ME 240

Dynamic and Control of Mechanical Systems

Lecture-2

G R Jayanth, [email protected]
 Acceleration of a point on a rigid body
• The acceleration of a point P on a rigid body, given the acceleration of another point Q is given by𝑷ሷ = 𝑸ሷ + 𝑸𝑷
ሷ
• Let us again focus on relative motion and assume that the body rotates about Q, and that the origin of the coordinate
system is situated at Q, so that 𝑄 = 0 and 𝑸𝑷ሷ = 𝑷ሷ
P
d 𝝎× 𝑷
QP
• Noting that 𝑷ሷ = d(𝑷)/dt
ሶ , we see that 𝑷ሷ = = 𝝎ሶ × 𝑷+ 𝝎 × 𝑷ሶ = 𝝎ሶ × 𝑷+ 𝝎 × (𝝎 × 𝑷)
dt

• The first term is due to the angular acceleration and is perpendicular to the radius vector 𝑷 and the angular acceleration Q
vector.
S
• The second term is the centripetal acceleration term. Notice that in general, the centripetal term is not directed along the
radius vector 𝑷: this happens only if 𝝎 and 𝑷 are perpendicular to one another.
• As effortless as this derivation has been, by simply following the rules of differentiation, they have been done in a very
mechanical, non-intuitive manner. We can “breathe” some life into them by deriving the same from first principles,
𝒗𝑷 (𝑡 + ∆𝑡)
particularly for rotation about the Z-axis.
• Consider the velocity vector 𝒗𝑃 = 𝑷ሶ at time t. After some time ∆𝑡, the vector would have increased in length by |∆𝒗𝑃 | =
𝛼|𝑷|∆𝑡, where 𝛼 = |𝝎|. ሶ Since the rigid body is also rotating about A, the new would be tangential to 𝑷(𝑡 + ∆𝑡), which
would be the rotated with respect to 𝑷(𝑡) by ∆𝜃 = 𝜔∆𝑡. vector 𝑷(𝑡 + ∆𝑡)
𝒗𝑷 (𝑡)
• Hence the net change in the velocity vector In the tangential direction is given by ∆𝒗𝑃𝑡 = |𝒗𝑷 𝑡 + ∆𝑡 |cos(∆𝜃/2)-
|𝒗𝑷 𝑡 |cos(∆𝜃/2) ≈ |𝒗𝑷 𝑡 + ∆𝑡 |- |𝒗𝑷 𝑡 | = 𝛼|𝑷|∆𝑡
∆𝒗𝑃𝑡 𝑡𝑎𝑛𝑔𝑒𝑛𝑡𝑖𝑎𝑙
• Thus the tangential acceleration is given by𝑎𝑡 = = 𝛼|𝑷|. This corresponds to the first term 𝝎ሶ × 𝑷, in the more general
∆𝑡
case
∆𝜃 ∆𝜃
• Likewise, in the normal direction the net change in velocity is ∆𝒗𝑃𝑛 = −[|𝒗𝑷 𝑡 + ∆𝑡 | sin + |𝒗𝑷 𝑡 sin ]
2 2 𝒗𝑷 (𝑡 + ∆𝑡)
• In other words, ∆𝒗𝑃𝑛 ≈ −[|𝒗𝑷 𝑡 + ∆𝑡 +|𝒗𝑷 𝑡 ]sin
∆𝜃
≈ −𝟐|𝒗𝑷 𝑡 | sin
∆𝜃
≈ -|𝒗𝑷 𝑡 | ∆𝜃=-𝜔2 |𝑷| ∆𝑡
𝒗𝑷 (𝑡)
2 2

• Thus, the normal acceleration is given by ∆𝒗𝑃𝑛 / ∆𝑡= -𝜔2 |𝑷| which is the familiar expression for centripetal acceleration for 𝑛𝑜𝑟𝑚𝑎𝑙
the case of planar motion. In the more general case, this corresponds to the second term 𝝎 × (𝝎 × 𝑷)
 YB
Velocity of a particle moving on a rigid body
• Next, let us consider the case when a particle located at R moves with respect to the rigid body. To measure
YA
this relative motion, we shall define a new coordinate system (XB-YB) that is coincident with Q and is fixed to R
the rigid body (and therefore rotates and translates along with it). rAC XB
• It is well worth getting used to having multiple coordinate frames because we would find later that it is very
convenient to describe the motion of an interconnection of rigid bodies using such frames.
Q
• Let the particle have a position 𝑹=𝑅𝑥 𝒙 ෝ𝑩 + 𝑅𝑦 𝒚 ෝ𝑩 + 𝑅𝑧 𝒛ො 𝑩 , with respect to A in this coordinate system. Then
ሶ 𝑅ሶ 𝑥 𝒙
𝑹= 𝒙𝑩ሶ ) + 𝑅ሶ 𝑦 𝒚
ෝ𝑩 + 𝑅𝑥 (ෝ 𝒚𝑩ሶ ) + 𝑅ሶ 𝑧 𝒛ො 𝑩 + 𝑅𝑧 (ො𝒛𝑩ሶ )
ෝ𝑩 + 𝑅𝑦 (ෝ
B
• Grouping the terms we have,𝑹=𝒗 𝒙𝑩ሶ ) + 𝑅𝑦 (ෝ
ሶ 𝑟𝑒𝑙 + 𝑅𝑥 (ෝ 𝒚𝑩ሶ ) + 𝑅𝑧 (ො𝒛𝑩ሶ ), where 𝒗𝑟𝑒𝑙 = 𝑅ሶ 𝑥 𝒙
ෝ𝑩 + 𝑅ሶ 𝑦 𝒚
ෝ𝑩 + 𝑅ሶ 𝑧 𝒛ො 𝑩 XA
represents the rate of change of coordinates of the particle in the coordinate system fixed to the rigid body,
but moving with respect to the ground.
• We note that since the new coordinate system is rotating, so do its unit vectors. Hence we would have
ෝ𝑩ሶ = 𝝎 × 𝒙
𝒙 ෝ𝑩ሶ = 𝝎 × 𝒚
ෝ𝑩 , 𝒚 ෝ𝑩 , 𝒛ො 𝑩ሶ = 𝝎 × 𝒛ො 𝑩

• ሶ 𝑟𝑒𝑙 + 𝝎 × (𝑅𝑥 𝒙
Thus, we would have 𝑹=𝒗 ෝ𝑩 + 𝑅𝑦 𝒚
ෝ𝑩 + 𝑅𝑧 𝒛ො 𝑩 )=𝒗𝑟𝑒𝑙 + 𝝎 × 𝑹

• Note that 𝑹ሶ is being measured in the XB-YB coordinate system and not in the XA-YA coordinate system. For
the measurements to be made in the XA-YA coordinate system, we need to assume that the XB-YB coordinate
system, which is rotating with the rigid body, is instantaneously coincident with the X A-YA coordinate
system, in which case, their unit vectors would be identical at just this instant. In such a case, the absolute
velocity of the point R can be written as 𝒗𝑹 = 𝒗𝑸 + 𝝎 × 𝑹 + 𝒗𝑟𝑒𝑙

• Alternately, we need to know the relative orientation of the XB-YB coordinate system with respect to the XA-
YA coordinate system, so that vectors in one coordinate system can be represented correctly in another.
• In either case, the central point is that we need to be consistent in representing all vectors in the same
coordinate system (either XA-YA or XB-YB).
Acceleration of a particle moving on a rigid body
ሶ 𝑟𝑒𝑙 + 𝝎 × 𝑹.
• Next, let us compute the acceleration of the same particle. To do this, we start with the equation 𝑹=𝒗
ሷ 𝒗ሶ 𝑟𝑒𝑙 + 𝑑(𝝎 × 𝑹)/𝑑𝑡
• Differentiating this, we get 𝑹=
𝑑 𝝎×𝑹
• Next, we have ሶ 𝝎ሶ × 𝑹+ 𝝎 ×(𝒗𝑟𝑒𝑙 + 𝝎 × 𝑹) = 𝝎ሶ × 𝑹 + 𝝎 × (𝝎 × 𝑹)+𝝎 × 𝒗𝑟𝑒𝑙
= 𝝎ሶ × 𝑹 + 𝝎 × 𝑹=
𝑑𝑡
𝑑 𝑅ሶ 𝑥 ෝ
𝒙𝑩 +𝑅ሶ 𝑦 𝒚
ෝ𝑩 +𝑅ሶ 𝑧 𝒛ො 𝑩
• Likewise, we have 𝒗ሶ 𝑟𝑒𝑙 = = 𝑅ሷ 𝑥 ෝ
𝒙𝑩 + 𝑅ሷ 𝑦 𝒚 𝒙𝑩ሶ ) + 𝑅ሶ 𝑦 (ෝ
ෝ𝑩 + 𝑅ሷ 𝑧 𝒛ො 𝑩 +𝑅ሶ 𝑥 (ෝ 𝒚𝑩ሶ ) + 𝑅ሶ 𝑧 (ො𝒛𝑩ሶ )
𝑑𝑡

𝒙𝑩ሶ = 𝝎 × ෝ
• We note again that since the new coordinate system is rotating, so do its unit vectors. Hence we would have ෝ ෝ𝑩ሶ = 𝝎 × 𝒚
𝒙𝑩 , 𝒚 ෝ𝑩 , 𝒛ො 𝑩ሶ = 𝝎 × 𝒛ො 𝑩 ,
𝒙𝑩ሶ ) + 𝑅ሶ 𝑦 (ෝ
so that 𝑅ሶ 𝑥 (ෝ 𝒚𝑩ሶ ) + 𝑅ሶ 𝑧 (ො𝒛𝑩ሶ )= 𝝎 × (𝑅ሶ 𝑥 ෝ
𝒙𝑩 + 𝑅ሶ 𝑦 𝒚
ෝ𝑩 + 𝑅ሶ 𝑧 𝒛ො 𝑩 )
• So, we would have 𝒗ሶ 𝑟𝑒𝑙 = 𝒂𝑟𝑒𝑙 +𝝎 × 𝒗𝑟𝑒𝑙 , where 𝒂𝑟𝑒𝑙 =𝑅ሷ 𝑥 ෝ
𝒙𝑩 + 𝑅ሷ 𝑦 𝒚
ෝ𝑩 + 𝑅ሷ 𝑧 𝒛ො 𝑩
• Combining both the terms, we have 𝑹ሷ = 𝝎ሶ × 𝑹 + 𝝎 × (𝝎 × 𝑹)+ 𝟐𝝎 × 𝒗𝑟𝑒𝑙 + 𝒂𝑟𝑒𝑙
• So, the overall acceleration of the point C is given by 𝑹ሷ = 𝑸ሷ + 𝑸𝑹
ሷ

• The first term in 𝑹ሷ is due to angular acceleration. The second term is the centripetal acceleration. The third term is called the “coriolis” acceleration.
The fourth term is the acceleration measured in the (XB-YB) coordinate system.
• Coriolis acceleration applies a pseudo force on any object moving in a rotating coordinate system, called the Coriolis force, analogous to centrifugal
force caused by centripetal acceleration. Note that unlike centrifugal force, this force is independent of the distance from the center of rotation, but is
dependent on the relative-velocity of motion.
• Coriolis force can have some dramatic consequences on our life. Note that earth is a rotating body. So movement of both air or water causes them to
experience Coriolis force. Cyclones are due to Coriolis effect. The trade winds are due to Coriolis effect and these end up transporting the dust from
Sahara desert in Africa to the Amazon rain forest in South America and fertilize the rain forest. Likewise, rivers erode one bank more than the opposite
bank, depending on their direction of flow, due to a larger Coriolis force on one side than on the other. A pendulum that is supposes to swing in a plane
will, due to Coriolis force, experience an out-of-plane force as well, as a consequence of which, the plane of oscillation rotates at the same rate as that
of earth’s rotation. This famous experiment, called Foucault’s pendulum experiment, unequivocally established that the earth was spinning about itself,
contrary to the illusion that the sun and other celestial bodies are spinning around earth. Coriolis force is also employed by insects for guidance and
navigation. An organ called “halteres” is used by insects to determine, and subsequently control their angular velocity.
 First principles derivation of the Coriolis acceleration
Q
• As with the previous case, the expression for Coriolis acceleration, was derived effortlessly by following
the rules of differentiation of cross-products, but equally, without much physical insight. dθ
• For the planar case, it is again possible to derive the Coriolis acceleration from first principles.
• For the sake of simplicity, we shall assume that 𝒂𝑟𝑒𝑙 = 0, and we shall also consider that the velocity is
tangential to the radius vector at this instant. Though this reduces the generality, we can readily extend
the analysis for the more general case once we understand the logic.
dθ’
r r’
• Consider the figure on the right. At the present time, the relative velocity is aligned with 𝜔𝑟. At a time dt v+ωr’
later, the rotating stage would be at an position dθ = 𝜔𝑑𝑡. Furthermore, in the same time dt, if we were
sitting on the rotating disk, we would notice that the particle would have moved along the direction of R’
velocity by dx = 𝑣𝑟𝑒𝑙 𝑑𝑡. To obtain the actual position at a time dt later, we rotate this position R’ forward
by dθ. Therefore, the total velocity of the object at a time dt later is shown below: vrel+ωr
′ ′
• We see that the net change in velocity is given by 𝜔𝑟 𝑑𝜃 + 𝑑𝜃 + 𝑣𝑟𝑒𝑙 𝑑𝜃
dx
• Noting that dθ = 𝜔𝑑𝑡, and 𝑑𝜃 ′ = = 𝑣𝑟𝑒𝑙 𝑑𝑡/𝑟′, we get the net change in velocity to be 𝜔2 𝑟 ′ 𝑑𝑡 +
r′
2𝑣𝑟𝑒𝑙 𝜔𝑑𝑡.
• Thus the net acceleration is 𝜔2 𝑟 ′ + 2𝑣𝑟𝑒𝑙 𝜔 ≈ 𝜔2 𝑟 + 2𝑣𝑟𝑒𝑙 𝜔
• The first term is the centripetal acceleration. The second term is the Coriolis acceleration.
Reference

J. L. Meriam, and L. G. Kraige, Engineering Mechanics: Dynamics, Wiley, 2013/7th Edition