D'Alembert-Lagrange!s Principal Equations, Their Origin and Applications

D'Alembert-Lagrange!
s Principal Equations,
Their Origin and Applications
K. C. Park, Carlos A. Felippa

Center for Aerospace Structures
University of Colorado at Boulder
Boulder, CO 80309
Roger Ohayon
Structural Mechanics and Coupled Systems Laboratory
Conservatoire National des Arts et Metiers (CNAM)
Paris, France.
USNCCM9
San Francisco, 23-26 July 2007
Principal Figures for Todays Talk
DAlemberts Principle was Lagrange came to Paris from

reported to Acadmie des Turin in 1787 and published
Sciences in 1742. his book in 1788.
Trait de Dynamique (1758) Mchanique Analytique (1788)

Developments in the Formulation of Dynamical Systems
! It began with a single, free and rigid body

with a point mass;
! Then, a rigid body with the rotational inertia properties;
! A rigid-link with joints;
! Lumped mass-spring models
! Continuum (flexible) models
! FEM/Rigid models with joints and constraints;
! And, the race for complex models is on!

Theme of Todays Talk
As a result of the race for ever more complex modeling

and analysis, often physical insight with simplicity
in modeling is lost or abandoned on the wayside?
And, we are left in the thick flood of numbers and
gigabit analysis data.
One always returns to where one has started.

--Denise Levertov(1923-1997)
Hence, in dynamics, we may return to the simplicity and

fundamental form of dAlembert and Lagranges
Principal Equations.
What are DAlembert-Lagrange Principal Equations?
Definition:
For a N-degree of freedom system,
regardless whether it is rigid or flexible,
dAlembert-Lagranges principal equations
are obtained by summing all the forces and
all the moments(with respect to a point) in the system.
Symbolic Expressions:
Sum of forces (3 equations at most):

! (fi - mi ai ) = 0
Sum of moments (3 equations at most):

! {Mi + ri x (fi - mi ai )} = 0
Semantic Questions related to
the dAlembert-Lagrange Principal Equations:
Is a floating structure in an equilibrium condition?
Or
Is a free-free substructure, partitioned from an assembled

system, in its equilibrium state?
The answer is not in the blowing wind,

but in the mathematical expressions of the dAlembert-
Lagrange Principal Equations and their physical meaning.
A bottom-up approach to
DAlembert-Lagranges Principal Equations
***Warning: This model does not capture the boomerang motions!

Equations of Motion for 6-DOF Boomerang Model
Definition of dAlemberts Forces
Note that the inertia forces are associated with the

negative sign (-) which means: they are resisting forces
in equilibrium equations.
Sum of x-direction forces:
Note that uc changes with time in general for flexible structures.

Sum of y-direction forces
Sum of moments around (x1 , y1 )
Come on, K. C.! Every undergraduate knows this stuff . .

And dont waste my time here in a nice place, please.
A top-down generalization via
Operational Form of Previous Equations
(We do not teach undergraduates this way!)
DAlembert-Lagrange Principal Equations
for 6-DOF Boomerang Model
What are they?

Where does one utilize them for?
What added value do they offer, if any?
What are they?
The transposition of the summation operator below
consists of three rigid-body modes.
Mathematically, ST fD = 0 is a projection of dAlemberts

forces onto the rigid-body modes.
What are they? - Contd
The dAlembert-Lagrange principal equations,
ST fD = 0,
represent the mean motions of flexible dynamical systems

for which the instantaneous mass center of the total system
is given, for the example problem, by
uc = (m1 u1 + m2 u2 + m3 u3 )/M, M = (m1 + m2 + m3)
so that the corresponding x-direction equation is given by
STx fxD = fx - M (d2 / d t2)uc = 0

What are they? - Concluded
Physically, the three DAlembert-Lagrange

Principal Equations given by
are self-equilibrium equations, and

ST is the self-equilibrium operator whose transposition,
S, in turn consists of the rigid-body modes, R, of the system.
R =S
Variational Representation of DAlembert-Lagrange
Principal Equations
Introduce the self-equilibrium displacement
so that one obtains the following variational statement:
which clearly reveals that

ST, in fact, is the self-equilibrium operator of the system.
What are they in partitioned mechanical systems?
What are they in partitioned systems?
What are they in partitioned systems? - Concluded.
For the partitioned modeling of flexible mechanical systems,

ST fD = 0 provides self-equilibrium condition for each partition.
Examples for which self-equilibrium conditions are important:
Where does one utilize them for?
1. If the applied forces are known, they provide the mean

motions of the system.
2. If part or all of the forcing functions are unknown and the

mean motions are measured, they provide a least-squares
solution of the applied forces and moments. For the
example problem, they can provide a least-squares estimate
of aerodynamic forces acting on the boomerang.
3. From the theoretical point of view, the dAlembert-

Lagrange principal equations provide the solvability
conditions for completely free or partially constrained
flexible systems, either quasi-static or dynamic. We will
examine this aspect later in the talk.
What added value do they offer, if any?
1. In the modeling of multi-physics problems, they provide

the principal (rigid-body modes) interface forces and moments,
viz., average interface forces and moments.
2. In multi-body dynamics, if properly formulated and

implemented, they provide the fundamental rigid-body motions
and the corresponding joint forces that can aid first-hand
physical insight for subsequent optimization, control and
baseline solutions for detailed analysis.
3. They facilitate the divide-and-conquer paradigm for the

modeling and solution of complex systems: a key property for
partitioned modeling and analysis.
4. In the iterative solution of large-scale problems, they provide a

crucial starting vector and subsequent filtering of residuals for
faster iterative convergence.
DAlembert-Lagrange Principal Equations for General 3-D
Systems (Lets get serious on their usage)
Step 1: Variational Statement of the dAlambert-Lagrange Principal Equations:

Lets get serious on their usage - contd
The preceding variational observation allows us to decompose the displacement as
Step 2: The summation operator for 3-dimensional problem:

For each discrete nodal point we have
Variational formulation of partitioned system
Partition Interface constraint functional
And B extracts the interface degrees of freedom for each partition.

Variational Statement for Partitioned Systems
Introduce the decomposition of the displacement into deformation

and rigid-body(DAlembert-Lagrange principal) modes:
Four-Variable Variational Formulation
for Partitioned Systems
DAlembert-Lagrange Principal Equations for
Partitioned Systems
This equation provides the mean gross motions of the total system!
Application 1: mass properties of complex structures
(Example - Bar element)
It is not trivial at all for computing the rotatory inertia and

its coupling matrix with the translational mass properties!
Translational mass matrix of a bar element:
Summation operator for the case when the moment is computed

around node 1
Application 1: mass properties of complex structures - Contd
(Bar element example)
The above matrix correctly captures the rotatory inertia

when the rotational motion is measured with respect to
node 1 and the corresponding coupling matrix with the
Translational mass. Complex cases yield the correct mass
properties. Computations of the rotatory inertia for
complex systems, e.g., automobile, satellites, etc.,
with the present procedure is straightforward and consistent.
All one needs for computing the inertia properties of

complex systems are the translational mass matrix and
the moment summation operator!
Small application: Generation of Diagonal Mass Matrix
Application 2: Solvability for unconstrained systems
under quasi-static equilibrium states
Partitioned equations of motion for structures:
This equation is indefinite and consequently requires a delicate

care for its solution!
Application 2: Solvability for unconstrained systems
under quasi-static equilibrium states
Eliminate q to obtain:
This equation is regular and consequently can be solved

in a routine manner!
Application 3: Consistent starting vector for
an iterative solution of partitioned system equations
Step 1: Solve for a least-square value of the interface forces:

Application 3: Consistent starting vector for
an iterative solution of partitioned system equations
Step 2: Project the new iterate to be orthogonal to the interface

rigid-body modes:
in order to minimize the residual:

Future Potential Applications
A. Least-squares nominal interface forces that may

provide a preliminary design modification or control
strategy for systems with constraints.
B. Augmented solution of (q, ") for dual control

strategy development, i.e., for principal motions
and deformational motions in tandem.
C. Filtering of mean motion signals from output signals.
D. Advanced multi-physics modeling

Discussions
1. The dAlembert-Lagrange principal equations consists of 6 rigid-body motions

regardless how large the flexible mechanical structural systems may be, and
they provide the mean motions of the overall system dynamics.
2. The rotatory inertia and its coupling terms with the translational mass
properties are obtained as part of the derivational process of the dAlembert-
Lagrange principal equations presented herein.
3. The dAlembert-Lagrange principal equations constitute the key solvability

condition for systems partially constrained or in completely free-free state.
4. For an iterative solution of coupled multi-physics problems, the solution of the

dAlembert-Lagrange principal equations provides a consistent starting vector,
thus accelerating the iterative process.
5. There remains a challenge to expand the usage of the dAlembert-Lagrane

principal equations, some of which have been outlined herein.
Viva la dynamique lmentaire!
Fin!
Question:
Can you readily get the necessary physical
insight from 108-DOF simulation results?
Wisdom from Native American Culture:

I can count five from the right fingers, and
I can count another five from the left fingers.
And, I dont know how to count further!

D'Alembert-Lagrange!s Principal Equations, Their Origin and Applications

Uploaded by

Copyright:

Available Formats

D'Alembert-Lagrange!s Principal Equations, Their Origin and Applications

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

D'Alembert-Lagrange!s Principal Equations, Their Origin and Applications

Uploaded by

Copyright:

Available Formats

D'Alembert-Lagrange!

K. C. Park, Carlos A. Felippa

DAlemberts Principle was Lagrange came to Paris from

Trait de Dynamique (1758) Mchanique Analytique (1788)

! It began with a single, free and rigid body

! Then, a rigid body with the rotational inertia properties;

! A rigid-link with joints;

! Lumped mass-spring models

! Continuum (flexible) models

! FEM/Rigid models with joints and constraints;

! And, the race for complex models is on!

As a result of the race for ever more complex modeling

One always returns to where one has started.

Hence, in dynamics, we may return to the simplicity and

Sum of forces (3 equations at most):

Sum of moments (3 equations at most):

Is a floating structure in an equilibrium condition?

Is a free-free substructure, partitioned from an assembled

The answer is not in the blowing wind,

***Warning: This model does not capture the boomerang motions!

Note that the inertia forces are associated with the

Note that uc changes with time in general for flexible structures.

Come on, K. C.! Every undergraduate knows this stuff . .

What are they?

consists of three rigid-body modes.

Mathematically, ST fD = 0 is a projection of dAlemberts

The dAlembert-Lagrange principal equations,

represent the mean motions of flexible dynamical systems

uc = (m1 u1 + m2 u2 + m3 u3 )/M, M = (m1 + m2 + m3)

so that the corresponding x-direction equation is given by

STx fxD = fx - M (d2 / d t2)uc = 0

Physically, the three DAlembert-Lagrange

are self-equilibrium equations, and

Introduce the self-equilibrium displacement

so that one obtains the following variational statement:

which clearly reveals that

For the partitioned modeling of flexible mechanical systems,

1. If the applied forces are known, they provide the mean

2. If part or all of the forcing functions are unknown and the

3. From the theoretical point of view, the dAlembert-

1. In the modeling of multi-physics problems, they provide

2. In multi-body dynamics, if properly formulated and

3. They facilitate the divide-and-conquer paradigm for the

4. In the iterative solution of large-scale problems, they provide a

Step 1: Variational Statement of the dAlambert-Lagrange Principal Equations:

Step 2: The summation operator for 3-dimensional problem:

And B extracts the interface degrees of freedom for each partition.

Introduce the decomposition of the displacement into deformation

It is not trivial at all for computing the rotatory inertia and

Translational mass matrix of a bar element:

Summation operator for the case when the moment is computed

The above matrix correctly captures the rotatory inertia

All one needs for computing the inertia properties of

Partitioned equations of motion for structures:

This equation is indefinite and consequently requires a delicate

This equation is regular and consequently can be solved

Step 1: Solve for a least-square value of the interface forces:

Step 2: Project the new iterate to be orthogonal to the interface