The angular momentum of an object about an arbitrary point $O$ is given by, $$ \mathbf{H_O}=[I_O]\mathbf{\omega}$$ where $[I_O]$ is the moment of inertia tensor about $O$. And for the center of mass $G$, $$ \mathbf{H_G}=[I_G]\mathbf{\omega} $$

My book says that a new array of moments and products of inertia would be obtained if a different system of axes were used. The transformation characterized by this new array, however, would still be the same. The angular momentum $\mathbf{H_G}$ about the mass center corresponding to a given angular velocity $\mathbf{\omega}$ is independent of the choice of the coordinate axes.

So if I setup the axes with origin at $O$, according to the first equation $\omega $ would transform into $\mathbf{H_O}$ and according to the book that would equal $\mathbf{H_G}$. But if that is the case then won't the angular momentum about any arbitary point be equal to the angular momentum about the center of mass?

  • $\begingroup$ I've removed a number of comments that were attempting to answer the question and/or responses to them. Please keep in mind that comments should be used for suggesting improvements and requesting clarification on the question, not for answering. $\endgroup$
    – David Z
    Jul 12 '20 at 23:10

The first equation is wrong. Angular momentum of a rigid body is first defined at the center of mass G as

$$ \boldsymbol{H}_G = \mathbf{I}_G \boldsymbol{\omega} \tag{1}$$

and then transferred to another point O with the following law

$$ \boldsymbol{H}_O = \boldsymbol{H}_G + \boldsymbol{r}_{G/O} \times \boldsymbol{p} \tag{2} $$where $\boldsymbol{p}$ is linear momentum, and $\boldsymbol{r}_{G/O}$ is the location of the center of mass G with respect to the reference point O.


Consider a rigid body as a system of particles moving together. The total mass of the body is $m = \sum_i^n m_i$ and the position of each particle as measured from the reference point O is decomposed into two parts, one the location of the center of mass G and then the relative location of the particle to the center of mass.

$$ \boldsymbol{r}_i = \boldsymbol{r}_{G} + \boldsymbol{d}_i \tag{3} $$

For brevity I drop the O subscripts, and thus $\boldsymbol{r}_G = \boldsymbol{r}_{G/O}$ and later $\boldsymbol{v}_G = \tfrac{\rm d}{{\rm d}t} \boldsymbol{r}_{G/O}$.

The center of mass is defined as the location which makes $\sum_i^n m_i \boldsymbol{d}_i=0$ and thus

$$ \sum_i^n m_i \boldsymbol{r}_i = \boldsymbol{r}_G \sum_i^n m_i + \sum_i^n m_i \boldsymbol{d}_i = m\, \boldsymbol{r}_G \tag{4}$$

The time derivative of each position is velocity and therefore the total momentum of the rigid body is derived from the time derivative of (4)

$$ \boldsymbol{p} = \sum_i^n m_i \boldsymbol{v}_i = m\,\boldsymbol{v}_G \tag{5}$$

There is a little trickery in the above expression because the time derivative of (3) is

$$ \boldsymbol{v}_i = \boldsymbol{v}_G + \boldsymbol{\omega} \times \boldsymbol{d}_i \tag{6} $$

which means that $\require{cancel} \boldsymbol{p} = \sum_i^n m_i \boldsymbol{v}_i = (\sum_i^n m_i) \boldsymbol{v}_G + \boldsymbol{\omega} \times (\cancel{\sum_i^n m_i \boldsymbol{d}_i}) = m \boldsymbol{v}_G $

The same trickery is to be applied for angular velocity about the reference point

$$ \require{cancel} \begin{aligned}\boldsymbol{H}_O & =\sum_{i}^{n}\boldsymbol{r}_{i}\times(m_{i}\boldsymbol{v}_{i})=\sum_{i}^{n}(\boldsymbol{r}_{G}+\boldsymbol{d}_{i})\times m_{i}(\boldsymbol{v}_{G}+\boldsymbol{\omega}\times\boldsymbol{d}_{i})\\ & =\sum_{i}^{n}\left(\boldsymbol{r}_{G}\times m_{i}(\boldsymbol{v}_{G}+\boldsymbol{\omega}\times\boldsymbol{d}_{i})+\boldsymbol{d}_{i}\times m_{i}(\boldsymbol{v}_{G}+\boldsymbol{\omega}\times\boldsymbol{d}_{i})\right)\\ & =\boldsymbol{r}_{G}\times(\sum_{i}^{n}m_{i})\boldsymbol{v}_{G}+\boldsymbol{r}_{G}\times\left(\boldsymbol{\omega}\times\cancel{\sum_{i}^{n}m_{i}\boldsymbol{d}_{i}}\right)\\ & +\left(\cancel{\sum_{i}^{n}m_{i}\boldsymbol{d}_{i}}\right)\times\boldsymbol{v}_{G}+\sum_{i}^{n}\boldsymbol{d}_{i}\times m_{i}(\boldsymbol{\omega}\times\boldsymbol{d}_{i})\\ & =\boldsymbol{r}_{G}\times\boldsymbol{p}+\mathbf{I}_{G}\,\boldsymbol{\omega} = \boldsymbol{r}_{G}\times\boldsymbol{p} + \boldsymbol{H}_G\;\;\; \checkmark \end{aligned} \tag{7}$$

Given the definition $$ \boldsymbol{H}_G = \mathbf{I}_{G}\,\boldsymbol{\omega}=\sum_{i}^{n}m_{i}\left(-\boldsymbol{d}_{i}\times(\boldsymbol{d}_{i}\times\boldsymbol{\omega})\right) \tag{8} $$

or expanded out

$$ \mathbf{I}_G = \sum_i^n m_i \begin{bmatrix} y_i^2+z_i^2 & -x_i y_i & -x_i z_i \\ -x_i y_i & x_i^2+z_i^2 & -y_i z_i \\ -x_i z_i & -y_i z_i & x_i^2+y_i^2 \end{bmatrix} \tag{9} $$ where $\boldsymbol{d}_i = \pmatrix{x_i \\ y_i \\z_i}$.

Now that angular momentum at the center of mass $\boldsymbol{H}_G$, mass moment of inertia at the center of mass $\mathbf{I}_G$ and angular momentum at the reference point $\boldsymbol{H}_O$ is defined, we can derive the rule of transferring the mass moment of inertia tensor to the reference point, such that $$ \boldsymbol{H}_O = \underbrace{ \left( \mathbf{I}_G - m [\boldsymbol{r}_G \times] [\boldsymbol{r}_G\times] \right) }_{\mathbf{I}_O} \boldsymbol{\omega} = \mathbf{I}_O\, \boldsymbol{\omega} \tag{10}$$


Consider the body rotating about the reference point, setting the kinematics of the center of mass as $\boldsymbol{v}_G = \boldsymbol{\omega} \times \boldsymbol{r}_G$ and thus momentum as $$\boldsymbol{p} = m ( \boldsymbol{\omega} \times \boldsymbol{r}_G) = m (-[\boldsymbol{r}_G \times] \boldsymbol{\omega}) \tag{11}$$ where $[\boldsymbol{r}\times] = \begin{bmatrix}0 & -z & y \\ z & 0 & -x \\ -y & x &0 \end{bmatrix}$ is the 3×3 cross product matrix operator.

Now use equations (1) and (2) to define angular momentum about the reference point, and factor out the rotational velocity vector to create a vector = matrix * vector equation.

$$ \begin{aligned} \boldsymbol{H}_O & = [\mathbf{I}_G] \boldsymbol{\omega} + [\boldsymbol{r}_G \times] \boldsymbol{p} \\ & =[\mathbf{I}_G] \boldsymbol{\omega} + [\boldsymbol{r}_G\times] (-m [\boldsymbol{r}_G \times] \boldsymbol{\omega} ) \\ & = \left( [\mathbf{I}_G] - m [\boldsymbol{r}_G \times] [\boldsymbol{r}_G\times] \right) \boldsymbol{\omega} \;\;\checkmark \end{aligned} \tag{12} $$

Parallel Axis Theorem

So the transfer of mass moment of inertia to a different location is described by the so-called parallel axis theorem, which in vector form is

$$ [\mathbf{I}_O] = [\mathbf{I}_G] - m [\boldsymbol{r}_G \times][\boldsymbol{r}_G \times] \tag{13} $$

which in turn allows us to find angular momentum at the reference point as just simply

$$ \boldsymbol{H}_O = [\mathbf{I}_O]\, \boldsymbol{\omega} \tag{14} $$

  • $\begingroup$ I thought that the equation that you wrote could be used but is not how $\mathbf{H_O}$ is defined, as here on page 3 they have written the first equation. $\endgroup$ Jul 13 '20 at 1:00
  • $\begingroup$ I had made some notation errors in my answer. I edit some fixes. The above is the parallel axis theorem $\boldsymbol{H}_O = \mathbf{I}_O \boldsymbol{\omega}$ if the center of mass is rotating about the reference axis, and thus $\boldsymbol{p} = m \boldsymbol{v}_G = m ( \boldsymbol{\omega} \times \boldsymbol{r}_G)$ or $$\boldsymbol{H}_O = \boldsymbol{H}_G + m (- \boldsymbol{r}_G \times( \boldsymbol{r}_G \times \boldsymbol{\omega})) $$ and $$ \mathbf{I}_O = \mathbf{I}_G - m [\boldsymbol{r}_G \times] [\boldsymbol{r}_G\times] $$ $\endgroup$ Jul 13 '20 at 1:53
  • $\begingroup$ So $\mathbf{H_O}=\mathbf{I_O}\mathbf{\omega}$ is not always true? $\endgroup$ Jul 15 '20 at 6:41
  • $\begingroup$ It is true, as the definition of $\mathbf{I}_O$ using the parallel axis theorem. Consider $$ \boldsymbol{H}_O = \mathbf{I}_G \boldsymbol{\omega} + m (- \boldsymbol{r}_G \times( \boldsymbol{r}_G \times \boldsymbol{\omega})) = \mathbf{I}_O \boldsymbol{\omega} $$ where $$ \boldsymbol{H}_O = \underbrace{ \left( \mathbf{I}_G - m [\boldsymbol{r}_G \times] [\boldsymbol{r}_G\times] \right) }_{\mathbf{I}_O} \boldsymbol{\omega}$$ $\endgroup$ Jul 15 '20 at 11:50
  • $\begingroup$ @AkshatSharma, I updated the answer with more details on how the transfer law for rot. momentum leads to the transfer law for mass moment of inertia. $\endgroup$ Jul 15 '20 at 12:14

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.