost all flows in the sea are not irrotational, raises the question of extending Hamilton Principle to rotational free-surface flows. The equations governing the fluid motion within the fluid domain, in the Eulerian description, have been derived by means of Hamilton Principle since late 50s. Nevertheless, a complete variational formulation of the problem, including the derivation of boundary conditions, seems to be lacking up to now. Such a formulation is given in the present work. The differential equations governing the fluid motion are derived as usually, starting from the typical Lagrangian, constrained with the conservation of mass and the conservation of fluid parcels identity. To obtain the boundary conditions, generic differential-variational constraints are introduced in the boundary variational equation, leading to a reformulation which permits us to derive both kinematic and dynamic conditions on all boundaries of the fluid, including the free surface. An interesting feature, appearing in the present variational derivation of kinematic boundary conditions, is a dual possibility of obtaining either the usual kinematic condition (the same as in irrotational flow) or a condition of different type, corresponding to zero tangential velocity on the boundary. The deeper meaning and the significance of these findings seem to deserve further analysis.

Introduction

Motivation

The significance of understanding and predicting phenomena related with the wave motion in the sea can be hardly overestimated nowadays. It suffices, perhaps, to mention that they are of central interest in Naval, Coastal, and Offshore Engineering, and also of great importance for oceanography. A distinctive feature of these nonlinear waves is the presence of the unknown free surface, which has to be predicted simultaneously with the underlying wave field. This feature greatly complicates both the mathematical analysis and the numerical computation of free-surface problems.

It turns out that these difficulties become milder and easier-to-handle under specific mathematical formulations. For example, in the case of an ideal (inviscid) incompressible liquid undergoing an irrotational wave motion, the Hamiltonian variational formulation, initiated by Petrov [1] [1] and Zakharov [2] [2], has produced effective equations for studying non-linear waves and very efficient schemes for numerical computations. The closely related unconstrained variational principle of Luke [3] [3] facilitated further development of nonlinear wave theories and numerical techniques. In the last 60 years, there is a huge amount of works developing and exploiting irrotational wave theories, those exhibiting a Hamiltonian structure being the most successful.

However successful the irrotational model may be, it is well known that rotationality is always present in the sea. And this is true, not only for phenomena like wave breaking and air-sea interaction, but also for mild realistic sea waves. Oceanographers introduced the concept of the wave-induced non-breaking turbulence in 2004 [4, 5] and found significant improvements in upper-ocean circulation predictions when they incorporated it in general circulation models [6]. The authors of the latter paper, trying to explain the scarcity of rotational wave theories, argue that “the success of potential theories of nonlinear waves … made the applications of non-potential wave theories seem redundant and eventually even led to them being nearly forgotten.”

History and background literature

The need for rotational wave theories, in conjunction with the extraordinary success of the Hamiltonian variational formulation for irrotational waves, raises the question of extending the method to the more realistic (and more complicated) case of rotational free-surface flows. In this case, however, the choice between Lagrangian and Eulerian description of the fluid motion becomes significant. In the second approach, the physical fields, say the velocity u\mathbf{u}, the density ρ\rho and the pressure pp, are expressed as functions of spatial coordinates (x,z)=(x1,x2,z)(x,z)=(x_1,x_2,z) and time tt, while in the former one the main field is the position of fluid parcels, X(a,t)\mathbf{X}(a,t), considered as a function of their initial positions a=(a1,a2,a3)\mathbf{a}=(a_1,a_2,a_3) and the time. It is thus clear that the Lagrangian description, being non convenient in applications and almost forgotten in engineering hydrodynamics, is directly amenable to methods of Analytical Mechanics, based on virtual displacements. Lagrange himself derived the hydrodynamic equations (in the Lagrangian description) by using D’Alembert Principle of virtual work in 1815 [7] (Sec. 11), but their derivation from Hamilton’s Principle should be waiting for more than a century before appearing in 1929 [8]. In the present paper, we are primarily interested in the variational formulation of rotational flows, in their Eulerian description; the Lagrangian one will be touched upon only at the extend it is necessary for developing the former.

Direct application of Hamilton’s Principle to the derivation of equations governing rotational flows in the Eulerian description, stumbles upon a fundamental controversy. The validity of Hamilton’s Principle is crucially dependent on applying virtual variations of the positions of fluid parcels, X(a,t)X(a,t), for fixed (a,t)(a,t). However, fluid parcels’ positions are completely absent in the Eulerian formulation, and the physical fields, say u=u(x,z,t)u=u(x,z,t), are defined in terms of the spatial variables. Accordingly, the natural variations of the involved fields, say δu(x,z,t)\delta u(x,z,t), are variations for fixed (x,z,t)(x,z,t). The situation described above will be referred to in the sequel as the variational controversy (of Eulerian fluid dynamics).

The first attempt to derive the Eulerian equations of fluid dynamics by means of Hamilton’s Principle was made by Herivel in 1955 [9]. He recognized the need of introducing constraints in the standard Lagrangian function (kinetic minus potential energy), and he implemented as constraints the conservation of mass and the conservation of entropy. The obtained variational equations apparently produce Euler’s equation of momentum; however, the underlying velocity representation turns out to be restricted to irrotational flows, when the entropy is assumed to be constant, as it should 11; see also the discussion in [10] (p. 5). Let it be noted that, from a purely mechanical point of view, entropy should not be of significance for the dynamics of an ideal mechanical system, as the considered fluid flow is. The variational controversy has been resolved, at least for the bulk motion of the fluid, by means of the clever proposal of Lin [11, 12]22 who introduced the purely mechanical constraint of the conservation of fluid parcels’ identity. See also Sec. 3.1. Although Lin’s constraint is commonly used in conjunction with the entropy constraint [10, 13], in fact, it can replace the latter, leaving us with a purely mechanical variational formulation.

The Herivel-Lin approach became the standard one after the publication of the works of Serrin [11] and Eckart [14]. Many papers have appeared since then, exploiting various aspects and deepening our understanding of this variational formulation. Penfield [15], Bretherton [16] and Salmon [17] discussed the importance of using Hamilton’s Principle (the Herivel-Lin approach) for studying Eulerian rotational flows of ideal fluids. The necessity of introducing Lin’s constraint has also been highlighted, by means of transformations between the two descriptions of fluid flow [17, 18]. In an alternative direction, the Eulerian variational formulation was obtained from that in the Lagrangian description, by using canonical transformations [19, 20]. The papers by Seliger and Whitham [10] and Fukagawa and Fujitani [13] dealt with many aspects of Hamilton’s Principle for rotational flows, including its relation with the Clebsch approach (see below, at the end of the Introduction). They also discussed the variational formulation with reduced versions of Lin’s constraint, that is, velocity representation with fewer potentials; see, also, the discussion in [17] (Sec. 5). However, as first illustrated in [16], and later proved with more advanced mathematical tools in [21, 22, 23], these consideration apply only to a restricted class of flows, with zero helicity and without points of vanishing vorticity; see, also [24] and [25] (Sec. 6.17).

All papers mentioned above, dealing with the Herivel-Lin Hamiltonian variational formulation of three-dimensional (3D) rotational flows, do not touch upon the issue of boundary conditions. In fact, the only work that the present authors found, with some discussion on boundary conditions in this context, is the book by Berdichevsky [26] (Sec. 9.3), where he derives a form of the free-surface dynamic condition, having all the kinematic conditions a priori imposed. In a different direction, which uses Hamilton’s Principle in conjunction with the constrained variations of the Euler-Poincaré framework, a limited number of works dealing with free-surface boundary conditions appeared recently (see [27, 28], and references therein). Although these works are interesting and illuminative regarding various aspects of the variational formulation of unsteady vortical free-surface flows, they are limited to two-dimensional (2D) flows, utilizing a stream function. Again, kinematic conditions need to be imposed as constraints in the variational formulation.

Contribution of the present paper

The goal of the present paper is to provide a complete derivation of the equations of motion and of the boundary conditions for 3D rotational flows with a free surface and a moving seabed, by means of Hamilton’s Principle. The fluid is assumed to be barotropic, which permits us to consider it as a purely mechanical system, avoiding the use of entropy and other thermodynamic concepts. To achieve this goal, we start with Herivel-Lin’s version of Hamilton‘s Principle (the standard Lagrangian, constrained by the conservation of mass and the conservation of fluid parcels’ identity), calculate the total variation of the actual functional with respect to all functional arguments (physical fields and fields of constraints), and deduce the equations of motion within the 3D fluid domain. Up to this point we follow the standard approach, as in [10], with the minor difference that we do not use entropy nor the entropic constraint. As usual, the Clebsch velocity representation, u=−∇k+∑iAi∇ai\boldsymbol{u}=-\nabla k+\sum_i A_i\nabla a_i, where kk, AiA_i, aia_i are scalar functions (Clebsch potentials), is derived variationally. It is also observed that the number of pairs AiA_i, aia_i, entering into the velocity representation, do not affect the variational treatment. A possible interpretation of this situation may be that the velocity representation belongs to the underlying kinematics and, thus, it is not fully controlled by the variational principle.

The calculation of the total variation of the action functional provides us with boundary terms as well, from which it is expected to derive the boundary conditions. This seems not to be possible, however, if we consider the boundary variations of the involved fields to be independent on the boundary, as it happens within the fluid domain. This controversy is addressed, by the present authors, to the fundamental variational controversy of the Eulerian fluid dynamics. The main argument (conjecture) is that Lin’s constraint, expressed by volume integrals, does not effectively act on the boundary, which is a lower-dimensional manifold. To cope with this difficulty, use is made of differential-variational constraints, which express the local Eulerian field variations by means of the virtual displacement δX\delta X of the fluid parcels. The latter are pointwise valid, thus valid on the boundary as well. After substituting the variations of the Eulerian fields on the boundary with the corresponding expressions in terms of δX\delta X, the number of independent boundary terms decreases and their new forms, in conjunction with standard variational arguments, provide us with the expected boundary conditions on the free surface, on the moving seabed, and on any lateral rigid-wall boundary. These results can be considered as an a posteriori justification of our conjecture, on the inadequacy of Lin’s volume-integral constraint for the boundary conditions. An interesting and rather unexpected feature, appearing in the variational derivation of kinematic boundary conditions, is the possibility of a (non-exclusive) duality. That is, the variational equation supports two types of kinematic conditions: either the usual kinematic condition (the same as in irrotational flow) or a condition of different type, corresponding to zero tangential velocity on the boundary. The deeper meaning and the significance of these findings seem to deserve further analysis.

The paper is organized as follows: In Sec. 2, we describe the geometry of the fluid domain, and present the usual differential formulation (governing equations and boundary conditions) of the problem. The Hamiltonian action functional, along with appropriate constraints, are presented in Sec. 3. Both the usual integral constraints and additional, differential-variational constraints are discussed in this section. The general variational equation is derived in Sec. 4, and it is exploited for obtaining the equations of motion within the fluid domain, including explicit representations of the velocity and pressure fields in terms of Clebsch potentials. Sec. 5 is devoted to the treatment of the boundary variational equation. Its primitive form, as obtained by the standard approach of the Calculus of Variations, is reformulated by using the differential-variational constraints, resulting in a new form from which the complete set of boundary conditions is derived, for each part of the boundary (free surface, moving seabed, fixed lateral rigid-wall boundaries). Last, in Sec. 6 we discuss the findings of this work and arrive at some concluding remarks.

Before concluding this introductory section, we believe it is appropriate to discuss some

In fact, the history of the variational formulation for rotational flows in the Eulerian description started in the 19{sup}th century, with the pioneering work of Clebsch in 1859 {citep}[29]. In this work, the author applies the theory of Pfaffian forms to obtain a representation of the velocity field in terms of three scalar functions (called Clebsch potentials), and observes that the hydrodynamic equations (expressed in terms of these potentials) may be obtained as the extremal condition of an integral functional. This functional is essentially the space-time integral of the pressure field. The main points of the Clebsch approach were summarized by Bateman (1929, 1944) {citep}[30, 31] in a form adapted to compressible flows. Seliger and Whitham {citep}[10] and Fukagawa and Fujitani {citep}[13] have discussed the relation of the Clebsch approach with Hamilton’s Principle, without considering boundary conditions. Luke {citep}[3] provided a short note suggesting the possibility of exploiting the Clebsch-Bateman variational principle for barotropic free-surface flows, and this suggestion was recently elaborated further by Timokha {citep}[32]`. In this approach, the dynamic free-surface condition is easily obtained, because the Lagrangian density is just the pressure field. We do not follow this line of thought in the present paper, choosing instead the Hamilton’s Principle as a basis of our search for the variational formulation, which provides direct connection with the foundational ground of Analytical Mechanics (canonical transformations, Noether’s theorem, etc.).

Another direction which is not considered herein (3) is 2D free-surface flows with vorticity, although there has been a lot of progress in this direction in the last two decades. This progress has mainly been based on the existence of a scalar stream function, which permits a drastic simplification of the formulation, not available in three dimensions.

Differential formulation of the problem

Generalities. Description of the fluid domain

In this work, our attention is focused on an inviscid, compressible (barotropic) fluid, undergoing a rotational flow in a horizontally unbounded domain. The fluid domain is limited by a free surface (upper boundary), an impermeable moving bottom (seabed, lower boundary), and possibly-vertical lateral boundaries, restricting the fluid domain horizontally in some directions (horizontal sectors); see Figure 1. For simplicity, the lateral boundaries, if existing, are assumed to be rigid walls. To give an exact mathematical formulation of this fluid-dynamics problem, an orthogonal Cartesian system Ox1x2zOx_1x_2z is introduced, with x=(x1,x2)\boldsymbol{x}=(x_1,x_2) being the horizontal spatial variables, and zz being the vertical variable, pointing upwards, i.e. in the opposite direction with respect to the constant gravity acceleration g\mathbf{g}. The level z=0z=0 is taken to coincide with the quiescent free surface. The moving seabed is located at z=−h(x,t)z=-h(\mathbf{x},t), where h(x,t)h(\mathbf{x},t) is a known depth function of the horizontal variable x\mathbf{x} and the time tt, whereas the free surface is described by the equation z=η(x,t)z=\eta(\mathbf{x},t), where η(x,t)\eta(\mathbf{x},t) is an (unknown) surface-elevation function. Accordingly, the fluid domain is shaped as (see Figure 1)

Image not extracted

Figure 1. Geometrical configuration of the fluid domain V(t)V(t).

V(t)={(x,z)∈R3:x=(x1,x2)∈D⊆R2, z∈[−h(x,t),η(x,t)]},(1)V(t)=\{(\mathbf{x},z)\in\mathbb{R}^{3}:\mathbf{x}=(x_{1},x_{2})\in D\subseteq\mathbb{R}^{2},\ z\in[-h(\mathbf{x},t),\eta(\mathbf{x},t)]\}, \tag*{(1)}

where the horizontal domain DD is the projection of the free surface on the plane (x1,x2)(x_{1},x_{2}). DD is assumed to be unbounded and simply connected.

The vectors

Nη=∇(z−η(x,t))=(−∂η∂x1,−∂η∂x2,1)(2)\mathbf{N}_{\eta}=\nabla(z-\eta(\mathbf{x},t))=\left(-\frac{\partial\eta}{\partial x_{1}},-\frac{\partial\eta}{\partial x_{2}},1\right) \tag*{(2)}

and

Nh=−∇(z+h(x,t))=(−∂h∂x1,−∂h∂x2,−1)(3)\mathbf{N}_{h}=-\nabla(z+h(\mathbf{x},t))=\left(-\frac{\partial h}{\partial x_{1}},-\frac{\partial h}{\partial x_{2}},-1\right) \tag*{(3)}

are normal (perpendicular), but non-unitary, on the free surface and the seabed, respectively, pointing outwardly with respect to the fluid. They commonly appear in our subsequent calculations, and this is why they are given special names.

For convenience, the free surface and the seabed are also denoted by ∂Vf\partial V_f and ∂Vsb\partial V_{sb}, respectively. The lateral boundary of the fluid domain V(t)V(t) is assumed to be vertical, and, as a whole, is denoted by ∂Vlat\partial V_{\mathrm{lat}}. It consists of two types of boundaries: rigid-wall boundaries, denoted by ∂Vw\partial V_w, and infinite “boundaries”, meaning that the flow extends to infinity in the corresponding horizontal directions, denoted by ∂V∞\partial V_{\infty}. The outward unit normal vector on the lateral boundary is denoted by nlat\mathbf{n}_{\mathrm{lat}}, or nw\mathbf{n}_w when the considerations are restricted on the rigid-wall part.

Of the two descriptions normally considered in hydrodynamics, the Eulerian one is of primary interest to us, although the Lagrangian description will also be touched upon, since it is intimately related with Hamilton’s Variational Principle. Although our main goal herein is to provide a variational formulation of rotational free-surface flows, in this section we shall briefly present the classical differential formulation of the problem. There are good reasons for doing this: i) Rederiving standard differential equations and boundary conditions from the variational principle is a desirable justification of both approaches, ii) Some of these equations, especially those expressing the kinematics of the flow, will be used as constraints in the variational principle, iii) An interesting, unexpected feature appears in the variational derivation of kinematic boundary conditions. Two types of them are produced variationally: the standard ones, as given in the present section, and an alternative type, consisting of zero tangential velocity. This phenomenon is further commented in Remark 6, at the end of § 6.

Equations governing the bulk flow

The involved fields in the Eulerian description are: the fluid velocity u=u(x,z,t)\boldsymbol{u}=\boldsymbol{u}(x,z,t), the fluid density ρ=ρ(x,z,t)\rho=\rho(x,z,t), and the fluid pressure p=p(x,z,t)p=p(x,z,t), which, along with the free-surface elevation η=η(x,t)\eta=\eta(x,t), constitute the unknowns of the problem. The differential equations, governing the fluid flow in V(t)V(t), are the following (see e.g. [33, 34, 35]):

∂ρ∂t+∇⋅(ρu)=0,(4)\frac{\partial\rho}{\partial t}+\nabla\cdot(\rho\boldsymbol{u})=0, \tag*{(4)}
∂u∂t+(u⋅∇)u+∇pρ=−g,(5)\frac{\partial\boldsymbol{u}}{\partial t}+(\boldsymbol{u}\cdot\nabla)\boldsymbol{u}+\frac{\nabla p}{\rho}=-\boldsymbol{g}, \tag*{(5)}

which expresses the momentum law, and the constitutive equation (barotropic fluid)

p=p(ρ)=ρ2∂E(ρ)∂ρ,(6)p=p(\rho)=\rho^2\frac{\partial E(\rho)}{\partial\rho}, \tag*{(6)}

where E=E(ρ)E=E(\rho) is the internal energy per unit mass of the fluid (defined by phenomenological considerations). Note that, in Eqs. (3) and (4), ∇=(∂ ⋅∂x1,∂ ⋅∂x2,∂ ⋅∂z)\nabla=\left(\frac{\partial\,\cdot}{\partial x_1},\frac{\partial\,\cdot}{\partial x_2},\frac{\partial\,\cdot}{\partial z}\right) is the 3D gradient operator.

Remark 1: In many relevant works, the internal energy is assumed to depend on the density, ρ\rho, and the specific entropy, ss, of the fluid, i.e. E=E(ρ,s)E=E(\rho,s), in accordance with general thermodynamic considerations. We do not follow this tradition for various reasons. First, the related necessary assumption of isentropic flow cancels any essential contribution from the introduction of the entropy in the variational analysis, since it produces only irrotational flows; see e.g. [10] (p. 5). Second, the option of studying non-isentropic flows (which induce heat transfer) based on classical Hamilton’s Principle is controversial [9] (p. 345). Third, the introduction of Lin’s constraint, expressing the conservation of parcels’ identity (see § 4.1, § 4.2), resolves the problem of variationally obtaining the equations of rotational flow, initially supposed to be resolved by using the specific entropy as an independent field.

Boundary conditions

As concerns the boundary conditions, we have to distinguish the various types of boundaries. On the free surface, the following two conditions apply

∂η∂t+u1∂η∂x1+u2∂η∂x2−u3=0,on z=η(x,t),(7)\frac{\partial\eta}{\partial t}+u_1\frac{\partial\eta}{\partial x_1}+u_2\frac{\partial\eta}{\partial x_2}-u_3=0,\qquad\text{on }z=\eta(\boldsymbol{x},t), \tag*{(7)}
p=p‾=given,on z=η(x,t),(8)p=\overline{p}=\text{given},\qquad\text{on }z=\eta(\boldsymbol{x},t), \tag*{(8)}

where p‾=p‾(x,t)\overline{p}=\overline{p}(\boldsymbol{x},t) is a known applied-pressure field. The first of them is the kinematic condition, ensuring that the free surface moves in accordance with the motion of the fluid parcels lying on it (that is, it is a material surface), while the second one is of dynamic nature, and it is called the dynamic free-surface condition.

Remark 2: In the case of barotropic, irrotational flow, Bernoulli’s equation (5) permits us to write the dynamic condition (8) as an evolution equation with respect to the velocity potential, greatly facilitating the variational formulation, the mathematical study, and the numerical treatment of the problem [3, 36, 37, 38, 39, 40]. Such an equation is not a priori available for the problem under consideration. However, an extended version of Bernoulli’s equation, appropriate for the examined problem (rotational flows), can be derived from the variational formulation presented herein (6); see Sec. 4.4.

On the moving seabed and on the lateral rigid-wall boundaries, we have the kinematic conditions

∂h∂t+u1∂h∂x1+u2∂h∂x2+u3=0,on z=−h(x,t),(9)\frac{\partial h}{\partial t}+u_1\frac{\partial h}{\partial x_1}+u_2\frac{\partial h}{\partial x_2}+u_3=0,\qquad\text{on }z=-h(\boldsymbol{x},t), \tag*{(9)}

and

u⋅nw=0,on ∂Vw,(10)\boldsymbol{u}\cdot\boldsymbol{n}_{w}=0,\qquad\text{on }\partial V_w, \tag*{(10)}

ensuring zero mass flow through the impermeable boundaries. Recall that, in Eq. (10), nw\boldsymbol{n}_w is the normal unit vector on the rigid-wall boundary ∂Vw\partial V_w. Since the motion of these boundaries is predetermined, a second (dynamical) condition is not needed.

Finally, as concerns the lateral “infinite” boundary ∂V∞\partial V_\infty, that is, the flow at infinity in the unbounded horizontal directions, we assume that the velocity field and the free-surface elevation tend to zero with rates ensuring that the energy integral is finite. In the opposite case, that is, when a wave system of infinite energy has been developed and moves towards infinity, the variational method is not directly applicable. Nevertheless, such cases can be included in the present formulation by using domain decomposition techniques. Such kinds of techniques have been well developed in the context of irrotational free-surface flows; see e.g. [41, 42, 43] and [44] (Ch. 7).

pρ+∂Φ∂t+12u2+E+Ω=C(t)\frac{p}{\rho}+\frac{\partial\Phi}{\partial t}+\frac{1}{2}\boldsymbol{u}^{2}+E+\Omega=C(t)

[33] (Article 20), where Φ\Phi is the velocity potential (u=∇Φ\boldsymbol{u}=\nabla\Phi), Ω=Ω(x,z)\Omega=\Omega(\boldsymbol{x},z) is the potential of an external conservative force-field, and C(t)C(t) is an arbitrary function of time.

A brief discussion of the Lagrangian approach

Analytical mechanics is based on the consideration of material particles’ positions xi(t)\boldsymbol{x}_i(t) and their virtual variations δxi(t)\delta\boldsymbol{x}_i(t), subjected to all relevant constraints. For a continuum, especially a fluid, the index identifying the positions of material parcels is a vector field, leading to an expression of the form xa(t)=x(a,t)\boldsymbol{x}_a(t)=\boldsymbol{x}(\boldsymbol{a},t). The common choice of index (label) a\boldsymbol{a} is the initial position of the fluid parcels, which means that x(a,t0)=a\boldsymbol{x}(\boldsymbol{a},t_0)=\boldsymbol{a}. Clearly, in this consideration, a\boldsymbol{a} is an independent variable extending over the geometric domain V(t0)V(t_0), not varying with time. For each t≥t0t\geq t_0, the transformation

V(t0)∋a⟶x(a,t)∈V(t)(11)V(t_0)\ni\boldsymbol{a}\longrightarrow\boldsymbol{x}(\boldsymbol{a},t)\in V(t) \tag*{(11)}

is assumed regular, smooth and invertible (diffeomorphism). Thus, the inverse transformation,

References

  1. [1]A. A. Petrov, “Variational statement of the problem of liquid motion in a container of finite dimensions,” J. Appl. Math. Mech., vol. 28, no. 4, pp. 917–922, 1964, doi: https://doi.org/10.1016/0021-8928(64)90077-2.
  2. [2]V. E. Zakharov, “Stability of periodic waves of finite amplitude on the surface of a deep fluid,” J. Appl. Mech. Tech. Phys., vol. 9, no. 2, pp. 190–194, 1968, doi: 10.1007/BF00913182.DOI
  3. [3]J. C. Luke, “A variational principle for a fluid with a free surface,” J. Fluid Mech., vol. 27, no. 2, pp. 395–397, 1967, doi: 10.1017/S0022112067000412.DOI
  4. [4]F. Qiao, Y. Yuan, Y. Yang, Q. Zheng, C. Xia, and J. Ma, “Wave-induced mixing in the upper ocean: Distribution and application to a global ocean circulation model,” Geophys. Res. Lett., vol. 31, no. 11, 2004.DOI
  5. [5]A. V. Babanin, “On a wave-induced turbulence and a wave-mixed upper ocean layer,” Geophys. Res. Lett., vol. 33, no. 20, 2006.DOI
  6. [6]A. V. Babanin and D. Chalikov, “Numerical investigation of turbulence generation in non-breaking potential waves,” J. Geophys. Res. Ocean., vol. 117, no. C11, 2012.
  7. [7]J. L. Lagrange, Mécanique analytique, no. τ. 2. Paris: Ve Courcier, 1815.
  8. [8]L. Lichtenstein, Grundlagen der Hydromechanik, 1st ed. Springer Berlin, Heidelberg, 1929.
  9. [9]J. W. Herivel, “The derivation of the equations of motion of an ideal fluid by Hamilton’s principle,” in Mathematical Proceedings of the Cambridge Philosophical Society, 1955, vol. 51, no. 2, pp. 344–349.DOI
  10. [10]R. L. Seliger and G. B. Whitham, “Variational principles in continuum mechanics,” Proc. R. Soc. London. Ser. A. Math. Phys. Sci., vol. 305, no. 1480, pp. 1–25, 1968.
  11. [11]J. Serrin, “Mathematical principles of classical fluid mechanics,” in Fluid Dynamics I/Strömungsmechanik I, Springer, 1959, pp. 125–263.
  12. [12]C. C. Lin, “Liquid Helium, Proceedings of the Enrico Fermi International School of Physics, Course XXI,” 1963.
  13. [13]H. Fukagawa and Y. Fujitani, “Clebsch potentials in the variational principle for a perfect fluid,” Prog. Theor. Phys., vol. 124, no. 3, pp. 517–531, 2010.DOI
  14. [14]C. Eckart, “Variation principles of hydrodynamics,” Phys. Fluids, vol. 3, no. 3, pp. 421–427, 1960.
  15. [15]P. Penfield Jr, “Hamilton’s principle for fluids,” Phys. Fluids, vol. 9, no. 6, pp. 1184–1194, 1966.
  16. [16]F. P. Bretherton, “A note on Hamilton’s principle for perfect fluids,” J. Fluid Mech., vol. 44, no. 1, pp. 19–31, 1970.DOI
  17. [17]R. Salmon, “Hamiltonian fluid mechanics,” Annu. Rev. Fluid Mech., vol. 20, no. 1, pp. 225–256, 1988.DOI
  18. [18]F. Bampi and A. Morro, “The connection between variational principles in Eulerian and Lagrangian descriptions,” J. Math. Phys., vol. 25, no. 8, pp. 2418–2421, 1984.
  19. [19]W. Van Saarloos, “A canonical transformation relating the Lagrangian and Eulerian description of ideal hydrodynamics,” Phys. A Stat. Mech. its Appl., vol. 108, no. 2–3, pp. 557–566, 1981.
  20. [20]A. Griffa, “Canonical transformations and variational principles for fluid dynamics,” Phys. A Stat. Mech. its Appl., vol. 127, no. 1–2, pp. 265–281, 1984.DOI
  21. [21]C. R. Graham and F. S. Henyey, “Clebsch representation near points where the vorticity vanishes,” Phys. Fluids, vol. 12, no. 4, pp. 744–746, 2000.DOI
  22. [22]Z. Yoshida, “Clebsch parameterization: basic properties and remarks on its applications,” J. Math. Phys., vol. 50, no. 11, p. 113101, 2009.DOI
  23. [23]M. Scholle, F. Marner, and P. H. Gaskell, “Potential fields in fluid mechanics: A review of two classical approaches and related recent advances,” Water, vol. 12, no. 5, p. 1241, 2020.
  24. [24]K. Ohkitani, “Study of the 3D Euler equations using Clebsch potentials: dual mechanisms for geometric depletion,” Nonlinearity, vol. 31, no. 2, p. R25, 2018.
  25. [25]A. Feldmeier, Theoretical Fluid Dynamics. Springer International Publishing, 2020.
  26. [26]V. Berdichevsky, Variational Principles of Continuum Mechanics: I. Fundamentals. Springer Berlin Heidelberg, 2009.
  27. [27]H. A. Ardakani, T. J. Bridges, F. Gay-Balmaz, Y. . H. Huang, and C. Tronci, “A variational principle for fluid sloshing with vorticity, dynamically coupled to vessel motion,” Proc. R. Soc. A, vol. 475, no. 2224, p. 20180642, 2019.DOI
  28. [28]T. J. Bridges, “The Pressure Boundary Condition and the Pressure as Lagrangian for Water Waves,” Water Waves, vol. 1, no. 1, pp. 131–143, 2019.DOI
  29. [29]A. Clebsch, “Ueber die Integration der hydrodynamischen Gleichungen.,” J. für die reine und angew. Math., vol. 56, pp. 1–10, 1859.
  30. [30]H. Bateman, “Notes on a differential equation which occurs in the two-dimensional motion of a compressible fluid and the associated variational problems,” Proc. R. Soc. London. Ser. A, Contain. Pap. a Math. Phys. Character, vol. 125, no. 799, pp. 598–618, 1929.
  31. [31]H. Bateman, “The transformation of partial differential equations,” Q. Appl. Math., vol. 1, no. 4, pp. 281–296, 1944.
  32. [32]A. N. Timokha, “The Bateman–Luke variational formalism for sloshing of an ideal incompressible liquid with rotational flows,” Збірник Праць Інституту математики НАН України, vol. 12, no. 5, pp. 267–274, 2015.
  33. [33]H. Lamb, Hydrodynamics, 6th ed. C.U.P., 1932.
  34. [34]J. Z. Wu, H. Ma, and M. D. Zhou, Vorticity and Vortex Dynamics. Springer Berlin Heidelberg, 2006.
  35. [35]J. Spurk and N. Aksel, Fluid Mechanics. Springer Berlin Heidelberg, 2008.
  36. [36]W. Craig and D. P. Nicholls, “Traveling gravity water waves in two and three dimensions,” Eur. J. Mech., vol. 21, no. 6, pp. 615–641, 2002.DOI
  37. [37]M. J. Ablowitz, A. S. Fokas, and Z. H. Musslimani, “On a new non-local formulation of water waves,” J. Fluid Mech., vol. 562, pp. 313–343, 2006.
  38. [38]A. C. L. Ashton and A. S. Fokas, “A non-local formulation of rotational water waves,” J. Fluid Mech., vol. 689, pp. 129–148, 2011, doi: 10.1017/jfm.2011.404.DOI
  39. [39]C. E. Papoutsellis, A. G. Chalaramopoulos, and G. A. Athanassoulis, “Implementation of a fully nonlinear Hamiltonian Coupled-Mode Theory, and application to solitary wave problems over bathymetry,” Eur. J. Mech. - B/Fluids, vol. 72, pp. 199–224, 2018, doi: 10.1016/j.euromechflu.2018.04.015.DOI
  40. [40]A. I. Dyachenko, P. M. Lushnikov, and V. E. Zakharov, “Non-canonical Hamiltonian structure and Poisson bracket for two-dimensional hydrodynamics with free surface,” J. Fluid Mech., vol. 869, pp. 526–552, 2019.
  41. [41]Y. Cao, R. F. Beck, and W. W. Schultz, “An absorbing beach for numerical simulations of nonlinear waves in a wave tank,” in Proc. 8th Intl. Workshop Water Waves and Floating Bodies, 1993, pp. 17–20.
  42. [42]D. Fructus, D. Clamond, J. Grue, and Ø. Kristiansen, “An efficient model for three-dimensional surface wave simulations: Part I: Free space problems,” J. Comput. Phys., vol. 205, no. 2, pp. 665–685, 2005.
  43. [43]Y. Zhang, A. B. Kennedy, N. Panda, C. Dawson, and J. J. Westerink, “Generating–absorbing sponge layers for phase-resolving wave models,” Coast. Eng., vol. 84, pp. 1–9, 2014.
  44. [44]C. E. Papoutsellis, “Nonlinear water waves over varying bathymetry: Theoretical and numerical study using variational methods”, PhD Thesis, National Technical University of Athens, Athens, 2016.
  45. [45]A. Bennett, Lagrangian Fluid Dynamics. Cambridge University Press, 2006.
  46. [46]L. A. Pars, A Treatise on Analytical Dynamics. London: Heinemann Educational Books, Ltd., 1965.
  47. [47]F. R. Gantmacher, Lectures in Analytical Mechanics. Beekman Books, Incorporated, 1975.
  48. [48]S. Mottaghi, R. Gabbai, and H. Benaroya, An Analytical Mechanics Framework for Flow-Oscillator Modeling of Vortex-Induced Bluff-Body Oscillations. Springer International Publishing, 2019.
  49. [49]L. Mittag, M. J. Stephen, and W. Youngrau, “Variational Principles in Hydrodynamics,” in Variational Principles in Dynamics and Quantum Theory, 3rd ed., Philadelphia: Saunders, 1968, pp. 142–161.
  50. [50]C. Eckart, “The electrodynamics of material media,” Phys. Rev., vol. 54, no. 11, p. 920, 1938.
  51. [51]I. M. Gelfand and S. V. Fomin, Calculus of variations. Prentice-Hall, 1963.
  52. [52]D. Henry, J. Hale, and A. L. Pereira, Perturbation of the Boundary in Boundary-Value Problems of Partial Differential Equations. Cambridge University Press, 2005.
  53. [53]G. Badin and F. Crisciani, Variational Formulation of Fluid and Geophysical Fluid Dynamics: Mechanics, Symmetries and Conservation Laws. Springer, 2018.
  54. [54]T. Kambe, Geometrical Theory Of Dynamical Systems And Fluid Flows (Revised Edition). World Scientific Publishing Company, 2009.
  55. [55]G. Grimberg and E. Tassi, “Comment on Clebsch’s 1857 and 1859 papers on using Hamiltonian methods in hydrodynamics,” Eur. Phys. J. H, vol. 46, no. 1, pp. 1–31, 2021.

Paper details

vlnc:W_bmdt985u
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 0

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