J. Peraire
16.07 Dynamics
Fall 2004
Version 1.1
Lecture D17 - Conservation Laws for Systems of Particles
In this lecture, we will revisit the application of Newton’s second law to a system of particles and derive
some useful relationships expressing the conservation of angular momentum.
Center of Mass
Consider a system made up of n particles. A typical particle, i, has mass mi, and, at the instant considered,
occupies the position ri relative to a frame xyz. We can then define the center of mass, G, as the point
whose position vector, rG, is such that,
rG = 1m(
nsummationdisplay
i=1
miri) . (1)
Here, m is the total mass of the system given by m =
nsummationdisplay
i=1
mi.
It is important to note that the center of mass is a property of the system and does not depend on the
reference frame used. In particular, if we change the location of the origin O, rG will change, but the
absolute position of the point G will not. Often, it will be convenient to describe the motion of particle i as
the motion of G plus the motion of i relative to G. To this end, we introduce the relative position vector,
r′i, and write,
ri = rG +r′i . (2)
It follows immediately, from 1 and 2, that,
nsummationdisplay
i=1
mir′i =
nsummationdisplay
i=1
mi(ri ?rG) =
nsummationdisplay
i=1
miri ?mrG = 0 . (3)
1
Forces
We will consider two types of forces acting on the particles:
External forces arising outside the system. We will denote the resultant of all the external forces acting
on particle i as Fi.
Internal forces due to pairwise particle interactions. Let fij denote the force that particle j exerts on
particle i. We shall assume that this force is directed along the line joining particle j and particle i.
This implies that (ri ?rj) ×fij = 0. We shall also assume that these internal forces obey Newton’s
third law. That is, fij = ?fji, where fji is the force that particle i exerts on particle j. The total
internal force on particle i is then,
nsummationdisplay
j=1, jnegationslash=i
fij ,
and, if we sum over all particles, we have
nsummationdisplay
i=1
nsummationdisplay
j=1 jnegationslash=i
fij = 0 ,
since, for every force, fij, there is an equal and opposite force, fji.
Conservation of Linear Momentum
The linear momentum of the system is defined as,
L =
nsummationdisplay
i=1
mivi . (4)
From equation 2, we have that vi = ˙ri = ˙rG + ˙r′i, which, combined with the above equation, gives,
L =
nsummationdisplay
i=1
mi(˙rG + ˙r′i) =
nsummationdisplay
i=1
mivG + ddt(
nsummationdisplay
i=1
mir′i) = mvG , (5)
sincesummationtextni=1 mir′i = 0. We now consider the time variation of the linear momentum. If we assume that the
reference frame xyz is inertial, then, starting from equation 4, we have,
˙L =
nsummationdisplay
i=1
miai =
nsummationdisplay
i=1
(Fi +
nsummationdisplay
j=1 jnegationslash=i
fij) =
nsummationdisplay
i=1
Fi = F,
where F is the sum of all external forces acting on the system. Thus, for a system of particles, we have that,
˙L = F . (6)
Note that, from equation 5, we can also write ˙L = maG.
2
Conservation of Angular Momentum
Since the angular momentum is defined with respect to a point in space, we will consider several cases, using
a different reference point for each case.
Conservation of Angular Momentum about a Fixed Point O
The angular momentum of a system of particles about a fixed point, O, is the sum of the angular momentum
of the individual particles,
HO =
nsummationdisplay
i=1
(ri ×mivi) . (7)
The time variation of HO can be written as,
˙HO =
nsummationdisplay
i=1
( ˙ri ×mivi)+
nsummationdisplay
i=1
(ri ×mi ˙vi) = 0+
nsummationdisplay
i=1
(ri ×(Fi +
summationdisplay
j=1, jnegationslash=i
fij)) =
nsummationdisplay
i=1
(ri ×Fi) .
Here, we have used two facts. First, the cross product of two parallel vectors is zero. Second, ri × fij +
rj × fji = (ri ? rj) × fij = 0, and, therefore, the internal forces have no net effect on the total angular
momentum change of the particle system. Thus, we have that
˙HO = MO , (8)
where MO =summationtextni=1(ri ×Fi) is the total moment, about O, of the applied external forces. Note that this
moment needs to be computed as the sum of the individual moments. In general MO negationslash= rG ×F, unless the
line of action of the the external forces resulant, F, passes through G.
Conservation of Angular Momentum about G
The angular momentum about G is given by,
HG =
nsummationdisplay
i=1
(r′i ×mivi) . (9)
Taking the time derivative of equation 2, we obtain
vi = ˙ri = ˙rG + ˙r′i = vG +v′i . (10)
Inserting this expression into equation 9, we obtain
HG =
nsummationdisplay
i=1
(r′i ×mi(˙rG + ˙r′i)) =
nsummationdisplay
i=1
(r′i ×miv′i) , (11)
sincesummationtextni=1(r′i ×mi˙rG) = ?˙rG ×summationtextni=1 mir′i = 0 (see equation 3). We note that equations 9 and 11 give us
alternative representations for HG. Equation 9 is called the absolute angular momentum (since it involves
absolute velocities, vi), whereas equation 11 is called the relative angular momentum (since it involves
velocities, v′i, relative to G). It turns out that when G is chosen to be the origin for the relative velocities,
3
both the absolute and relative angular momentum are identical. As we shall see, this will simplify our
analysis significantly. In general, the absolute and relative angular momentum with respect to an arbitrary
point are not the same.
We can now go back to equation 9 and consider the time variation of HG,
˙HG =
nsummationdisplay
i=1
(˙r′i ×mi(vG + ˙r′i)) +
nsummationdisplay
i=1
(r′i ×mi ˙vi) = 0+
nsummationdisplay
i=1
(r′i ×Fi) . (12)
In the above equation, the term ˙r′i × mi˙r′i is clearly zero, and summationtextni=1(˙r′i × mivG) = ?vG ×summationtextni=1 mi˙r′i =
?vG ×d(summationtextni=1 mir′i)/dt = 0. Thus, we have that
˙HG = MG . (13)
Here, MG =summationtextni=1(r′i ×Fi), is the total moment, about G, of the applied external forces.
The above expression is very powerful and allows us to solve, with great simplicity, a large class of problems
in rigid body dynamics. Its power lies in the fact that it is applicable in very general situations:
? In the derivation of equation 13, we have made no assumptions about the motion of the center of mass,
G. That is, equation 13 is valid even when G is accelerated.
? We have implicitly assumed that the reference frame used to describe r′i in equation 10 is non-rotating
with respect to the fixed frame xyz (otherwise, we would have written ˙r′i = v′i + ω′ ×r′i, with ω′, the
angular velocity of the frame considered). It is not difficult to show that equation 13 is still valid if the
reference frame rotates, provided the angular velocity is constant. If the reference frame rotates with
a constant angular velocity, the angular momentum will differ from that of equations, 9 and 11 by a
constant, but equation 13 still will be valid (see approach 2 in example below).
? The angular momentum of the particle system about the center of mass, G, can be evaluated using two
alternative representations, 9 and 11. In particular, equation 11 allows us to determine HG without
making any reference to the absolute velocities, vi, or position vectors, ri.
Finally, by combining equations 7 and 9, the angular momentum about a fixed point, O, can be expressed
as a function of the angular momentum about the center of mass, as,
HO = rG ×mvG +HG.
Example Rotating dumbbell [1]
We consider a steadily spinning turntable with angular velocity ?, with a symmetric dumbbell mounted at
a distance R from the center. The dumbbell is free to spin about its midpoint. We assume that the rod
connecting the two masses is massless. Therefore, we can model the dumbbell as a system consisting of
two particles. We want to determine the relative motion of the dumbbell. We will consider three different
approaches:
4
1) Use equation 13 with parallel axes
In this case, we are clearly under the assumptions of equation 13, and, therefore, we have that the sum
of the moments of all external forces must equal the change in angular momentum. Since the dumbbell
is symmetric, the center of mass is at the support point. The only external force to the system (force
at the support) has a zero moment with respect to G, that is, MG = 0. Therefore, ˙HG = 0, or HG =
constant. The constant value of HG is determined by the initial conditions. The velocity of the masses
relative to G is L ˙θ′, and the magnitude of the angular momentum of the system about G is simply,
HG = 2LmL˙θ′ = 2mL2 ˙θ′ .
Since HG is constant, we have that the angular velocity, ˙θ′, relative to the parallel axes x′y′, must also
be a constant.
2) Equation 13 with rotating axes
Here, we assume that the local axes x′y′ are rigidly attached to the turntable. Since the angular
velocity of the turntable is constant, we can also apply equation 13, and conclude that ˙θ must be a
constant. Note that although θ is different from θ′ (from approach 1), ˙θ′ and ˙θ only differ by a constant
(˙θ = ˙θ′ +?), and, hence, this result is compatible with that established in the previous approach (i.e.
the integration constants will be different). Note that, in this case, the angular momentum relative
5
to the rotating x′y′ axes differs from the angular momentum in approach 1 (where the axes remained
parallel) by a constant (= 2Lm?).
3) Solve directly from Newton’s second law
Here, we will calculate the acceleration of each mass with respect to a fixed reference frame. Using the
same axes as in approach 2, the acceleration of particle 1, a1, relative to the fixed axes xy, is
a1 = aG + (a1)x′y′ + 2?×(v1)x′y′ +?×(?×r′1) ,
Here, (v1)x′y′ and (a1)x′y′ are, respectively, the velocity and acceleration of particle 1 observed by the
rotating frame, and are given by
v1 = L˙θuθ
a1 = ?L˙θ2ur + L¨θuθ ,
where ur and uθ are the local polar coordinates unit vectors for particle 1. Also,
aG = ??2Rj′ = ??2R(sinθur + cosθuθ) ,
where j′ is the unit vector in the direction of y′. Thus,
a1 = (??2Rsinθ ?L˙θ2 ?2?L˙θ ??2L)ur + (??2Rcosθ + L¨θ)uθ .
Similarly, for particle 2,
a2 = (??2Rsinθ2 ?L˙θ22 ?2?L ˙θ2 ??2L)ur2 + (??2Rcosθ2 + L¨θ2)uθ2.
Here, θ2 = θ + pi is the circumferential coordinate for mass 2, and ur2 (= ?ur) and uθ2 (= ?uθ) are
the local unit radial and circumferential vectors for mass 2.
Since a1 and a2 are referred to the fixed frame, we can obtain the forces acting on the masses by
applying Newton’s second law. Thus, F1 = ma1 and F2 = ma2, where F1 and F2 denote the forces
on particles 1 and 2, respectively. Now, we observe that any force on the masses will be transmitted
through the rod.
6
Since we are assuming that the mass of the rod connecting the two masses is zero, the tangential
component of the two forces acting on the rod should be in equilibrium (otherwise the rod would
experience an infinite angular acceleration). Thus, (F1)θ + (F2)θ = 0. This results in,
m(??2Rcosθ + L¨θ) + m(??2Rcosθ2 + L¨θ2) = m(??2Rcosθ + L¨θ) + m(?2Rcosθ + L¨θ) = 0 ,
which, in turn, implies that ¨θ = 0, consistent with the previous results.
Conservation of Angular Momentum about an arbitrary point P
In some situations, it may be convenient to refer the applied moments to an arbitrary point, P.
The sum of the moments, about point P, MP , will be given by
MP = MG +r′′G ×F .
Using equations 13 and 6, we have that,
˙HG +r′′G ×maG = MP .
It is also possible to derive expressions for the time variation of the absolute angular momentum, HP, and
relative angular momentum, (HP)rel taken about point P. Note that, in general, HP negationslash= (HP)rel; and that
˙HP negationslash= MP and ( ˙HP)rel negationslash= MP. The expressions for the time variation of the angular momentum about
an arbitrary point, P, involve the velocity and acceleration of point P, and can be found in reference [2].
Kinetic Energy for Systems of Particles
Here, we derive the expression for the kinetic energy of a system of particles that will be used in the
following lectures. A typical particle, i, will have a mass mi, an absolute velocity vi, and a kinetic energy
Ti = (1/2)mivi ·vi = (1/2)miv2i .
7
The total kinetic energy of the system, T, is simply the sum of the kinetic energies for each particle,
T =
nsummationdisplay
i=1
Ti =
nsummationdisplay
i=1
1
2miv
2
i .
It is convenient to decompose the velocity of each particle, vi, into the velocity of the center of mass, vG,
and the velocity relative to the center of mass, ˙r′i. Then,
T =
nsummationdisplay
i=1
1
2mi(vG + ˙r
′
i) · (vG + ˙r
′
i) =
nsummationdisplay
i=1
1
2mi(v
2
G + 2vG · ˙r
′
i + ˙r
′
i
2) = 1
2mv
2
G +
nsummationdisplay
i=1
1
2mi ˙r
′
i
2 ,
since vG ·summationtextni=1 mi˙r′i = 0, andsummationtextni=1 mi = m. Thus, we see that the kinetic energy of a system of particles
equals the kinetic energy of a particle of mass m moving with the velocity of the center of mass, plus the
kinetic energy due to the motion of the particles relative to the center of mass, G.
ADDITIONAL READING
J.L. Meriam and L.G. Kraige, Engineering Mechanics, DYNAMICS, 5th Edition
4/1, 4/2, 4/3 (kinetic energy expression only), 4/4, 4/5 (momentum only)
References
[1] M. Martinez-Sanchez, Unified Engineering Notes, Course 95-96.
[2] J.L. Merriam and L.G. Kraige Engineeering Mechanics, Dynamics, Fifth Edition, Wiley, 2002.
8