D Alembert Principle
D Alembert Principle
D Alembert Principle
DAlemberts principle
The principle of virtual work states that the sum of the incremental virtual works done by
all external forces F i acting in conjunction with virtual displacements si of the point on
which the associated force is acting is zero:
X
W =
F i si = 0.
(1.1.1)
i
This technique is useful for solving statics problems, with static forces of constraint. A static
force of constraint is one that does no work on the system of interest, but merely holds a
certain part of the system in place.
In a statics problem there are no accelerations. We can extend the principle of virtual work to
dynamics problems, i.e., ones in which real motions and accelerations occur, by introducing
the concept of inertial forces. For each parcel of matter in the system with mass m, Newtons
second law states that
F = ma.
(1.1.2)
We can make this dynamics problem look like a statics problem by defining an inertial force
F = ma
(1.1.3)
F total = F + F = 0.
(1.1.4)
DAlemberts principle is just the principle of virtual work with the inertial forces added to
the list of forces that do work:
X
X
W =
F i si +
F j sj = 0.
(1.1.5)
i
mg
s
Figure 1.1.1: Sketch of bead of mass m sliding frictionlessly on a vertical hoop of radius R
under the influence of gravity.
1.1.1
A trivial example would be a mass m falling under the effect of a constant gravitational field
g. With z positive upward, the force on the mass is mg and the work due to this force
under vertical displacement z is WG = mgz. The inertial force is m(d2 z/dt2 ) and the
work is WI = m(d2 z/dt2 )z. Setting the sum of the two to zero gives us
mgz m(d2 z/dt2 )z = mg + m(d2 z/dt2 ) z = 0,
(1.1.6)
from which we infer the expected result
d2 z
= g.
dt2
(1.1.7)
DAlemberts principle offers no advantages over normal procedures in this case. However,
it becomes more economical in problems with constraints.
1.1.2
Figure 1.1.1 illustrates the slightly more interesting problem of a bead sliding frictionlessly
around a vertically oriented hoop of wire. Here the force of gravity is not in the direction
of motion. The component of gravity normal to the hoop does no work on the bead. Nor
does the force of the hoop on the bead that constrains the bead to move in a circle. The
work on the bead due to gravity for a small displacement s = R along the wire is
WG = mgR sin . The acceleration of the bead also has two components, a radial
component ar = v 2 /R, where the tangential velocity is v = ds/dt = R(d/dt), and a
tangential component at = d2 s/dt2 = R(d2 /dt2 ). The radial component of the inertial
force mv 2 /R does no work. However, the tangential component mat = mR(d2 /dt2 )
does: WI = mR(d2 /dt2 )s = mR2 (d2 /dt2 ). DAlemberts principle thus gives us
mgR sin mR2 (d2 /dt2 ) = 0
(1.1.8)
s
m
M
(1.1.9)
(This could equally well have been expressed in terms of s rather than .)
1.1.3
Atwoods machine
DAlemberts principle makes the venerable Atwoods machine, illustrated in figure 1.1.2,
trivial to analyze. The state of the machine is determined by the positions of the two masses
along the U-shaped coordinate s looping over the frictionless pulley with the string. The
work done by gravity on the left-hand mass under the displacement s is WL = mgs,
while the gravity acting on the right-hand mass produces work WR = +M gs. The inertial
work on the two masses gives WI = (M + m)(d2 s/dt2 )s, resulting in
W = (M m)g (M + m)(d2 s/dt2 ) s = 0
(1.1.10)
or
M m
d2 s
=g
.
(1.1.11)
2
dt
M +m
Note that the force of constraint by the pulley (which is assumed to be free to rotate but
held rigidly in place and with negligible moment of inertia) does not enter the problem at
all.
1.2
Distributed forces
So far we have dealt with a limited number of forces that each act at a point on the system
of interest. However, forces like gravity as well as inertial forces act on each mass point in
the system, making the calculation of the virtual work more complex.
dr
d
ri
mass element
fixed axis
1.2.1
Lets return to the simple case of a mass falling in a uniform gravitational field. If the mass is
extended but not rotating, the vertical displacement of each mass element mi is the same:
z. The work due to gravity on each point is therefore the same: WGi = mi gz. In this
case the work due to the inertial force associated with the vertical acceleration is also simple:
WIi = mi (d2 z/dt2 )z where we have recognized that the accelerations associated with
each point d2 zi /dt2 are all the same as that of some reference point z fixed to the object.
There are also forces exerted on each mass element by neighboring mass elements that do
work. However, as we now sum over all mass elements, these internal forces cancel by virtue
of Newtons third law. Performing this sum, we find that
X
W = g + d2 z/dt2
mi z = g + d2 z/dt2 mz,
(1.2.1)
i
where m is the total mass of the object. This gives rise to the expected result
d2 z
= g.
dt2
1.2.2
(1.2.2)
Figure 1.2.1 shows a rotating disk with angular acceleration induced by a force F applied
on the right side of the disk. The work done by the force with an angular displacement
is WF = RF . the work done by the inertial force on mass element dmi is WIi =
dmi ri (d2 /dt2 )ri where ri is the radius at which the mass element is located. Ignoring
internal forces between mass elements as before, the total virtual work is therefore
#
"
X
2 2
2
W = RF
dmi ri (d /dt )
(1.2.3)
i
center of mass
(x,y)
(X,Y)
(x,y)
mass
element
Figure 1.2.2: Sketch for parallel axis theorem. The fixed point denoted by the cross becomes
the center of mass.
and the governing equation is
I
d2
= RF
dt2
(1.2.4)
dmi ri2 .
(1.2.5)
Using the methods of calculus, we write dmi = ri drddz where the (constant) mass density is and the variable z is normal to the page. Upon letting the differentials become
infinitesimal, I becomes
Z w Z 2 Z R
wR4
,
(1.2.6)
r3 drddz =
I=
2
0
0
0
where w is the thickness of the disk. Recognizing that the mass of the disk is M = R2 w,
elimination of the density results in
M R2
.
(1.2.7)
I=
2
The moments of inertia for other common shapes may be obtained by similar methods.
1.2.3
In the above case the moment of inertia was calculated about the center of mass of the
disk. Sometimes we need to make this calculation for rotations about some other point. The
parallel axis theorem shows how to do this once the moment of inertia about the center of
mass has been calculated.
Referring to figure 1.2.2, the moment of inertia about the origin of the illustrated mass,
which is in the form of an irregularly shaped plate of thickness h normal to the page, is
Z
Z Z
Z Z
2
2
I = h r dm = h
r dxdy = h
(x2 + y 2 )dxdy,
(1.2.8)
where in general the density (x, y) may depend on position in the plane and where the
bounds on the integral extend over the mass. Let us substitute (x, y) = (X, Y ) + (x0 , y 0 )
where (X, Y ) points to a fixed location. The above integral then becomes
Z Z
2
2
I =(X + Y )h
dxdy
Z Z
+2h
(Xx0 + Y y 0 )dx0 dy 0
Z Z
+h
(x02 + y 02 )dx0 dy 0 .
(1.2.9)
(The change of variables in the integral is trivial, as dx = dx0 and dy = dy 0 .) The first term
on the right side of the equals sign is just the mass M of the object times the square of the
distance from the origin to the fixed point R2 = X 2 + Y 2 . The third term is the moment of
inertia about the fixed point. Anticipating that this point will become the center of mass,
we donote the moment of inertia about this point ICM .
The second term can be written in the form of a dot product
Z Z
2(X, Y ) h
(x0 , y 0 )dx0 dy 0
(1.2.10)
which must be zero for all choices of the original origin if (X, Y ) is to become the location of
the center of mass. To see why this is so, substitute (x0 , y 0 ) = (x, y) (X, Y ) in the integral
in equation (1.2.10) and change from primed to unprimed integrand variables:
Z Z
Z Z
h
(x, y)dxdy (X, Y )h
dxdy = 0.
(1.2.11)
The coefficient of (X, Y ) is just the mass of the object, whereas the first term is just the
mass times mass-weighted average of (x, y). This is just the mass times the center of mass,
and therefore
Z Z
h
(x, y)dxdy
(1.2.12)
(X, Y ) =
M
is just the position of the center of mass. Equation (1.2.9) becomes
I = M R2 + ICM
(1.2.13)
a statement of the parallel axis theorem: the moment of inertia about an arbitrary axis (for
a rotation axis normal to the x-y plane) is just the moment of inertia for rotations about the
center of mass (also normal to the x-y plane) plus the mass times the square of the distance
between the arbitrary axis and the center of mass.
1.2.4
The problem of a disk rolling down a ramp under the influence of gravity can be solved
economically using the parallel axis theorem. Two real forces act on the disk, gravity and
M
R
W =
M R2
2
3M R2 d2
M gR sin
2 dt2
(1.2.14)
= 0.
(1.2.15)
(1.2.16)
Thus, the acceleration down the ramp in rolling motion is less than would be expected from
frictionless sliding, g sin .
1.2.5
The inertial virtual work in general for a solid body in two-dimensional rotation and translation can be written as the sum of the works for each mass element of the solid:
WI =
X
i
mi
d2 r i
r i
dt2
(1.2.17)
where r i is the position vector of the ith mass element. Let us make an initially arbitrary
split of this position vector as follows:
r i = R + r 0i
(1.2.18)
where R represents a pivot point, to be chosen later. Substituting this into equation (1.2.17)
results in a rather complex expression
2
X
dR
d2 R
d2 r 0
d2 r 0
0
0
WI =
mi
(1.2.19)
R + 2 r + 2 R + 2 r .
2
dt
dt
dt
dt
i
This can be reduced to
WI = M
X
d2 R
d2 R
d2 r 0CM
d2 r 0i
0
m
r 0i
i
CM
2
dt2
dt2
dt2
dt
i
(1.2.20)
where the position of the center of mass relative to the pivot point is
r 0CM =
1 X
mi r 0i .
M i
(1.2.21)
We are now faced with choosing a pivot point that simplifies equation (1.2.20). The most
obvious choice is to take the pivot point as the center of mass of the object. With this choice,
r 0CM = 0, since the position of the center of mass relative to the center of mass is the zero
vector. In this case, the inertial virtual work is
WI = M
d2 RCM
d2
CM
CM
dt2
dt2
(1.2.22)
where RCM is the position of the center of mass of the object relative to the original origin.
The first term in equation (1.2.22) is the virtual inertial work that would occur if all of
the mass were concentrated at the center of mass. The second term is the virtual work for
rotations about the center of mass, such as was computed in section 1.2.2. The variable ICM
is the moment of inertia about the center of mass and is the corresponding rotation angle.
Another choice that is useful in certain situations occurs when the pivot point R is fixed in
space, so that d2 R/dt2 = 0 and R = 0. If the mass element located at the pivot point is
instantaneously stationary, then equation (1.2.20) reduces to
d2
WI = I 2
dt
(1.2.23)
at that instant, as was done in section 1.2.4 for the disk rolling down the ramp. In that case
the pivot point is the point of contact of the disk with the ramp at a particular time and
the moment of inertia I is that appropriate to rotations about that point. Since R is fixed,
all of the motion of each mass element is contained in changes in r i with time.
In section 1.2.4 we implicitly assumed that the gravitational force, which acts separately on
each mass element of the disk, can be considered to act solely on the center of mass, i.e., the
center of the disk. This seems obvious in the case of a disk, but is less obvious for a more
irregularly shaped object. We now prove that the gravitational force effectively acts on the
center of mass of an object regardless of its shape.
Fa
stationary
moving
Fs
Fk
Figure 1.3.1: Two cases of friction. On the left the object is stationary and the frictional
force Fs is in balance with the applied force Fa . On the right the object is moving and the
frictional force Fk is related to the normal force on the object N . The applied force is not
necessarily equal to the frictional force.
The total gravitational work due to small virtual displacements of mass elements can be
written
X
WG =
dmi g r i
i
=M g RCM
(1.2.24)
since the mass-weighted average of the positions of mass elements is simply the center of
mass. One note of caution: The above result holds for any force that acts uniformly across
the body in question. However, if the force is non-uniform, then the resulting virtual work
is more complex. An important example is the total gravitational force on a body in a
non-uniform gravitational field.
1.3
Effects of friction
The laws of friction are empirical and are not satisfied exactly. There are two cases, static
and kinetic friction. For kinetic friction on an object in contact with a surface, the frictional
force is directed opposite the velocity of the object (assuming that the surface is at rest) and
has magnitude of
F k = k N
(1.3.1)
where N is the normal force exerted by the surface on the object and k is the dimensionless
coefficient of kinetic friction. In the static case the frictional force is that needed to keep the
object stationary in view of the other forces acting on the object. There is an upper limit
on the magnitude of the static frictional force equal to
F s s N
(1.3.2)
(1.3.3)
10
b
CM???
x
a
1.4
Problems
1. Compute the location of the center of mass of the triangular piece of steel of uniform
thickness illustrated in figure 1.4.1.
pivot point
R/2
Figure 1.4.3: The disk is free to rotate in the plane of the page around the illustrated pivot
point.
11
hinge
broken
brace
L
m
m
2