Lecture 29
Lecture 29
Chasles Theorem: The most general displacement of a rigid body is a rotation about some axis plus a translation.
It ought to be possible to divide the problem into 2 separate problems: 1. Translation, 2. Rotation. Obviously if one point is fixed, this separation is trivially true. Also possible for general problem in some cases. Ch. 4 discussion: 6 generalized coordinates are needed to describe rigid body motion: 3 Cartesian coordinates of a fixed point in the body + 3 Euler angles describing the body rotation about an axis through the fixed point.
In general: Equations of motion for all 6 coordinates do not separate into 3 eqtns for translation + 3 for rotation!!
Most common special case where fixed point in rigid body Center of Mass (CM):
Can use some Ch. 1 results for angular momentum & kinetic energy which simplify the problem & separate equations of motion for translational & rotational coordinates into 2 sets of 3, so translation & rotation can be treated separately!
O = arbitrary origin (fixed axes), R = CM position with respect to O. V = CM velocity, M = CM mass. ri = position of particle i with respect to CM (body axes) pi = mivi = momentum of same particle in this system.
A Ch. 1 result for the angular momentum of a system of particles: L = R MV + i[ri pi] (1) The total angular momentum about point O = the angular momentum of the motion of the CM about O + the angular momentum of the motion of particles about the CM
L = R MV + i[ri pi] (1) Total angular momentum about O = angular momentum of the CM about O + angular momentum of particles about the CM
(1) In general, L depends on the origin O, through vector R. Only if the CM is at rest with respect to O, will first term in (1) vanish. Then & only then will L be indep of the point of reference. Then & only then L = angular momentum about the CM. If we choose the origin of the body axes to be the CM, then clearly, as far as L is concerned, we can treat the translation & rotation of the body separately. Then, the 1st term in (1) depends only on (X,Y,Z) & the 2nd term depends on (,,) : L = R MV + L(,,)
O = arbitrary origin (fixed axes), R = CM position with respect to O. V = CM velocity, M = CM mass. ri = position of particle i with respect to CM (body axes) vi = velocity of same particle in this system.
The total Kinetic Energy of a many particle system is equal to the Kinetic Energy of the CM plus the Kinetic Energy of motion about the CM.
T = ()MV2 + ()i mi(vi)2 (2) Total KE with respect to O = KE of the CM+ KE of motion about the CM.
(2) In general, T depends on origin O, through vector V. Only if the CM is at rest with respect to O, will first term in (2) vanish. Then & only then will T be indep of point of reference. Then & only then T = KE about the CM. If choose origin of body axes to be the CM, then clearly, as far as T is concerned, we can treat the translation of body & rotation of body the separately. Then, the 1st term in (2) depends only on (X,Y,Z), 2nd term depends on (,,): T = ()MV2 + T(,,)
We can make this separation for the angular momentum L & the KE T. To do dynamics, using say, the Lagrangian method, we would like to make this same kind of separation in Lagrangian L = T -V
Obviously, we need to be able to separate the PE V in the same way. We dont have a theorem which does this in general for V, like we do for L & T! However, from experience, in a large number of problems (special cases) of practical interest, such a separation is possible.
PE in a uniform gravitational field, PE of a magnetic dipole in uniform magnetic field, others,... In such cases, we can write: L = L (X,Y,Z) + L(,,)
In cases where: L = L (X,Y,Z) + L(,,) Lagranges Eqtns of motion for translation (X,Y,Z) obviously separate from Lagranges Eqtns of motion for rotation (,,) & we can treat the translation & rotation as independent. Now, work on expressions for angular momentum L & KE T for motion about some fixed point in a rigid body. Work on 2nd terms in previous expressions for these.
Called L & T in Eqtns (1) & (2) above. Drop prime notation.
Make repeated use of the Ch. 4 result relating the time derivatives in space & body axes (angular velocity ):
(d/dt)s = (d/dt)r +
(I)
Goldstein proves that angular velocity is indep of choice of the origin of body axes. See pages 185-186
for details. Intuitively obvious by definition of rigid body? Angular momentum L of rigid body about a fixed point in the body. Dropping the primes of before on positions, momenta, & velocities: L = i[ri pi], pi = mivi
Follow text & write using summation convention: L = mi[ri vi] (1)
ri = fixed with respect to body axes (definition of rigid!) The only contribution to vi is from the body rotation. Use (I) vi = (dri/dt)s = (dri/dt)r + ri = 0 + ri
(2)
Use the vector identity for double cross product: L = mi[(ri)2 - ri(ri)] (3) Look at the x, y, & z components:
Lx = xmi[(ri)2 - (xi)2] - ymixiyi - zmixizi Ly = ymi[(ri)2 - (yi)2] - xmiyixi - zmiyizi Lx = zmi[(ri)2 - (zi)2] - ymiziyi - xmizixi (4) (4) Each component of L = Linear combination of components of angular velocity . Or: The angular momentum vector is related to the angular velocity vector by a linear transformation!
Inertia tensor Ijk (summation convention, i = particle label, j, k = Cartesian x,y,z = 1,2,3 labels!) :
Or:
Recall: The coordinate system is the BODY axis system (the primes from the last chapter are dropped). Since the body is rigid, all elements Ijk are constants in time. Clearly, they DO depend on the origin of coordinates (different if taken about the CM or elsewhere!). We can summarize Lj = Ijkk in an even more compact form: L I where I Inertia Matrix This is in the form of the orthogonal transformation of Ch. 4. Clearly we need to take the interpretation that I acts on the vector to produce L, rather than the interpretation of I acting on the coordinate system.
Now, well derive a form for the KE T for a rigid body in terms of I and . Start with KE of motion about a point in form (summation convention, dropped prime from earlier) : T ()mi(vi)2
T ()mi(vi)2 = ()mivivi
Rigid body The only contribution to vi is from body rotation. As before use vi = (dri/dt)s = (dri/dt)r + ri = 0 + ri Convenient to use with only one of the vi in the dot product: T = ()mivi( ri ) Permuting vectors in triple product: T = ()[mi(ri vi )] ()L Using L = I this becomes:
T = ()I
Note: Order matters & product with when is to left of I & to right of I is clearly different. T is clearly a scalar, as it should be! Often, I may write it as simply T = ()I (leaving out )
T = ()I
n unit vector along rotation axis, so that = n Can write (1) as:
(1)
(3)
T = ()I
n unit vector along rotation axis:
(1)
T = ()2nIn ()I2 (2) I nIn Moment of Inertia (about rotation axis) Clearly, moment of inertia I depends on the direction n of the rotation axis. Also, for a rigid body in motion, that direction & thus the direction of can be time dependent. In general, I = I(t)
In the (important!) special case where the body is constrained to rotate about a fixed axis, clearly then, I = constant. When this is the case, T in the form of (1) can be used in the Lagrange formalism if we can write as the time derivative of some angle.
Moment of inertia about axis a: Ia = imi(R n)2 + imi(ri n)2 + 2 imi(R n)(ri n) Moment of inertia about axis b: Ib = imi(ri n)2 Note that: imi(R n)2 M(R n)2 Ia = Ib + M(R n)2 + 2imi(R n)(ri n) Rewrite 3rd term as: 2(R n)(imiri n) Note that, by definition of CM, imiri = 0 2imi(R n)(ri n) = 0 Ia = Ib + M(R n)2 For = angle between R & rotation axis n, distance between axes a & b = |R n| = Rsin = r. Ia = Ib + MR2sin2
Or:
Summary: Rotational KE in terms of inertia tensor: T = ()I (1) Can rewrite in terms of tensor elements as (summation convention): T = ()jIjkk (2) Again (summation convention, i = particle label, j, k = Cartesian x,y,z = 1,2,3 labels!) : Ijk mi[(ri)2jk - (ri)j(ri)k] Continuous bodies: Sum over particles integral over
volume V. Define: (r) Mass density at position r