2 Planar Mechanics of a Rigid Body

Notes for this topic are building on chapters 16 & 17 of the reference text (Engineering
Mechanics – Dynamics, by RC Hibbeler), with added information & examples for 2CD4.
Figures are from the 13th edition of Hibbeler.

This set of notes explores how to analyze the motion of rigid bodies (i.e., not just a particle, but a
system of particles connected by internal constraints that hold them at fixed relative distance to
each other), and presents a number of useful results and simplification to make that analysis
easier as long as they're moving in a plane (rather than in 3 dimensions).

2.1 16. Planar Kinematics of a Rigid Body

This first section (2.1, based on Chapter 16 in Hibbeler) works out relations between position,
velocity, and acceleration of points on a rigid body in terms of those of other points on that body.
e.g., if you know the position, velocity, and acceleration of point A on a rigid body, what are the
position, velocity, and acceleration of another point B on that same body? This section won't
examine how these are linked to the causes of that motion (e.g., forces and moments), we'll get
into that in the next section (2.2; Hibbeler Chapter 17).

2.1.1 16.1: Planar Rigid-Body Motion

Translation: line in the body remains parallel to its original orientation.

Rotation about fixed axis: all particles in body (except those on the axis) move in a circular path.
General plane motion: combination of translation in a plane and rotation about an axis
perpendicular to the plane.

2.1.2 16.2: Translation

Position of any 2 points in the body are related at any time by: ("position of B
relative to A")
In [rectilinear or curvilinear] translation, is a constant (both in magnitude & direction), so
this means velocity and acceleration must be the same for all points in the body:

2.1.3 16.3: Rotation [of a rigid body] about a Fixed Axis

[differential] Angular displacement:

Angular velocity:

Angular acceleration:
combining these and using the chain rule, we can also get the following (just like for linear
motion):

If constant angular acceleration, then we obtain these angular kinematic equations:

Consider a point at position , a displacement from the axis of rotation:

For this point,

Angular acceleration vector is the time derivative of angular velocity vector:

(or )

We can express linear acceleration in terms of normal and tangential components; and

, where , have

, and
Or, in vector formulation,

=
2.1.4 *16.4: Absolute Motion Analysis
Even if something is undergoing general plane motion, you can sometimes analyze it by
combining the rotation about a fixed axis with displacement by displacing that axis (see section
16.5).

For now, we can deal with kinematics where the rotation axis is fixed:

2.1.4.1 example 16.5 from Hibbeler

In the following figure, find when if the piston is expanding at a rate of

Solution:
B is described by (fixed radius rotation about O), and since OB and OA are fixed lengths, we
can use some trig to relate and s. Specifically, (coslaw)
Taking derivatives, , so
> restart:
OA:=2: OB:=1: ang:=Pi/6:
coslaw:=sqrt(OA^2+OB^2-2*OA*OB*cos(theta)):
s30:=evalf(subs(theta=ang, coslaw));
thetadot:=solve(2*s*sdot=2*OA*OB*sin(theta)*thetadot, thetadot);
thetadot30:=evalf(subs([theta=ang, s=s30, sdot=.5], thetadot));

Solution in general worked out with Maple:

> restart:
lam1:=(s(t))=sqrt(2^2+1^2-2*2*1*cos(theta(t)));
#assume(s(t),positive);
diff(lam1, t);
omega(t):=subs(diff(s(t),t)=.5, solve(%, diff(theta(t),t)));
s30:=solve(subs(theta(t)=30*Pi/180, lam1), s(t)); evalf(%);
omega30:=subs([theta(t)=30*Pi/180], omega(t)); evalf(%);
alpha(t):=diff(omega(t),t);
alpha30:=subs([theta(t)=30*Pi/180, diff(theta(t),t)=omega30],
alpha(t)); evalf(%);
Make the angle a variable, and you can work out some interesting results; this mechanism
requires infinite angular acceleration around 0 and 180o, but very small angular acceleration for
most of the motion:
> restart:
lam1:=(s(t))=sqrt(2^2+1^2-2*2*1*cos(theta(t)));
#assume(s(t),positive);
#Ang:=60*Pi/180:
diff(lam1, t);
omega(t):=subs(diff(s(t),t)=.5, solve(%, diff(theta(t),t)));
s30:=solve(subs(theta(t)=Ang, lam1), s(t)); evalf(%);
omega30:=subs([theta(t)=Ang], omega(t)); evalf(%);
alpha(t):=diff(omega(t),t);
alpha30:=subs([theta(t)=Ang, diff(theta(t),t)=omega30],
alpha(t)); evalf(%);
plot(alpha30, Ang);
2.1.5 16.5: Relative-Motion Analysis: Velocity
If a rigid body is undergoing general plane motion, then it's often convenient to have formulas to
express velocity and acceleration of points in that body in terms of another point in it. In
particular, if the rigid body is rotating about its point O' then we solve problems more easily if
we develop a way to describe the motion of any point in terms of how point O' is translating,
since the body's rotation doesn't move point O'.
Let r be the location relative to a fixed reference O and let be the location relative to a
translating reference at point A (some point on the rigid body, possibly that the body is rotating
about):

Absolute position of B can be expressed via A's absolute position and B's position relative to A:

Displacement is then
And velocity is:

If the rigid body AB is rotating about A with angle vs. time (where is the

angle about the x' axis that B is at relative to A), then we can write:
and relative velocity is:

Regardless of whether phase offset (i.e., initial angle) the velocity would still be

Or, by taking the cross product directly (for planar motion),

This result on velocity ( ) is useful for directly determining velocity
of any point on the body if you know the velocity of one of its pins, as in the next example.

2.1.5.1 Example 16.6 from Hibbeler

Find vB when theta = 45o:

Could take either A or B as the pivot point for the previous equations. Since we know velocity
of A, that means less to determine in:

We know as well that , and that by the right

hand rule.
Also, know that from the constraint imposed by the rail.
This gives an equation we can solve for omega:

> restart:
solve([
vBx=omega*0.2/sqrt(2),
0=-2+omega*0.2/sqrt(2)]);

So, vBx = 2 m/s.

(Note, we also could have found this at this angle by symmetry, so that serves as a good check of
the more general procedure used here)

Alternatively, without the info from this section, you could start with position and differentiate it:

We can avoid the angle derivative by instead first substituting and using that

which gives the same result as above:

> restart:
yA:=1/5*cos(Pi/4); vA:=2:
vB:=5/sqrt(1-(5*yA)^2)*(yA*vA); evalf(%);

2.1.5.2 Example 16.7 from Hibbeler

Find the instantaneous velocity of point A if the cylinder rolls on the conveyor belt without
slipping:
Solution:
Let's try using the origin of the cylinder to translate to point A's velocity.

where

and

So, .

Note: could also have used point B as the intermediate reference, since it isn't moving relative to
the conveyor so . In that case, observe that is the same for every point on the
cylinder, so we can still say

(omega is negative because it's into the page, so in the -k direction for the cross product result we
worked out earlier).
2.1.6 *16.6: Instantaneous Centre of Zero Velocity
In general planar motion a point that has 0 velocity at a given instant is an instantaneous centre
of zero velocity (IC) (e.g., the contact point for a wheel in rolling motion on a fixed surface).
Using an IC as the translation point A, we'd get that the velocity of any other point is

Since , this means , in a direction perpendicular to :

FYI: There's various ways to find the IC depending on what you know:
1) Use and solve for if you know the velocity of some point A and know
omega:
2) If you know the direction of two nonparallel velocity vectors, put those as tangents to a circle
centred on IC:

3) If you know relative magnitude of two parallel velocity vectors, use triangles to locate IC:
Note: An IC is convenient for finding instantaneous velocity of other points, but it doesn't (in
general) have 0 acceleration, so isn't necessarily so convenient for finding accelerations of other
points.
2.1.6.1 Example 16.13

(Note: the piston is the block at C. Note also that its angular velocity is labelled, but maybe it's
wrong, so try to check that as well.)

Solution 1: without IC
We can find the velocity at B from trig:
Since we know the velocity at point B, we can say that
and since is fixed, we can further say that:
But this doesn't by itself let us determine ; that relies on the constraint that points C and A
stay vertically aligned due to the piston's channel. Note that the velocities are related to rates of
change of these angles:
and that (sine law)
So then, (since BC and AB are fixed lengths)

> restart: with(LinearAlgebra):

deg2rad:=3.14159/180:
theta:=13.6*deg2rad:
phi:=45*deg2rad:
rAB:=.25: #ft
rBC:=.75:
rABVec:=rAB*<-sin(phi), cos(phi),0>;
rBCVec:=rBC*<sin(theta), cos(theta),0>;
omegaAB:=10: omegaABVec:=<0,0,-omegaAB>;
vB:=omegaABVec &x rABVec;
omegaBC:=rAB/rBC*(cos(phi)/cos(theta))*omegaAB;
omegaBCVec:=<0,0,omegaBC>:
vC:=vB+omegaBCVec &x rBCVec;
Solution 2: with IC
In this problem if we can locate the IC for BC then we can determine in a more
straightforward way from .
We know both magnitude and direction of , but only know magnitude of , so we can use
this method:

In this problem, that means the IC is located as follows:

We can find the angles in this new triangle by
1. noting is off from , so the angle at IC is also 45°,
2. Saying the angle at C must be 90° – 13.6° = 76.4°
3. Saying the final angle (at B) must add to the other two to produce 180°; i.e., 58.6°.
Then the law of sines lets us use rBC to find the other lengths:

With these, we can determine that and :

> restart:
deg2rad:=3.14159/180:
rBC:=.75: vB:=.25*10;
#angles:
ic:=45*deg2rad: c:=(90-13.6)*deg2rad: b:=3.14159-ic-c: b/deg2rad;
solve([
rBIC/sin(c)=rBC/sin(ic),
rCIC/sin(b)=rBC/sin(ic),
vB=omegaBC*rBIC,
vC=omegaBC*rCIC
]);
2.1.7 16.7: Relative-Motion Analysis: Acceleration
Differentiating the velocity equation leads to:

Especially if the is a constant (e.g., A and B are on the same rigid body and B rotates about
A, even though A is moving, as with the previous few sections), it's convenient to express the
relative acceleration in terms of tangential and normal components:

If is fixed length (e.g., both A and B parts of the same rigid object), then the only relative
motion of B from A's perspective can be rotation, and we get the same results as in 16.3:

So, [if A and B are a fixed length from each other, as in a rigid body] their accelerations are
related by

where and are the angular velocity and acceleration of (i.e., of the rigid body) about
point A.
2.1.7.1 Example 16.18

Solution, method 1:
Now we can say , and though we know (from example 16.13),
we need to find out . Using the IC this time is complicated by the fact that the IC is
changing locations in time.

We can still say (as in the solution to e.g. 6.13 from the previous section) that

So then, modifying the previous solution we can find

> restart: with(LinearAlgebra):
deg2rad:=3.14159/180:
theta:=13.6*deg2rad:
phi:=45*deg2rad:
rAB:=.25: #ft
rBC:=.75:
rABVec:=rAB*<-sin(phi), cos(phi),0>:
rBCVec:=rBC*<sin(theta), cos(theta),0>:
omegaAB:=10: omegaABVec:=<0,0,-omegaAB>:
vB:=omegaABVec &x rABVec:
omegaBC:=rAB/rBC*(cos(phi)/cos(theta))*omegaAB;
omegaBCVec:=<0,0,omegaBC>:
#vC:=vB+omegaBCVec &x rBCVec:
alphaAB:=20:
aBVec:=<0,0,-alphaAB> &x rABVec - omegaAB^2*rABVec;
alphaBC:=solve(
cos(phi)/rBC*(-alphaAB)-sin(phi)/rBC*(omegaAB)^2
=cos(theta)/rAB*(-alphaBC)-sin(theta)/rAB*(omegaBC)^2);
aCVec:=aBVec + <0,0,alphaBC> &x rBCVec - omegaBC^2*rBCVec;

Solution method 2:

In the previous solution, the law of sines relation that led to equations we could
solve for and were based on C only moving vertically. Another approach (especially
with a computer algebra system) is to leave these angular velocities as unknowns and use the
resulting vector equations for velocity and acceleration to determine what they'd need to be to
make C only move vertically, e.g.,
> restart: with(LinearAlgebra):
deg2rad:=3.14159/180:
theta:=13.6*deg2rad:
phi:=45*deg2rad:
rAB:=.25: #ft
rBC:=.75:
rABVec:=rAB*<-sin(phi), cos(phi),0>:
rBCVec:=rBC*<sin(theta), cos(theta),0>:
omegaAB:=10: omegaABVec:=<0,0,-omegaAB>:
vB:=omegaABVec &x rABVec:
#omegaBC:=rAB/rBC*(cos(phi)/cos(theta))*omegaAB;
omegaBCVec:=<0,0,omegaBC>:
vC:=vB+omegaBCVec &x rBCVec:
omegaBC:=solve(vC[1]=0);
vC:=vB+omegaBCVec &x rBCVec;
alphaAB:=20:
aBVec:=<0,0,-alphaAB> &x rABVec - omegaAB^2*rABVec;
aCVec:=aBVec + <0,0,alphaBC> &x rBCVec - omegaBC^2*rBCVec:
alphaBC:=solve(aCVec[1]=0);
aCVec:=aBVec + <0,0,alphaBC> &x rBCVec - omegaBC^2*rBCVec;

2.1.8 16.8: Relative-Motion Analysis using Rotating Axes

Sometimes it's convenient to use a coordinate system that both translates and rotates (rather than
the translating ones we dealt with so far); e.g., when we need to consider the rotation of the earth
during a rocket trajectory, or analyzing two points on a mechanism that are not located on the
same body (so you can't say the distance between them is constant like before).
Position is still , where , and & j are unit vectors of the x & y
axes.
(except now isn't a constant, and neither are or due to the rotation of the x & y axes).
Let and be the angular velocity and angular acceleration of the x,y axes.

Velocity is then
where

= "velocity of B relative to an observer moving (including rotating) with A in the xy coordinate

system" + cross product of angular velocity of that system and position of B relative to A.
Note: we don't need to specify here whether is in terms of xy or XY coordinate systems
since the vector is the same either way, but when using its components (e.g., to add them to

) you'll need to make sure the coordinate systems you're using match.

Together, this is

Acceleration:
and

while (just like for above when finding

velocity)
So, in total we'll have:

Compared to the result for a rotating but not translating coordinate change,

we have two new terms:

1. is the Coriolis acceleration (from B moving relative to xy frame and

rotation of xy frame)

2. acceleration of B with respect to A as observed from perspective of xy frame.

Note that if A and B are both on a rigid body then is fixed, and if the body also rotates with

the reference frame we have so this reduces to the previous

.

2.1.8.1 e.g., rotating reference frame

A tour bus takes a sharp highway exit ramp in the shape of a quarter-circle curving left at a
much-higher-than-recommended speed of 120 kph executing the 90° left turn in only 4 seconds.
a) Determine the angular velocity of the tour bus during this turn
b) Determine the acceleration experienced by the driver (which is seated stationary relative to the
bus's reference frame) in both the bus's reference frame and a stationary reference frame fixed to
the ground. What force must the bus exert on the bus driver to provide this acceleration?
c) During this motion, a bandmember walks from the front of the bus towards the back of the bus
at a constant speed of 1 m/s relative to the bus. Determine the necessary average frictional force
on their shoes to maintain this constant speed in this reference frame.
d) Does "centrifugal force" exist in the reference frame during this turn? If so, what is it for the
driver and the walking bandmember?

Solution:
a)
First, work out the bus's angular velocity and radius of the turn it's taking:
> restart:
vbus:=120*1e3/3600;
t_turn:=4:
s_turn:=vbus*t_turn:
Omega:=(Pi/2.)/t_turn;
R:=vbus/Omega;

b)
Let the stationary reference frame (point A) be at the centre of rotation of this curved path, and
let the bus's reference frame be translating and rotating with the bus (point B).

Then , , and

For any point C in the bus, if we know its velocity and acceleration with respect
to the bus in the bus's coordinate system (xy) then we can determine its velocity and acceleration
with respect to the ground by writing:

Relative to the bus, the driver is holding still and has no velocity or acceleration:

Still, by virtue of the bus's motion the driver experiences an acceleration in the ground's
reference frame of

the position of the driver within the bus is small compared to the position of the bus relative to
the centre of the turn, so this acceleration is very similar all throughout the bus. To be more
precise, we'd need to know exactly which part of the bus (e.g., the centre? front? back?) travels
the specified semicircle in this time interval going at this speed.

The bus must provide a force equal to this acceleration * the bus driver's mass to the bus driver
to hold them in their position within the bus.

c) As with the driver, except now (taking y as the direction increasing

towards the front of the bus)
So, the bandmember experiences a similar acceleration as the driver but
with a second a different acceleration term added due to their walking caused by the Coriolis

acceleration term (towards the right side of the

bus; opposite the turn direction).
In this case the Coriolis acceleration has reduce the centripetal acceleration because they're
essentially taking the turn slightly slower due to their walking.
(note that if their velocity was towards the front of the bus the Coriolis acceleration would point
to the left and increase centripetal acceleration).

Force experienced by the bandmember is similarly this total

multiplied by their mass. Slightly less than

was needed for the driver (but over a much smaller area and can be tough to balance through
with while walking!)

d)
From the perspective of the stationary reference frame on the ground, this acceleration is
provided by forces of the bus on the driver and bandmember. Within the [non-inertial] reference
frame of the bus, however, these occupants aren't accelerating, but still experience this force.
For to work within the reference frame then (i.e., to be able to say

) , it's necessary to add an additional "fictitious force" terms to the real force
terms. Specifically, for any body with mass m in the bus, we have

and therefore

where

For the driver, → centrifugal force.

This means you can still say and work totally within the non-inertial reference
frame as long as you also add an appropriate fictitious force to every object given by

2.2 17. Planar Kinetics of a Rigid Body: Force and Acceleration

This section looks at applying how force and moment are related to
2.2.1 17.1: Mass Moment of Inertia
I is a body's resistance to angular acceleration caused by moments:
(just as mass is its resistance to linear acceleration caused by forces: )

[mass] moment of inertia (MoI) is the integral of the second moment about an axis of all the

mass elements: (where the moment arm r is the distance from the axis to the mass
element)

Parallel axis theorem

If you already know the MoI about an axis through the object's mass centre, you can find it
through a parallel axis displaced by as follows:

where is the displacement from the axis through the centre of mass to the given point and
is the displacement of the of the axis to the centre of mass axis

The centre integral is zero since is a constant and integrating mass-weighted

displacement from the centre of mass over the object will give 0.

Radius of Gyration:
Sometimes texts give MoI using radius of gyration, k, where or .

For composite bodies, use superposition and the parallel axis theorem to find the total MoI.

2.2.1.1 Eg 17.4:
If the weight of member OA is 10 lb and the weight of BC is 8 lb, find
a) the MoI of this about O, and
b) the MoI of this about its centre of mass G
> restart:
MA:=10/32.2: LA:=2:
MoIOA:=1/3*MA*LA^2;
MB:=8/32.2: LB:=2*0.75:
MoIB:=1/12*MB*LB^2+MB*LA^2;
MoI:=MoIOA+MoIB;
y:=(LA/2*MA+LA*MB)/(MA+MB);
MoIOA:=1/12*MA*LA^2+MA*(y-1)^2;
MoIB:=1/12*MB*LB^2+MB*(LA-y)^2;
MoIB+MoIOA;

(in slug-ft^2)

2.2.2 17.2: Planar Kinetic Equations of Motion

We'll only consider situations where body experiences net force and motion that's all contained
in a single (fixed) reference plane (i.e., no action in the z-direction). Analysis for the general
case is in chapter 21 of Hibbeler.
Translational equation of motion for the mass centre of a rigid body:

Rotational motion:
Similarly, for a rigid body we can say : the net externally-applied moment about
the centre of mass is its mass-moment of inertia about the centre of mass times its
angular acceleration about it .
(And in general, the net externally-applied moment about any point P is its MoI about P times its
angular acceleration about it plus , the cross product of its mass-
centroid relative to P and the acceleration of point P times its total mass:

)
/*******Proof:
Consider a rigid body containing point P. Any other point in the body has a fixed distance from
point P (since the body is rigid).
The moment on mass element i in the body about point P from external and internal forces
acting on that mass element i is:

By Newton's second law in angular motion, this moment is equal to the element's rate of change
of angular momentum about point P:

(First term vanishes since , and second vanishes since the element we're considering is a
solid and not changing mass).

Writing the linear acceleration in terms of that of point P:

, so
Taking mi to dm and integrating over the object gives:

Now MP is only from external forces, since the internal ones come in pairs that are equal,
opposite, and colinear.

Notice that centre of mass is at , and the third term is the MoI about P, so
we can rewrite as follows:

If we happen to take the point P to coincide with the centre of mass, we'd get and the
simpler result
********/

Alternatively, we can rewrite in terms of IG and aG via

(parallel axis theorem)

and
(where is the location of the CoM relative to point P)
in components, we get:

So,

Therefore, adding moments about point P is equivalent to adding "kinetic moments" of

components of about P plus the "kinetic moment" of ,
2.2.3 17.3: Equations of Motion: Translation
For rectilinear translation, all particles travel along parallel straight-line paths, so

and the acceleration and net force must be parallel to the direction of motion.

About any other point A we'd get , and the kinetic moment works out to
, where d is the distance from point A to the straight-line path the
object is travelling:

(note: no need to add since is 0 for pure translation)

For curvilinear translation still have = 0, and

About any point B (other than the centre of mass) we can proceed like for rectilinear translation
and write that the net moment about B equals the kinetic moment about B:

but the cross product is not quite as easy as for pure translation; in terms of
tangential and normal kinetic force components:

we get

In either case, these alternate points to take the moment equation about can sometimes be useful
choices if the let us choose a point that most external forces are colinear with so the cross
products vanish. Then it comes down to:

2.2.4 17.4: Equations of Motion: Rotation about a Fixed Axis

Suppose a rigid body rotates in a plane about a fixed axis passing through O due to externally
applied forces and moments (couples):
Then the centre of mass experiences normal and tangential acceleration of and
,
and the net external forces and moments about the centre of mass must be equal to:

Alternatively, you can take the moment equation about a parallel axis by accounting for the
kinetic moments produced by about that point.
e.g., about point O, the normal acceleration produces no moment so we'd get

2.2.4.1 example 17.9

A 50 lb flywheel is unbalanced, and has a centre of mass .5’ to the right of its pivot point. If its
radius of gyration is kG = .6 ft about an axis through its CoM, find the horizontal and vertical
components of the reaction at O if it is released from rest.

Solution:
recall that radius of gyration is k, where or .
m = W/(32.2'/s2)
IG = m*k^2.

> restart:
k:=.6: w:=50: g:=32.2: rG:=.5:
m:=w/g; IG:=m*k^2;
omega:=0; #starting from rest
solve([
On = m*omega^2*rG,
w-Ot = m*rG*alpha,
rG*w = IG*alpha]);
2.2.4.2 Example 17.10

Solution:

> restart:
omega:=5: m:=20: L:=3: IO:=1/3*m*L^2: M1:=60: IG:=1/12*m*L^2:
at:=alpha*L/2: W:=m*9.81:
an:=omega^2*L/2:
solve([
Ot-W = m*at,
On = m*an,
-M1-L/2*W = IO*alpha]);
solve([
Ot-W = m*at,
On = m*an,
-M1-L/2*W = IG*alpha+m*L/2*at]);
solve([
Ot-W = m*at,
On = m*an,
-Ot*L/2 - M1 = IG*alpha]);
2.2.4.3 eg 17.11
This drum has a mass of 60 kg and radius of gyration . If the cord is massless and
the block has a mass of 20 kg, find the drum’s angular acceleration when the system is released
from rest.

> restart:
k:=.25: mD:=60: IG:=mD*k^2;
mB:=20: g:=9.81: r:=.4:
a:=alpha*r:
solve([
mB*g-T = mB*a,
T*r = IG*alpha]);
solve([
mB*g*r = IG*alpha+mB*a*r]);

2.2.4.4 eg 17.12 – DE!

Solution:

Here, you can first use the moment about the pivot point equation to find , then use the
tangential and normal force balance equations to determine At & An.
However, the IVP that results in time is a tough form for Maple to solve:
Attempting to solve the 2nd-order in time DE:
> restart:
IG:=1/12*m*L^2:
at:=L/2*alpha;
#assume(L, positive); assume(g, positive); assume(t, positive);
#doesn't help here
omega:=diff(theta(t),t);
alpha:=diff(omega,t);

#dsolve([at=g, theta(0)=0, D[1](theta)(0)=0]); #syntax test

an:=omega^2*r;
DE:=L/2*m*g*cos(theta(t)) = IG*alpha + m*L/2*at;
#L/2*m*g = (1/3*m*L^2)*alpha; #same eqn using IO instead of IG
dsolve([DE, theta(0)=0, D[1](theta)(0)=0]);
If you instead use and the torque equation to eliminate alpha you're left with a 1st
order IVP for , which is much easier to solve.
Method solving the first order in theta DE:
> restart:
IG:=1/12*m*L^2:
at:=L/2*alpha;
an:=(omega(theta))^2*L/2;
DE1:=L/2*m*g*cos(theta) = IG*alpha + m*L/2*at;
alpha:=solve(DE1, alpha);
DE2:= alpha = omega(theta)*D[1](omega)(theta);
dsolve([DE2, omega(0)=0]);
assign(%[1]);
An := m*g*sin(theta)+m*an;
At :=-m*g*cos(theta)+m*at;
2.2.5 17.5: Equations of Motion: General Plane Motion
Relative to an xy inertial frame, in general (i.e., even if not necessarily in pure translation or
rotation alone) a rigid body will still obey

and we could instead sum the moments about another point, as before:

One good point in rolling problems is the IC, the instantaneous centre of zero velocity (i.e., the
contact point); about this point,
(just like it did when summing moments about a pivot point).

Solution method:
1. FBD
a. Set up an xy inertial frame and draw the FBD
b. Specify the direction & sense of acceleration of the mass centre aG and angular
acceleration a of the body
c. Determine IG
d. If you want to use then try drawing the kinetic diagram to
help visualize the "moments" developed by the components and .
2. Equations of motion
a. Apply 3 eqns of motion
b. If friction, consider motion with no slipping / tipping and see if it works;
otherwise, slipping / tipping happens.
3. Kinematics
a. If equations of motion can't totally solve the problem, use kinematics
b. If body's motion is constrained to its supports, you can get extra equations by
using relating the accelerations of any two points A and B on the
body.
c. When something rolls on a flat surface without slipping then .
2.2.5.1 Eg 17.13
Find the angular acceleration of the spool if it has a mass of 8 kg and RoG of :

Soln:

> restart:
m:=8: k:=.35: IG:=k^2*m; g:=9.81:
solve([
m*aGy=100+T-m*g,
IG*alpha=.5*T-.2*100,
aGy=-.5*alpha]);
> restart:
m:=8: k:=.35: IG:=k^2*m; g:=9.81:
IIC:=m*.5^2+IG;
solve(.5*78.48-(.5+.2)*100=IIC*alpha);

2.3 2CD4 Tips for this Topic

In many coupled-body motion problems there are known external forces and constrained paths
the bodies can follow from rails, walls, or connections between them which mean knowing the
position of one body is enough to know the position of all of the other bodies. In these situations
you can usually solve the problem with the following approach:
1. Use geometry to relate position & angles (making use of topic 2.1 above)
a. differentiate (using chain rule) to then relate accelerations;
b. algebra to express all accelerations in terms of a single one (and any positions and
velocities you like)
2. Use N2 to relate forces & accelerations (making use of topic 2.2 above)
a. algebra to eliminate internal forces and find explicit eqn for one acceleration in
terms of external forces, constants, time, and any number of internal positions &
velocities.
b. Solve this DE with FlexPDE to find desired positions, velocities, and time at
desired endtime (by some fixed time or end condition)

Note that if the DE is simple enough (i.e., linear) Maple can likely solve it easily as well, and
may not require putting it in explicit form like FlexPDE requires. In any case, it can help to have
Maple assist with the algebra involved in coming up with the DE.