Big Picture

Real aerospace vehicles are never purely rigid or purely flexible. A launch vehicle bends under thrust and aerodynamic loading, a spacecraft bus carries long solar arrays, and aircraft wings undergo aeroelastic deformation while the fuselage executes rigid-body motion.

To model such systems realistically, we must describe rigid-body motion and elastic deformation together in a single variational setting. Extended Hamilton’s principle is the tool that produces, in one derivation:

• ordinary differential equations (ODEs) for discrete rigid-body coordinates,
• partial differential equations (PDEs) for distributed elastic fields,
• boundary and interface conditions at rigid–flexible junctions.

1. Hybrid Generalized Coordinates

We introduce two complementary types of generalized coordinates to represent hybrid aerospace structures.

1.1 Discrete (Lumped) Coordinates

The lumped coordinates are \( q_i(t), \; i = 1, \dots, n \).

They describe a finite set of degrees of freedom such as

• rigid-body translation and rotation of a spacecraft or launch vehicle,
• hinge and gimbal angles,
• actuator or shock-absorber strokes,
• control-surface deflections.

These coordinates naturally appear in standard finite-dimensional Lagrange equations.

1.2 Distributed (Field) Coordinates

The distributed (field) coordinates are functions of both space and time: \( w_j(x,t), \; x \in \Omega \).

They represent deformation fields, for example:

• bending deflection of a beam along its span,
• torsional twist of a panel or solar array,
• vibration of booms, antennas, and other slender structures.

Classical Lagrangian mechanics is well suited to the lumped coordinates \( \{q_i\} \), while classical field theory focuses on the fields \( \{w_j\} \). Hybrid aerospace systems contain both, so we need a variational principle that treats them in a unified way.

2. Hybrid Lagrangian: Rigid, Flexible, and Interface Terms

For a hybrid system, the total Lagrangian is written as

\[ L = L_D(q,\dot q, t) + L_B(q, w, \dot w, t) + \int_{\Omega} \mathcal{L}(q,\dot q, w, \dot w, w', \dot w', x, t)\, d\Omega. \]

2.1 Discrete (Rigid-Body) Contribution \(L_D\)

\(L_D\) collects the kinetic and potential energy of the rigid or lumped elements:

• spacecraft bus and reaction wheels,
• gimballed engines and thruster clusters,
• lumped actuators and concentrated masses.

2.2 Interface Contribution \(L_B\)

\(L_B\) encodes the energy associated with joints and attachment points where rigid and flexible subsystems meet. Typical contributions include

• root bending moment times root deflection,
• torques transmitted into flexible panels,
• hinge springs and dampers at the root of an appendage.

These terms are essential for correctly capturing the rigid–flexible interface conditions.

2.3 Distributed Lagrangian Density \( \mathcal{L} \)

The Lagrangian density \( \mathcal{L} \) is integrated over the elastic domain \( \Omega \) and typically contains:

• Distributed kinetic energy, e.g. \( \tfrac{1}{2} \rho A \dot w^{\,2} \),
• Strain energy built from spatial derivatives \( w', w'', \dots \) (bending, torsion, axial strain),
• Distributed external loads such as aerodynamic pressure, gravity-gradient forces, or thermal loads.

This decomposition keeps the bookkeeping clean:

• \( (q,\dot q) \): rigid/lumped variables,
• \( (w,\dot w) \): elastic field variables,
• interface and boundary contributions arise through \(L_B\) and the boundary terms generated from \( \mathcal{L} \).

3. Extended Hamilton’s Principle with Nonconservative Work

Let \( \delta W_{\text{nc}} \) denote the virtual work of nonconservative forces, including both lumped and distributed parts:

• actuator forces and torques,
• damping forces,
• aerodynamic work,
• boundary forces and moments.

The extended Hamilton principle states that the actual motion \( (q(t), w(x,t)) \) satisfies

\[ \delta \int_{t_0}^{t_f} \left( L - W_{\text{nc}} \right) dt = 0, \]

for all admissible variations such that

• \( \delta q_i(t_0) = \delta q_i(t_f) = 0 \),
• \( \delta w(x,t) \) respects any fixed boundary conditions on the field,
• boundary variations at the ends (e.g. \( x = 0, \ell \)) are consistent with the chosen constraints.

Because \( w(x,t) \) depends on both space \( x \) and time \( t \), the variational calculus involves integration by parts in both variables. This automatically generates:

• ODEs for the discrete coordinates,
• PDEs for the elastic fields,
• boundary and interface conditions at the edges and attachment points.

4. Equations for the Discrete Coordinates \( q_i(t) \)

Collecting the terms that multiply \( \delta q_i(t) \) and integrating by parts in time (using \( \delta q_i(t_0) = \delta q_i(t_f) = 0 \)) leads to a generalized Lagrange equation:

\[ \frac{d}{dt}\left( \frac{\partial L}{\partial \dot q_i} \right) - \frac{\partial L}{\partial q_i} = Q_i^{(T)}. \]

Here \( Q_i^{(T)} \) is the total generalized nonconservative force associated with \( q_i \):

• direct actuator forces or torques,
• damping and control forces in generalized form,
• reaction forces transmitted through the flexible structure and interface terms.

The equation has the same form as the standard finite-dimensional Lagrange equation, but all coefficients include the influence of the distributed subsystem. The rigid-body motion is therefore dynamically “aware” of the elastic deformation.

5. PDEs for the Distributed Coordinates \( w(x,t) \)

Next we collect all terms multiplying the interior variations \( \delta w(x,t) \). After integrating by parts in both time and space, the interior contribution leads to an Euler–Lagrange–type PDE:

\[ \frac{\partial}{\partial t} \left( \frac{\partial \mathcal{L}}{\partial \dot w} \right) - \frac{\partial \mathcal{L}}{\partial w} + \frac{\partial}{\partial x} \left( \frac{\partial \mathcal{L}}{\partial w'} \right) - \frac{\partial^2}{\partial x^2} \left( \frac{\partial \mathcal{L}}{\partial w''} \right) = f^{(T)}(x,t), \]

where \( f^{(T)}(x,t) \) is the net distributed nonconservative load per unit length (or area/volume) due to aerodynamic pressure, distributed damping, gravity-gradient forcing, and so on.

The structure of the PDE reflects the physics:

• time-derivative terms arise from distributed kinetic energy,
• spatial derivatives encode bending, torsion, and axial strain energies,
• the right-hand side collects applied and dissipative loads.

For a classical Euler–Bernoulli beam, choosing the standard forms for kinetic and bending energy in \( \mathcal{L} \) reduces this expression to the familiar fourth-order beam equation.

6. Boundary and Interface Conditions: Essential vs Natural

After handling the interior variations, the remaining terms in the variation involve the boundary quantities at the ends of the domain (for example \( x = 0 \) and \( x = \ell \)):

\[ \delta w(0,t),\;\delta w(\ell,t),\;\delta w'(0,t),\;\delta w'(\ell,t), \]

together with contributions from the interface Lagrangian \( L_B \). These terms must vanish independently of the interior variations and therefore generate the boundary and interface conditions.

In compact notation, the boundary contribution can be written as

\[ \left[ \left( \frac{\partial \mathcal{L}}{\partial w'} - \frac{\partial}{\partial x} \frac{\partial \mathcal{L}}{\partial w''} \right)\delta w + \left( \frac{\partial \mathcal{L}}{\partial w''} \right)\delta w' + (\text{boundary forces and moments}) \right]_{x=0}^{x=\ell} = 0. \]

6.1 Essential (Geometric) Boundary Conditions

If a displacement or slope is prescribed at a boundary, the corresponding variation is zero. Examples:

• Clamped (built-in) end at \( x = 0 \):
  \( w(0,t) = 0, \quad w'(0,t) = 0 \Rightarrow \delta w(0,t) = 0, \; \delta w'(0,t) = 0. \)
• Pinned end at \( x = 0 \):
  \( w(0,t) = 0 \) but the slope is free: \( \delta w(0,t) = 0 \) while \( \delta w'(0,t) \) is unrestricted.

6.2 Natural (Dynamic) Boundary Conditions

If a variation is not constrained at a boundary, its coefficient must vanish. This yields physical conditions such as:

• zero bending moment at a free end,
• zero shear force at a free end,
• continuity of bending moment and shear across a rigid–flexible interface,
• relations involving applied endpoint forces and moments.

A key advantage of extended Hamilton’s principle is that these conditions emerge automatically from the variational process—there is no need to guess them beforehand.

7. Energy Viewpoint and Hybrid Hamiltonian

For a conservative hybrid system (no explicit time dependence in \( L \), no nonconservative forces), the extended Hamilton framework connects directly to an energy integral.

We define the generalized momenta as

• Lumped momenta: \( p_i = \dfrac{\partial L}{\partial \dot q_i} \),
• Momentum density of the field: \( \pi(x,t) = \dfrac{\partial \mathcal{L}}{\partial \dot w(x,t)} \).

The hybrid Hamiltonian is then

\[ H = \sum_i p_i \dot q_i + \int_{\Omega} \pi(x,t)\, \dot w(x,t)\, d\Omega - L. \]

Under the usual “natural system” assumptions (kinetic energy quadratic in \( \dot q \) and \( \dot w \), and potential energy depending only on \( q \) and \( w \)), the Hamiltonian reduces to

\[ H = T + V, \]

which is conserved whenever nonconservative forces are absent. This energy functional forms the basis for:

• fully Hamiltonian formulations of hybrid systems,
• structure-preserving (symplectic) time integrators,
• modal analysis of flexible spacecraft and launch vehicles.

8. Summary Box — Extended Hamilton’s Principle for Hybrid Systems

• Introduce hybrid coordinates: lumped $q_i(t)$ and distributed $w_j(x,t)$.
• Form the hybrid Lagrangian $L = L_D + L_B + \int_{\Omega}\mathcal{L}\,d\Omega$.
• Apply extended Hamilton’s principle with nonconservative work $\delta W_{\text{nc}}$.
• Derive coupled equations of motion: ODEs for $q_i(t)$ and PDEs for $w_j(x,t)$.
• Obtain essential and natural boundary/interface conditions automatically from the variational boundary terms.
• Build the hybrid Hamiltonian $H = \sum_i p_i \dot q_i + \int \pi\,\dot w\,d\Omega - L$ for energy-based analysis and symplectic integration.

Extended Hamilton’s principle is therefore the core variational tool behind modern flexible multibody dynamics. It provides a clean, unified methodology for deriving consistent ODE–PDE systems and their boundary conditions for hybrid aerospace structures.