Main Portal UC3M 251-14169
Pattern, table Theorem, definition Derivation Algorithm Exam trap Prerequisite
Engineering Mechanics — Formula SheetUC3M Aerospace Engineering · Topics 1 to 5

How to solve a mechanics problem

Algorithm The lecturer's order

  1. Frames, bases and coordinates. Fixed frame $S_0 = \{O;\ \mathcal B_0\}$, moving frames $S_1 = \{A;\ \mathcal B_1\}, \dots$, intrinsic basis $\{\mathbf e_t, \mathbf e_n, \mathbf e_b\}$ if useful. Count degrees of freedom, write the constraints, list the parameters with SI units.
  2. Change of basis and position vector. Deduce every $[{}_0R_1]$ from a drawing of each rotation (axis out of the paper ⊙ or into it ⊗); build $\mathbf r_0^P$ as a chain of vectors, e.g. $\overrightarrow{OC} + \overrightarrow{CP}$.
  3. Kinematics. Velocity and acceleration: chain rule, product rule, $\boldsymbol\omega \times$ for rotating unit vectors, composition of motions and Coriolis.
  4. Forces and equations of motion. Free-body diagram: applied forces, constraint reactions (normal to the constraint), inertia forces if the frame is not inertial; project $m\mathbf a = \sum\mathbf F$ on the chosen basis. Tangent projections give the motion, normal projections the reactions.
  5. Answers and checks. Answer every part; limiting cases ($\theta = 0$, $\Omega = 0$), dimensions, $\det R = 1$, orthogonality, energy if it is conserved.

Pattern Notation of the course

  • $\mathbf r_0^P$, $\mathbf v_0^P$, $\mathbf a_0^P$: position, velocity, acceleration of the point $P$ (superscript) seen from $S_0$ (subscript).
  • $\left.\dfrac{d\mathbf b}{dt}\right|_0$: derivative of a vector seen from $S_0$; it differs from $\left.\dfrac{d\mathbf b}{dt}\right|_1$ when $S_1$ rotates.
  • $[{}_0R_1]$: change-of-basis matrix; its columns are $\mathbf i_1, \mathbf j_1, \mathbf k_1$ written in $\mathcal B_0$, and $[\mathbf b]_0 = [{}_0R_1][\mathbf b]_1$ [Notes Eqs. 1.12–1.13].
  • $\boldsymbol\omega_{10}$, $\boldsymbol\alpha_{10}$: angular velocity and acceleration of $S_1$ relative to $S_0$.
  • $T_0$ kinetic energy, $V_0$ potential, $E_0 = T_0 + V_0$, all measured in $S_0$.

Trap Mistakes that cost the most marks

  • Differentiating a unit vector of a moving basis as if it were constant: $\dot{\mathbf e}_R = \dot\theta\,\mathbf e_\theta \neq \mathbf 0$.
  • A sign in $[{}_0R_1]$ guessed instead of deduced: draw the view along the rotation axis and apply the right-hand rule.
  • Writing a reaction with a tangent component on a smooth constraint (it has none), or forgetting that friction is applied force when sliding.
  • Adding inertia forces when working in the inertial frame $S_0$ (they exist only in the equation written in $S_1$).
  • Missing the factor 2 of Coriolis, or the minus sign of the centripetal term $-R\dot\theta^2\mathbf e_R$.

Topic 1 — Fundamentals and particle kinematics

Theorem Rotation matrices

  • Orthogonal, proper: $[{}_1R_0] = [{}_0R_1]^{-1} = [{}_0R_1]^T$, $\det[{}_0R_1] = +1$ [Notes Eqs. 1.14–1.15].
  • Composition: $[{}_0R_2] = [{}_0R_1][{}_1R_2]$, not commutative [Notes Eq. 1.16].
  • Axis and angle: eigenvalues $1, e^{\pm i\beta}$; the axis is the eigenvector of $\lambda = 1$ and $\operatorname{tr}R = 1 + 2\cos\beta$ [Slide 6].
$$[{}_0R_1(z,\theta)] = \begin{pmatrix}\cos\theta & -\sin\theta & 0 \\ \sin\theta & \cos\theta & 0 \\ 0 & 0 & 1\end{pmatrix}$$

About $x$: rows $(1,0,0)$, $(0,\cos,-\sin)$, $(0,\sin,\cos)$. About $y$: rows $(\cos,0,\sin)$, $(0,1,0)$, $(-\sin,0,\cos)$: the minus moves to the lower row because $\mathbf k \times \mathbf i = \mathbf j$ [Notes Eqs. 1.18–1.20].

Derivation Building $[{}_0R_1]$ from a drawing

In short: write each new unit vector in the old basis using the right triangle of the rotation angle; those components are the columns.
rotation $\theta$ about $\mathbf k_0$ (⊙, out of the paper)
the view along the axis shows $\mathbf i_0$, $\mathbf j_0$ in the plane; positive $\theta$ is counterclockwise
$\mathbf i_1 = \cos\theta\,\mathbf i_0 + \sin\theta\,\mathbf j_0$
unit hypotenuse: adjacent side $\cos\theta$ on $\mathbf i_0$, opposite side $\sin\theta$ on $\mathbf j_0$
$\mathbf j_1 = -\sin\theta\,\mathbf i_0 + \cos\theta\,\mathbf j_0$
$\mathbf j_1$ is $\mathbf i_1$ turned $90°$ more: its $\mathbf i_0$ component points backwards
$\mathbf k_1 = \mathbf k_0$
the axis of rotation does not move
check $\det = \cos^2\theta + \sin^2\theta = 1$
a right-handed orthonormal basis

If the axis points into the paper (⊗) in the view drawn, the positive angle is clockwise there and the signs of the $\sin$ terms swap.

Derivation Cylindrical coordinates

In short: the polar unit vectors rotate with $\theta$, $\dot{\mathbf e}_R = \dot\theta\,\mathbf e_\theta$ and $\dot{\mathbf e}_\theta = -\dot\theta\,\mathbf e_R$, and the product rule does the rest.
$\mathbf e_R = \cos\theta\,\mathbf i_0 + \sin\theta\,\mathbf j_0,\ \mathbf e_\theta = -\sin\theta\,\mathbf i_0 + \cos\theta\,\mathbf j_0$
$\mathbf e_R$ along $OP$ projected on $Oxy$; $\mathbf e_\theta$ at $+90°$
$\dot{\mathbf e}_R = \dot\theta(-\sin\theta\,\mathbf i_0 + \cos\theta\,\mathbf j_0) = \dot\theta\,\mathbf e_\theta$
chain rule: $\frac{d}{dt}\cos\theta = -\sin\theta\,\dot\theta$
$\dot{\mathbf e}_\theta = -\dot\theta\,\mathbf e_R$
same calculation
$\mathbf r = R\,\mathbf e_R + z\,\mathbf k_0 \Rightarrow \mathbf v = \dot R\,\mathbf e_R + R\dot\theta\,\mathbf e_\theta + \dot z\,\mathbf k_0$
product rule on $R\,\mathbf e_R$
$\mathbf a = (\ddot R - R\dot\theta^2)\,\mathbf e_R + (R\ddot\theta + 2\dot R\dot\theta)\,\mathbf e_\theta + \ddot z\,\mathbf k_0$
differentiate again; $\dot R\dot\theta$ appears twice

$-R\dot\theta^2$ is the centripetal term; $2\dot R\dot\theta$ vanishes if the radius or the angle is constant. Also $R\ddot\theta + 2\dot R\dot\theta = \frac1R\frac{d}{dt}(R^2\dot\theta)$.

Table Spherical coordinates (latitude $\phi$)

$x = r\cos\phi\cos\theta$, $y = r\cos\phi\sin\theta$, $z = r\sin\phi$, with $\theta$ the azimuth in $Oxy$ and $\phi \in [-\pi/2, \pi/2]$ measured up from the equator.

QuantityExpression
basis rates$\dot{\mathbf e}_r = \dot\theta\cos\phi\,\mathbf e_\theta + \dot\phi\,\mathbf e_\phi$; $\dot{\mathbf e}_\theta = -\dot\theta\cos\phi\,\mathbf e_r + \dot\theta\sin\phi\,\mathbf e_\phi$; $\dot{\mathbf e}_\phi = -\dot\phi\,\mathbf e_r - \dot\theta\sin\phi\,\mathbf e_\theta$
velocity$\dot r\,\mathbf e_r + r\dot\theta\cos\phi\,\mathbf e_\theta + r\dot\phi\,\mathbf e_\phi$
$a_r$$\ddot r - r\dot\phi^2 - r\dot\theta^2\cos^2\phi$
$a_\theta$$2\dot r\dot\theta\cos\phi + r\ddot\theta\cos\phi - 2r\dot\theta\dot\phi\sin\phi$
$a_\phi$$2\dot r\dot\phi + r\ddot\phi + r\dot\theta^2\sin\phi\cos\phi$

Trap: with the colatitude convention ($\phi$ measured from $+z$) every $\cos\phi$ becomes $\sin\phi$; check which angle the statement uses.

Derivation Intrinsic (Frenet) acceleration

In short: $\mathbf v = \dot s\,\mathbf e_t$; differentiating, the turning of $\mathbf e_t$ gives the normal term $\dot s^2/\rho$.
$\mathbf v = \dfrac{d\mathbf r}{ds}\dfrac{ds}{dt} = \dot s\,\mathbf e_t$
chain rule; $d\mathbf r/ds$ is the unit tangent
$\mathbf a = \ddot s\,\mathbf e_t + \dot s\,\dfrac{d\mathbf e_t}{ds}\dfrac{ds}{dt}$
product rule, then chain rule on $\mathbf e_t(s(t))$
$\dfrac{d\mathbf e_t}{ds} = \dfrac{1}{\rho}\,\mathbf e_n$
first Frenet formula: definition of curvature $\kappa = 1/\rho$
$\mathbf a = \ddot s\,\mathbf e_t + \dfrac{\dot s^2}{\rho}\,\mathbf e_n$
$a_b = 0$: $\mathbf v$ and $\mathbf a$ lie in the osculating plane
$\mathbf v \times \mathbf a = \dfrac{\dot s^3}{\rho}\,\mathbf e_b \Rightarrow \kappa = \dfrac{\lVert\mathbf v \times \mathbf a\rVert}{v^3}$
$\mathbf e_t \times \mathbf e_t = \mathbf 0$, $\mathbf e_t \times \mathbf e_n = \mathbf e_b$

Then $\mathbf e_t = \mathbf v/v$, $\mathbf e_b = \mathbf v \times \mathbf a/\lVert\mathbf v \times \mathbf a\rVert$, $\mathbf e_n = \mathbf e_b \times \mathbf e_t$, $a_t = \dot v = \mathbf a\cdot\mathbf e_t$, $a_n = \sqrt{a^2 - a_t^2}$. Centre of curvature: $\mathbf r_C = \mathbf r + \rho\,\mathbf e_n$.

Table Frenet-Serret formulas

FormulaMeaning
$\dfrac{d\mathbf e_t}{ds} = \kappa\,\mathbf e_n$the tangent turns towards the centre of curvature
$\dfrac{d\mathbf e_n}{ds} = -\kappa\,\mathbf e_t + \tau\,\mathbf e_b$the normal turns in the osculating plane and leaves it with the torsion
$\dfrac{d\mathbf e_b}{ds} = -\tau\,\mathbf e_n$a plane curve has $\tau = 0$ and a fixed binormal
$\tau = \dfrac{(\mathbf v \times \mathbf a)\cdot\dot{\mathbf a}}{\lVert\mathbf v \times \mathbf a\rVert^2}$torsion from time derivatives

Trap Intrinsic kinematics

  • $v_n = v_b = 0$ always; $a_b = 0$ always.
  • $a_t = 0$ means constant speed, not zero acceleration (uniform circular motion has $a_n = v^2/\rho$).
  • Arc length: $s(t) = \int_{t_0}^{t}\lVert\mathbf v(\tau)\rVert\,d\tau$; for a helix $v$ is constant and $s = vt$.

Topic 2 — Point particle dynamics

Theorem Newton's laws and momentum

  • Second law in an inertial frame: $\mathbf F = \left.\dfrac{d\mathbf p_0}{dt}\right|_0 = m\,\mathbf a_0^P$, $\mathbf p_0 = m\,\mathbf v_0^P$ [Slides 4–6].
  • Projected on any basis: Cartesian $m\ddot x = F_x$; cylindrical $m(\ddot R - R\dot\theta^2) = F_R$, $m(R\ddot\theta + 2\dot R\dot\theta) = F_\theta$, $m\ddot z = F_z$ [Slide 9].
  • State form: with $\mathbf v = \dot{\mathbf r}$, six first-order equations and six initial conditions $\mathbf r(0)$, $\mathbf v(0)$ [Slide 10].
  • Isolated system: $\sum\mathbf F^{ext} = \mathbf 0 \Rightarrow \sum m_i\mathbf v_i = $ const [Slides 7–8].

Derivation Kinetic energy theorem

In short: dot Newton's law with $\mathbf v$: $m\,\mathbf a\cdot\mathbf v = \frac{d}{dt}(\tfrac12mv^2)$, so power equals the rate of change of kinetic energy.
$m\,\mathbf a\cdot\mathbf v = \mathbf F\cdot\mathbf v$
dot product of $m\mathbf a = \mathbf F$ with $\mathbf v$
$\dfrac{d}{dt}\big(\tfrac12m\,\mathbf v\cdot\mathbf v\big) = m\,\dot{\mathbf v}\cdot\mathbf v$
product rule; the dot product is symmetric
$\dfrac{dT_0}{dt} = P = \mathbf F\cdot\mathbf v_0^P$
power of the force [Slide 13]
$T_0(t_2) - T_0(t_1) = \int_{t_1}^{t_2}P\,dt = \int_{\mathbf r_1}^{\mathbf r_2}\mathbf F\cdot d\mathbf r = W_{12}$
integrate with Barrow's rule; $\mathbf v\,dt = d\mathbf r$

A force normal to the velocity (a smooth fixed constraint, the Coriolis force) does no work.

Derivation Potential energy and energy conservation

In short: if $\mathbf F = -\nabla V_0$, then $\mathbf F\cdot\mathbf v = -\frac{dV_0}{dt}$, so $T_0 + V_0$ is constant when only conservative forces work.
$\dfrac{d}{dt}V_0(\mathbf r(t)) = \nabla V_0\cdot\dot{\mathbf r}$
chain rule in three variables
$P = \mathbf F\cdot\mathbf v = -\nabla V_0\cdot\mathbf v = -\dfrac{dV_0}{dt}$
definition of potential
$\dfrac{d}{dt}(T_0 + V_0) = P_{NC}$
add the kinetic energy theorem; only non-conservative forces remain
$E_0 = T_0 + V_0 = $ const if $P_{NC} = 0$
non-conservative forces absent or normal to $\mathbf v$ [Slides 19–20]

Table Potentials of the course

ForcePotential $V_0$
uniform gravity $-mg\,\mathbf k_0$$mgz$
linear spring, stiffness $k$, natural length $\ell_0$$\tfrac12k(\ell - \ell_0)^2$
inverse square $-\dfrac{GMm}{r^2}\mathbf e_r$$-\dfrac{GMm}{r}$
central $F(r)\,\mathbf e_r$$V_0(r) = -\int F(r)\,dr$
$aR\,\mathbf e_R$ (Slide 17)$-\tfrac12aR^2$

Pattern Is the force conservative?

  • Cartesian test: $\nabla \times \mathbf F = \mathbf 0$ on a simply connected domain, equivalently $\oint\mathbf F\cdot d\mathbf r = 0$ [Slides 14–16].
  • Plane polar test for $\mathbf F = F_R\,\mathbf e_R + F_\theta\,\mathbf e_\theta$: $\dfrac{\partial(RF_\theta)}{\partial R} - \dfrac{\partial F_R}{\partial\theta} = 0$.
  • Slide 18: $\mathbf F = bR^2\,\mathbf e_\theta$ gives $\partial_R(bR^3) = 3bR^2 \neq 0$: not conservative (a full turn does work $2\pi bR^3$).
  • Friction, drag and any force that depends on the velocity are not conservative.

Derivation One-dimensional motion from the energy

In short: $\tfrac12m\dot x^2 + V_0(x) = E_0$ gives the speed at each $x$, the allowed regions and, by separation, the time.
$\ddot x = \dfrac{d\dot x}{dt} = \dfrac{d\dot x}{dx}\dfrac{dx}{dt} = \dot x\dfrac{d\dot x}{dx} = \dfrac{d}{dx}\big(\tfrac12\dot x^2\big)$
chain rule: the standard trick to integrate $m\ddot x = F(x)$ once
$\tfrac12m\dot x^2 = \int F\,dx = -V_0(x) + E_0$
integrate in $x$; the constant is the energy
$v(x) = \sqrt{2(E_0 - V_0(x))/m}$
real only where $E_0 \ge V_0(x)$ (allowed region)
$t = \int_{x_0}^{x}\dfrac{dx'}{\sqrt{2(E_0 - V_0(x'))/m}}$
separable: $dt = dx/v$

Turning points: $V_0(x) = E_0$. Between two turning points the motion is periodic.

Theorem Equilibrium and stability

  • Equilibrium: $F(x^*) = -V_0'(x^*) = 0$ [Slides 21–22].
  • Stable if $V_0''(x^*) > 0$ (minimum), unstable if $V_0''(x^*) < 0$ (maximum); if $V_0'' = 0$ look at higher derivatives.
  • Small oscillations about a minimum: $m\ddot\xi + V_0''(x^*)\,\xi = 0$, $\omega = \sqrt{V_0''(x^*)/m}$ (Taylor: $V_0'(x^* + \xi) \approx V_0''(x^*)\xi$).

Derivation Simple pendulum

In short: in the polar basis of the rod, the transverse equation gives $\ddot\theta + (g/\ell)\sin\theta = 0$ and the radial one the tension $T = m(g\cos\theta + \ell\dot\theta^2)$.
$\mathbf r = \ell\,\mathbf e_r$, $\theta$ from the downward vertical
one degree of freedom; $\ell$ fixed by the rod
$\mathbf a = -\ell\dot\theta^2\,\mathbf e_r + \ell\ddot\theta\,\mathbf e_\theta$
cylindrical formulas with $\dot R = \ddot R = 0$
$m\mathbf g = mg\cos\theta\,\mathbf e_r - mg\sin\theta\,\mathbf e_\theta$
gravity is adjacent to $\theta$ along the rod, opposite across it, and pulls back towards $\theta = 0$
$\mathbf e_\theta$: $m\ell\ddot\theta = -mg\sin\theta$
the tension $-T\,\mathbf e_r$ has no transverse component
$\mathbf e_r$: $-m\ell\dot\theta^2 = mg\cos\theta - T$
gives the tension (a reaction)
$E_0 = \tfrac12m\ell^2\dot\theta^2 + mg\ell(1 - \cos\theta)$
first integral: multiply the $\mathbf e_\theta$ equation by $\dot\theta$ and integrate

Small angles: $\sin\theta \approx \theta$, $\omega_0 = \sqrt{g/\ell}$, period $2\pi\sqrt{\ell/g}$. Equilibria $\theta = 0$ (stable) and $\theta = \pi$ (unstable).

Topic 3 — Constraints and reaction forces

Table Degrees of freedom and reactions

ConstraintEquationsDOFNormal spaceUnknown reactions
free particlenone3—0
surface$f(\mathbf r) = 0$2$\nabla f$1
curve$f = 0,\ g = 0$1$\nabla f$, $\nabla g$2

Normal of a surface: $\nabla f$, or $\mathbf t_\alpha \times \mathbf t_\beta$ with $\mathbf t_\alpha = \partial\mathbf r/\partial\alpha$ from a parametrisation. Tangent of a curve: $d\mathbf r/d\alpha$, or $\nabla f \times \nabla g$ [Slides 7–12, Notes 6.3].

Derivation Why a smooth fixed constraint does no work

In short: differentiating $f(\mathbf r(t)) = 0$ shows $\mathbf v \perp \nabla f$; the reaction is along $\nabla f$, so $\mathbf N\cdot\mathbf v = 0$.
$\dfrac{d}{dt}f(\mathbf r(t), t) = \nabla f\cdot\mathbf v + \dfrac{\partial f}{\partial t} = 0$
chain rule along the motion; the constraint holds at all times
scleronomic, $\partial f/\partial t = 0$: $\nabla f\cdot\mathbf v = 0$
the velocity is tangent to the constraint
$\mathbf N = N_1\nabla f \Rightarrow P_N = \mathbf N\cdot\mathbf v = 0$
smooth constraint: the reaction lies in the normal space [Notes Eq. 6.12]
rheonomic: $P_N = -N_1\,\partial f/\partial t \neq 0$
a moving constraint can do work (spool, Slide 22)

Algorithm Dynamical decoupling

  1. Write $m\,\mathbf a_0^P = \mathbf F_{applied} + \mathbf N$ with $\mathbf N = N_1\mathbf n$ (surface) or $N_1\mathbf n_f + N_2\mathbf n_g$ (curve) [Slide 13].
  2. Project on the tangent vectors: $\mathbf N$ disappears and the equations of motion of the free coordinates remain [Slide 14].
  3. Project on the normal vectors: each projection gives one reaction once the motion is known.
  4. Along a curve use the intrinsic frame: $F_t = m\ddot s$, $F_n + N_n = m\dot s^2/\rho$, $F_b + N_b = 0$ [Notes Eqs. 6.17–6.19].
  5. Conservative applied forces and fixed smooth constraints: use $E_0 = $ const as a first integral to get $\dot s^2$ (or $\dot\theta^2$) in the normal equation.

Table Coulomb friction

RegimeFriction forcePower
sliding, $\mathbf v^* \neq \mathbf 0$applied force $\mathbf F_R = -\mu N\,\dfrac{\mathbf v^*}{\lVert\mathbf v^*\rVert}$$-\mu Nv^* < 0$
sticking, $\mathbf v^* = \mathbf 0$constraint force in the tangent space, $\lVert\mathbf F_R\rVert \le \mu N$$0$

$\mathbf v^*$ is the velocity relative to the surface and $N$ the magnitude of the normal reaction. Switch to sliding when the tangential force needed to stay at rest exceeds $\mu N$; switch back to rest when $\mathbf v^* = \mathbf 0$ and the applied tangential force is at most $\mu N$ [Slides 18–19, Notes 6.6].

Derivation Atwood machine

In short: the inextensible string gives $\ddot x_2 = -\ddot x_1$; subtracting the two equations of motion eliminates the tension.
$x_1 + x_2 = $ const $\Rightarrow \ddot x_2 = -\ddot x_1$
string length fixed, $x_i$ measured downwards, ideal pulley
$m_1\ddot x_1 = m_1g - T,\ m_2\ddot x_2 = m_2g - T$
same tension on both sides of a massless frictionless pulley
$(m_1 + m_2)\ddot x_1 = (m_1 - m_2)g$
subtract the second equation from the first
$T = \dfrac{2m_1m_2}{m_1 + m_2}\,g$
back into the first equation

Checks: $m_1 = m_2$ gives rest or uniform motion and $T = m_1g$; $m_2 = 0$ gives free fall.

Table Classification of constraints

TypeFormConsequence
holonomic$f(\mathbf r, t) = 0$removes one coordinate
non-holonomic$\sum A_i\dot x_i + B = 0$, not integrableremoves a velocity, not a coordinate
scleronomic$\partial f/\partial t = 0$no work if smooth
rheonomic$\partial f/\partial t \neq 0$the reaction can do work
unilateral$f \ge 0$, $N \ge 0$the particle leaves the constraint when $N = 0$

Trap: for a unilateral contact (a bead on the outside of a sphere, an aircraft on the runway) compute $N$ as a function of the position and find where it reaches zero; beyond that point the equations with $N$ are no longer valid.

Topic 4 — Angular momentum and central forces

Derivation Angular momentum about a moving point

In short: differentiating $\mathbf H_{A0}^P = \overrightarrow{AP} \times m\mathbf v_0^P$ gives $\dot{\mathbf H}_{A0} = \mathbf M_A - \mathbf v_0^A \times m\mathbf v_0^P$.
$\mathbf H_{A0}^P = (\mathbf r_0^P - \mathbf r_0^A) \times m\,\mathbf v_0^P$
definition [Slide 3, Notes 4.8]
$\dot{\mathbf H}_{A0}^P = (\mathbf v_0^P - \mathbf v_0^A) \times m\mathbf v_0^P + \overrightarrow{AP} \times m\,\mathbf a_0^P$
product rule for the cross product (order kept)
$\mathbf v_0^P \times m\mathbf v_0^P = \mathbf 0$
parallel vectors
$\overrightarrow{AP} \times m\mathbf a_0^P = \overrightarrow{AP} \times \mathbf F = \mathbf M_A$
Newton's second law
$\dot{\mathbf H}_{A0}^P = \mathbf M_A - \mathbf v_0^A \times m\,\mathbf v_0^P$
the extra term vanishes if $A$ is fixed or $\mathbf v_0^A \parallel \mathbf v_0^P$ [Slide 4]

Conservation: $\mathbf M_A = \mathbf 0$ about a fixed point $\Rightarrow \mathbf H_{A0} = $ const (three scalar first integrals).

Derivation Central forces: planar motion and Kepler's second law

In short: $\mathbf M_O = \mathbf r \times F(r)\mathbf e_r = \mathbf 0$, so $\mathbf H_O$ is constant, the motion is planar and $h = r^2\dot\theta$ is constant.
$\mathbf r \times F(r)\,\mathbf e_r = \mathbf 0$
$\mathbf r$ and $\mathbf e_r$ are parallel
$\mathbf H_O = $ const, $\mathbf r \perp \mathbf H_O$
$\mathbf H_O = \mathbf r \times m\mathbf v$ is normal to $\mathbf r$: $\mathbf r$ stays in a fixed plane
$m(r\ddot\theta + 2\dot r\dot\theta) = 0$
transverse equation: a central force has no $\mathbf e_\theta$ component
$r\ddot\theta + 2\dot r\dot\theta = \dfrac1r\dfrac{d}{dt}(r^2\dot\theta) = 0$
product rule backwards
$h = r^2\dot\theta = $ const, $\dfrac{dA}{dt} = \tfrac12r^2\dot\theta = \dfrac h2$
the area swept in $dt$ is the triangle $\tfrac12r\cdot r\,d\theta$ [Slide 12]

Derivation Binet's equation

In short: with $u = 1/r$ and $d/dt = hu^2\,d/d\theta$, the radial equation becomes $u'' + u = -F/(mh^2u^2)$; for gravity $u'' + u = \mu/h^2$.
$\dot\theta = hu^2$
$h = r^2\dot\theta$ with $r = 1/u$
$\dot r = -\dfrac{\dot u}{u^2} = -\dfrac{u'\dot\theta}{u^2} = -h\,u'$
chain rule twice; $' = d/d\theta$
$\ddot r = -h\,u''\,\dot\theta = -h^2u^2u''$
$h$ is constant
$m(\ddot r - r\dot\theta^2) = F$ with $r\dot\theta^2 = h^2u^3$
radial equation [Slide 13, Notes 8.4]
$u'' + u = -\dfrac{F(1/u)}{mh^2u^2}$
divide by $-mh^2u^2$
$F = -\dfrac{\mu m}{r^2}$: $u = \dfrac{\mu}{h^2} + C\cos(\theta - \theta_0)$
constant forcing: particular solution plus harmonic part
$r = \dfrac{p}{1 + e\cos(\theta - \theta_0)},\ p = \dfrac{h^2}{\mu},\ e = \dfrac{Ch^2}{\mu}$
a conic with focus at $O$ (Kepler's first law); periapsis at $\theta = \theta_0$

Table Conics and energy

Orbit$e$$a$$\xi = \tfrac12v^2 - \mu/r$
circle$0$$p$$-\mu/(2a) < 0$
ellipse$0 < e < 1$$p/(1 - e^2) > 0$$-\mu/(2a) < 0$
parabola$1$$\infty$$0$
hyperbola$e > 1$$< 0$$-\mu/(2a) > 0$

Ellipse: $r_p = a(1 - e)$, $r_a = a(1 + e)$, $b = a\sqrt{1 - e^2}$, $p = a(1 - e^2) = b^2/a$, $h = \sqrt{\mu p}$, $v_p/v_a = r_a/r_p$.

Derivation Vis-viva and Kepler's third law

In short: evaluate $\xi$ at periapsis with $h = r_pv_p$ to get $\xi = -\mu/(2a)$; the area rate $h/2$ gives the period $\tau = 2\pi\sqrt{a^3/\mu}$.
$v_p = h/r_p,\ r_p = \dfrac{p}{1 + e},\ h^2 = \mu p$
at periapsis $\dot r = 0$, so $v = r\dot\theta = h/r$
$\xi = \dfrac{\mu(1 + e)^2}{2p} - \dfrac{\mu(1 + e)}{p} = -\dfrac{\mu(1 - e^2)}{2p}$
substitute and factor $\mu(1 + e)/(2p)$
$\xi = -\dfrac{\mu}{2a} \Rightarrow v^2 = \mu\Big(\dfrac2r - \dfrac1a\Big)$
$p = a(1 - e^2)$; vis-viva [Notes Eq. 8.19]
$\tau = \dfrac{\pi ab}{h/2} = \dfrac{2\pi ab}{\sqrt{\mu a(1 - e^2)}} = 2\pi\sqrt{\dfrac{a^3}{\mu}}$
ellipse area over areal velocity; $b = a\sqrt{1 - e^2}$ [Notes Eqs. 8.16–8.18]

Circular speed $v_c = \sqrt{\mu/r}$, escape speed $v_e = \sqrt{2\mu/r} = \sqrt2\,v_c$ [Notes Eqs. 8.20–8.21].

Table Two-body reduction

QuantityFormula
barycentre$\mathbf r_G = \dfrac{m_P\mathbf r_P + m_S\mathbf r_S}{m_P + m_S}$, $\ddot{\mathbf r}_G = \mathbf 0$
relative vector$\mathbf r = \mathbf r_P - \mathbf r_S$, $\ddot{\mathbf r} = -\dfrac{\mu}{r^3}\mathbf r$, $\mu = G(m_P + m_S)$
bodies about $G$$\mathbf r_P - \mathbf r_G = \dfrac{m_S}{m_P + m_S}\mathbf r$, $\mathbf r_S - \mathbf r_G = -\dfrac{m_P}{m_P + m_S}\mathbf r$

Pattern Effective potential in $u$

For any central force, $(u')^2 + W_{eff}(u) = 2\xi/h^2$ with $W_{eff}(u) = u^2 + 2V_0(1/u)/h^2$ ($V_0$ per unit mass). Motion only where $W_{eff} \le 2\xi/h^2$; the roots give $r_{min}$ and $r_{max}$. For gravity, $W_{eff} = u^2 - 2\mu u/h^2$.

Topic 5 — Relative motion

Derivation The Coriolis theorem

In short: write $\mathbf b$ in the moving basis; the derivative of each moving unit vector seen from $S_0$ is $\boldsymbol\omega_{10} \times$ that vector.
$\mathbf b = b_x\mathbf i_1 + b_y\mathbf j_1 + b_z\mathbf k_1$
components in $\mathcal B_1$
$\left.\dfrac{d\mathbf b}{dt}\right\rvert_0 = (\dot b_x\mathbf i_1 + \dot b_y\mathbf j_1 + \dot b_z\mathbf k_1) + b_x\dot{\mathbf i}_1 + b_y\dot{\mathbf j}_1 + b_z\dot{\mathbf k}_1$
product rule; seen from $S_0$ the unit vectors move
$\left.\dfrac{d\mathbf i_1}{dt}\right\rvert_0 = \boldsymbol\omega_{10} \times \mathbf i_1$ (same for $\mathbf j_1$, $\mathbf k_1$)
Poisson: a unit vector only rotates [Notes Eqs. 3.3–3.5, Slide 14]
$\left.\dfrac{d\mathbf b}{dt}\right\rvert_0 = \left.\dfrac{d\mathbf b}{dt}\right\rvert_1 + \boldsymbol\omega_{10} \times \mathbf b$
the bracket is the derivative seen from $S_1$ [Notes Eq. 3.8]

Consequence: $\boldsymbol\alpha_{10} = \dot{\boldsymbol\omega}_{10}\rvert_0 = \dot{\boldsymbol\omega}_{10}\rvert_1$ because $\boldsymbol\omega_{10} \times \boldsymbol\omega_{10} = \mathbf 0$ [Notes Eq. 3.9].

Table Composition of angular motions

ResultFormula
velocities add$\boldsymbol\omega_{20} = \boldsymbol\omega_{21} + \boldsymbol\omega_{10}$ [Notes Eq. 3.14]
accelerations do not$\boldsymbol\alpha_{20} = \boldsymbol\alpha_{21} + \boldsymbol\omega_{10} \times \boldsymbol\omega_{21} + \boldsymbol\alpha_{10}$ [Notes Eq. 3.15]
inverse$\boldsymbol\omega_{01} = -\boldsymbol\omega_{10}$, $\boldsymbol\alpha_{01} = -\boldsymbol\alpha_{10}$
simple rotation$\psi(t)$ about $\mathbf k_0$: $\boldsymbol\omega_{10} = \dot\psi\,\mathbf k_0$ [Slide 16]
Euler 3-1-3$\boldsymbol\omega_{30} = \dot\psi\,\mathbf k_0 + \dot\theta\,\mathbf i_1 + \dot\varphi\,\mathbf k_2$ [Slides 20–21]

The $\boldsymbol\omega_{10} \times \boldsymbol\omega_{21}$ term comes from differentiating $\boldsymbol\omega_{21}$ (known in $S_1$) from $S_0$ with the Coriolis theorem.

Derivation Relative velocity and acceleration

In short: differentiate $\mathbf r_0^P = \mathbf r_0^A + \mathbf r_1^P$ twice in $S_0$, applying the Coriolis theorem to every vector known in $S_1$.
$\mathbf r_0^P = \mathbf r_0^A + \mathbf r_1^P$
vector chain $\overrightarrow{OA} + \overrightarrow{AP}$ [Notes Eq. 3.16]
$\mathbf v_0^P = \mathbf v_0^A + \mathbf v_1^P + \boldsymbol\omega_{10} \times \mathbf r_1^P$
Coriolis theorem on $\mathbf r_1^P$ [Notes Eq. 3.26]
$\left.\dfrac{d\mathbf v_1^P}{dt}\right\rvert_0 = \mathbf a_1^P + \boldsymbol\omega_{10} \times \mathbf v_1^P$
Coriolis theorem on $\mathbf v_1^P$
$\left.\dfrac{d}{dt}(\boldsymbol\omega_{10} \times \mathbf r_1^P)\right\rvert_0 = \boldsymbol\alpha_{10} \times \mathbf r_1^P + \boldsymbol\omega_{10} \times (\mathbf v_1^P + \boldsymbol\omega_{10} \times \mathbf r_1^P)$
product rule, then Coriolis on $\mathbf r_1^P$
$\mathbf a_0^P = \mathbf a_0^A + \mathbf a_1^P + \boldsymbol\alpha_{10} \times \mathbf r_1^P + \boldsymbol\omega_{10} \times (\boldsymbol\omega_{10} \times \mathbf r_1^P) + 2\boldsymbol\omega_{10} \times \mathbf v_1^P$
the two $\boldsymbol\omega_{10} \times \mathbf v_1^P$ make the Coriolis term [Notes Eq. 3.27]

Names: $\mathbf a_0^A$ origin, $\mathbf a_1^P$ relative, $\boldsymbol\alpha \times \mathbf r$ Euler, $\boldsymbol\omega \times (\boldsymbol\omega \times \mathbf r) = -\omega^2\mathbf r_\perp$ centripetal, $2\boldsymbol\omega \times \mathbf v_1^P$ Coriolis. Pure translation: $\mathbf a_0^P = \mathbf a_0^A + \mathbf a_1^P$.

Table Inertia forces in a non-inertial frame

$m\,\mathbf a_1^P = \mathbf F + \mathbf F_i$ [Notes Eq. 4.63], with $\mathbf F_i = -m(\mathbf a_0^P - \mathbf a_1^P)$:

TermForcePotential, work
origin$-m\,\mathbf a_0^A$$V_1 = m\,\mathbf a_0^A\cdot\mathbf r_1^P$ if $\mathbf a_0^A$ is constant
Euler$-m\,\boldsymbol\alpha_{10} \times \mathbf r_1^P$no potential
centrifugal$-m\,\boldsymbol\omega_{10} \times (\boldsymbol\omega_{10} \times \mathbf r_1^P) = m\omega^2\mathbf r_\perp$$V_1 = -\tfrac12m\omega^2r_\perp^2$ if $\boldsymbol\omega_{10}$ is constant
Coriolis$-2m\,\boldsymbol\omega_{10} \times \mathbf v_1^P$never works: $\perp \mathbf v_1^P$

$\mathbf r_\perp$ is the component of $\mathbf r_1^P$ perpendicular to the rotation axis through $A$. Energy in $S_1$: $E_1 = \tfrac12m(v_1^P)^2 + V - \tfrac12m\omega^2r_\perp^2$ is conserved if $\mathbf a_0^A = \mathbf 0$, $\boldsymbol\omega_{10}$ constant, actual forces conservative and constraints fixed in $S_1$ [Slide 29].

Algorithm Relative-motion problem

  1. Choose $S_1$ attached to the moving body (tube, disc, platform) so that the particle's motion in $S_1$ is simple.
  2. Write $\boldsymbol\omega_{10}$, $\boldsymbol\alpha_{10}$, $\mathbf a_0^A$ in $\mathcal B_1$; deduce $[{}_0R_1]$ from the drawing.
  3. Relative kinematics: $\mathbf r_1^P$, $\mathbf v_1^P$, $\mathbf a_1^P$ in $\mathcal B_1$.
  4. Either absolute route ($m\mathbf a_0^P = \mathbf F$ with the composition formula) or relative route ($m\mathbf a_1^P = \mathbf F + \mathbf F_i$): both give the same equations.
  5. Project: the directions free for the particle give its equation of motion; the constrained directions give the reactions.
  6. Checks: $\Omega = 0$ recovers the fixed-frame problem; units of every term m/s²; $E_1$ conserved if the conditions above hold.

Trap Relative-motion traps

  • $\mathbf v_1^P$ is the velocity seen from $S_1$, not the component of $\mathbf v_0^P$ in $\mathcal B_1$.
  • $\boldsymbol\alpha_{10} \neq \mathbf 0$ whenever the direction of $\boldsymbol\omega_{10}$ changes, even with constant magnitude.
  • The centrifugal force points away from the axis, not away from $A$.
  • Both routes at once: inertia forces belong only to the equation written in $S_1$.

Pocket prerequisites

Prerequisite Vector algebra

IdentityUse
$\mathbf a \times (\mathbf b \times \mathbf c) = \mathbf b\,(\mathbf a\cdot\mathbf c) - \mathbf c\,(\mathbf a\cdot\mathbf b)$centripetal term $\boldsymbol\omega \times (\boldsymbol\omega \times \mathbf r) = \boldsymbol\omega(\boldsymbol\omega\cdot\mathbf r) - \omega^2\mathbf r$
$\mathbf a \times \mathbf b = -\mathbf b \times \mathbf a$, $\mathbf a \times \mathbf a = \mathbf 0$order in torques and Coriolis
$\mathbf i \times \mathbf j = \mathbf k,\ \mathbf j \times \mathbf k = \mathbf i,\ \mathbf k \times \mathbf i = \mathbf j$cyclic order of a right-handed basis
$\mathbf a\cdot(\mathbf b \times \mathbf c) = \det[\mathbf a, \mathbf b, \mathbf c]$torsion, volumes, coplanarity
$\dfrac{d}{dt}(\mathbf a \times \mathbf b) = \dot{\mathbf a} \times \mathbf b + \mathbf a \times \dot{\mathbf b}$angular momentum, relative kinematics

Prerequisite Calculus tools

  • Chain rule: $\dfrac{d}{dt}f(\theta(t)) = f'(\theta)\,\dot\theta$, so $\dfrac{d}{dt}\cos\theta = -\dot\theta\sin\theta$ and $\dfrac{d}{dt}(\dot\theta^2) = 2\dot\theta\ddot\theta$.
  • Integrating $\ddot\theta = g(\theta)$ once: $\ddot\theta = \dfrac{d}{d\theta}\big(\tfrac12\dot\theta^2\big)$, so $\tfrac12\dot\theta^2 - \tfrac12\dot\theta_0^2 = \int_{\theta_0}^{\theta}g(s)\,ds$.
  • Barrow's rule: $\int_a^b f(x)\,dx = F(b) - F(a)$; write the antiderivative before substituting the limits.
  • Small angles: $\sin\theta \approx \theta$, $\cos\theta \approx 1 - \theta^2/2$, $\tan\theta \approx \theta$.

Prerequisite Trigonometry

  • $\cos(\theta - \pi/2) = \sin\theta$, $\sin(\theta - \pi/2) = -\cos\theta$: relating an angle from the vertical to the polar angle.
  • $\sin(a \pm b) = \sin a\cos b \pm \cos a\sin b$, $\cos(a \pm b) = \cos a\cos b \mp \sin a\sin b$.
  • $\cos^2\theta = \tfrac12(1 + \cos 2\theta)$, $\sin\theta\cos\theta = \tfrac12\sin 2\theta$.
  • In every right triangle: adjacent $=$ hypotenuse $\times\cos$, opposite $=$ hypotenuse $\times\sin$.

Prerequisite Units and dimensions

QuantitySI unit
force, $m\mathbf a$, $m\omega^2r$, $m\omega v$$\mathrm{N = kg\,m\,s^{-2}}$
energy, work, $V_0$, $\tfrac12mv^2$$\mathrm{J = N\,m}$
power $\mathbf F\cdot\mathbf v$$\mathrm{W = J\,s^{-1}}$
angular momentum $\mathbf r \times m\mathbf v$; torque$\mathrm{kg\,m^2\,s^{-1}}$; $\mathrm{N\,m}$
$h = r^2\dot\theta$; $\mu = GM$$\mathrm{m^2\,s^{-1}}$; $\mathrm{m^3\,s^{-2}}$
$\omega$, $\alpha$, curvature $\kappa$$\mathrm{s^{-1}}$, $\mathrm{s^{-2}}$, $\mathrm{m^{-1}}$