Visual walkthrough — Converting between orbital elements and state vectors (r, v)

3.2.17 · D2 · Physics › Orbital Mechanics & Astrodynamics › Converting between orbital elements and state vectors (r, v)

Shuru karne se pehle: ek vector bas ek arrow hai jiske paas length aur direction hai. Hum ise likhte hain $\vec{r}$. Iski length hum likhte hain $r=|\vec{r}|$ (bina arrow ke = sirf ek number). Ek frame teen arrows ka ek set hai jo right angles par hain — jaise ek room ka corner — jiske against hum sab kuch measure karte hain. In do ideas ko pakde rakho; baaki sab hum khud kamaate hain.

---

Step 1 — Orbit ko flat draw karo, jahan yeh sabse simple hai

KYA. Hum ellipse ko ek flat table par rakhte hain taaki iski planet ke closest point (periapsis) seedhi ek chosen arrow ke saath lie kare. Us arrow ko hum $\hat{p}$ kehte hain ("p-hat", chota hat matlab "length 1 wala arrow, sirf ek direction"). Us se $90^{\circ }$ par wala arrow, table ke plane mein, hum $\hat{q}$ kehte hain. Milke $\hat{p},\hat{q}$ (aur ek teesra, $\hat{w}$, jo seedha table se upar nikalta hai) perifocal frame banate hain — woh frame jo sirf is ek orbit ke liye bani hai.

YE YAHAN KYUN. Kisi bhi aur frame mein ellipse tilted aur swivelled hoti hai aur likhna mushkil hota hai. Perifocal frame mein periapsis $\hat{p}$ se nail ki hui hai, isliye body ki position ek akele angle $\nu$ (true anomaly) se describe hoti hai jo us $\hat{p}$ axis se measure hoti hai. Ek pechida tangle ki jagah ek angle — isliye hum yahan se shuru karte hain.

PICTURE. Planet focus par baitha hai (black dot). Lal arrow $\hat{p}$ periapsis ki taraf aim kar raha hai. Body $\hat{p}$ se angle $\nu$ par sweep karke baitha hai.

---

Step 2 — Body kitni door hai? Orbit equation

KYA. Humein distance $r$ chahiye — focus se body tak — jab angle $\nu$ badalta hai. Conic equation yeh deta hai: $\underset{\text{abhi ki distance}}{\underbrace{r}}=\frac{\stackrel{\text{semi-latus rectum}}{\overbrace{p}}}{1+\underset{\text{shape}}{\underbrace{e}}\underset{\text{orbit par kahan}}{\underbrace{\cos \nu }}},p=a(1-e^2)=\frac{h^2}{\mu }$

Term by term: $p$ ek fixed length hai (orbit ka "width parameter") — $a$ (definition box wala size number) aur $e$ se bana, ya equivalently $h$ (sweep rate) aur $\mu$ (pull strength) se; $e$ ellipse ko squash karta hai; $\cos \nu$ periapsis par $+1$ ($\nu =0$) se apoapsis par $-1$ ($\nu =180^{\circ }$) tak swing karta hai.

YEH SHAPE KYUN. $\nu =0$ par: $\cos \nu =1$, toh $r=p/(1+e)=a(1-e)$ — sabse chhoti distance. $\nu =180^{\circ }$ par: $\cos \nu =-1$, toh $r=p/(1-e)=a(1+e)$ — sabse badi. Donon extreme ka average $a$ hota hai, jo confirm karta hai ki $a$ sach mein long axis ki half-length hai. Formula automatically ek baar per orbit andar-bahar breathe karta hai. Isliye hum ise har $\nu$ ke liye trust karte hain.

PICTURE. Dekho $r$ kaise shrink aur grow karta hai jab body ghoomti hai; lal near point mark karta hai.

---

Step 3 — Kitni tez, aur kis direction mein? Perifocal velocity

KYA. Velocity yeh hai ki position arrow har second kitna badalti hai, $\vec{v}=\frac{d\vec{r}}{dt}$. Hum ${\vec{r}}_{pqw}=r(\cos \nu ,\sin \nu ,0)$ differentiate karte hain. Yahan time ke saath do cheezein badal rahi hain: length $r$ aur angle $\nu$. Toh chain rule se (ek piece differentiate karo, doosra rakho) har component do parts mein split hoti hai — ek "$r$ badla" part aur ek "$\nu$ badla" part: $\frac{d{\vec{r}}_{pqw}}{dt}=\underset{\text{length badhi}}{\underbrace{\dot{r}\left[\begin{array}{c}\cos \nu \\ \sin \nu \\ 0\end{array}\right]}}+\underset{\text{angle sweep hua}}{\underbrace{r\dot{\nu }\left[\begin{array}{c}-\sin \nu \\ \cos \nu \\ 0\end{array}\right]}}$ Yahan $\dot{r}=dr/dt$ aur $\dot{\nu }=d\nu /dt$ (upar dot matlab "rate per second").

AREA LAW KYUN AATA HAI. Ab humein $\dot{r}$ aur $\dot{\nu }$ chahiye sirf $\nu$ ke functions ke roop mein — koi messy time nahi. Area law $r^2\dot{\nu }=h$ (constant sweep, Angular momentum conservation in orbits se) seedha $\dot{\nu }=h/r^2$ deta hai. $\dot{r}$ ke liye, orbit equation $r=p/(1+e\cos \nu )$ ko chain rule se differentiate karo aur phir $\dot{\nu }=h/r^2$ substitute karo: $\dot{r}=\frac{pe\sin \nu }{(1+e\cos \nu {)}^2}\dot{\nu }=\frac{r^{2}e\sin \nu }{p}\cdot \frac{h}{r^2}=\frac{h}{p}e\sin \nu$ Aur $r\dot{\nu }=r\cdot h/r^2=h/r=\frac{h}{p}(1+e\cos \nu )$. Donon ko upar wale split mein substitute karo: $\frac{d{\vec{r}}_{pqw}}{dt}=\frac{h}{p}[e\sin \nu (\cos \nu ,\sin \nu ,0)+(1+e\cos \nu )(-\sin \nu ,\cos \nu ,0)]$ Pehla component expand karo: $e\sin \nu \cos \nu -\sin \nu -e\cos \nu \sin \nu =-\sin \nu$. Doosra component: $e{\sin }^2\nu +\cos \nu +e{\cos }^2\nu =e+\cos \nu$. Cross terms cancel ho jaate hain — isi liye tidy form aati hai. Aakhir mein $h/p=\sqrt{\mu /p}$ kyunki $p=h^2/\mu \Rightarrow h=\sqrt{\mu p}$, toh $h/p=\sqrt{\mu p}/p=\sqrt{\mu /p}$. Result: ${\vec{v}}_{pqw}=\sqrt{\frac{\mu }{p}}\left[\begin{array}{c}-\sin \nu \\ e+\cos \nu \\ 0\end{array}\right]$

Term by term: $\sqrt{\mu /p}$ ek fixed speed scale hai (zyada pull $\mu$ ya tighter orbit $p$ → zyada fast); $-\sin \nu$ radial (andar/bahar) part hai; $e+\cos \nu$ tangential (sideways) part hai.

PICTURE. Split ke do arrows — ek position ke along ("length badhi"), ek perpendicular ("angle sweep hua") — add hokar total velocity banate hain. Periapsis par ($\nu =0$) radial part $-\sin 0=0$: velocity purely sideways hai, body na climb kar rahi hai na fall. Yeh orbit ka fastest point hai.

Recall Periapsis case check karo

$\nu =0$ par: ${\vec{v}}_{pqw}=\sqrt{\mu /p}(0,e+1,0)$ — sab tangential. ✓ $\nu =180^{\circ }$ (apoapsis) par: ${\vec{v}}_{pqw}=\sqrt{\mu /p}(0,e-1,0)$ — phir bhi tangential lekin slower (kyunki $e-1$ size mein $e+1$ se chhota hai). Donon turning points par zero radial speed hai, exactly jaise vis-viva picture demand karti hai.

---

Step 4 — Ab paper flat hai lekin space nahi: teen spins

KYA. Hamare vectors table par hain ($\hat{p},\hat{q},\hat{w}$). Real world ek fixed frame use karta hai — ECI frame, jisme $\hat{x}$ ek fixed star direction par point karta hai aur $\hat{z}$ planet ke spin axis ke upar (dekho Reference frames — perifocal (PQW) vs ECI). Humein table ko spin karna hoga jab tak woh line up na ho jaye. Teen spins karte hain, innermost-first:

1. Paper ko apne $\hat{w}$ axis ke baare mein $\omega$ se ghumaao — yeh periapsis ko plane ke andar sahi jagah point karta hai.
2. Paper ko node line ke baare mein $i$ se tilt karo — yeh plane ko horizontal se utha deta hai.
3. Poori cheez ko fixed $\hat{z}$ ke baare mein $\Omega$ se swivel karo — yeh tilt ko sahi compass direction mein point karta hai.

TEEN KYUN AUR IS ORDER MEIN. Har element ek alag "kis taraf?" sawaal ka jawab deta hai: $\omega$ = plane ke andar periapsis kahan hai, $i$ = plane kitna steeply tilted hai, $\Omega$ = tilt kis direction mein face kar raha hai. Tum swivel ($\Omega$) nahi kar sakte pehle kuch tilt ($i$) kiye bina jise swivel kar sako, aur sensibly tilt nahi kar sakte pehle periapsis apni jagah ($\omega$) ke bina. Order = $\omega$, phir $i$, phir $\Omega$ — inside out.

PICTURE. Wahi lal periapsis arrow, teeno spins ke through follow kiya gaya; planes milne wali jagah par node line marked.

---

Step 5 — $R_3$ aur $R_1$ actually kya hain, aur negative angles kyun

KYA. Ek rotation matrix numbers ka ek $3\times 3$ box hai jo ek vector ke teen components leta hai aur spin ke baad uske components return karta hai. Hum do flavours use karte hain, naam se pata chalta hai kaunse axis ke baare mein spin karte hain:

R_1(\theta)=\begin{bmatrix}1 & 0 & 0\\ 0 & \cos\theta & -\sin\theta\\ 0 & \sin\theta & \cos\theta\end{bmatrix}\ \text{(1st axis ke baare mein spin, }\hat x\text{)}$$ Pattern notice karo: jis axis ke baare mein spin karte ho woh row/column untouched rehta hai (diagonal par uska $1$, baaju mein zeros); baaki do entries $\cos$ aur $\sin$ se mix hoti hain. **NEGATIVE ANGLES KYUN.** Parent recipe $R_3(-\Omega)\,R_1(-i)\,R_3(-\omega)$ use karti hai — **minus** signs ke saath. Do cheezein hain jo tum spin kar sakte ho: *arrow* ya *woh frame jisme measure karte ho*. $R_3(+\theta)$ *arrow* ko $+\theta$ se rotate karta hai. Lekin hum *frame* rotate karte hain (perifocal → inertial); wahi physical arrow ko tab **nayi numbers** milti hain jaise usse $-\theta$ se ghuma diya ho. Toh har angle sign flip kar leta hai. Yeh poori conversion mein sabse common bug hai. **PICTURE.** Ek fixed lal arrow, do descriptions: rotate-the-arrow vs rotate-the-frame — mirror images. ![[deepdives/dd-physics-3.2.17-d2-s05.png]] > [!mistake] Sign trap > $R_3(+\omega)$ likhna *sahi lagta hai* ("maine $\omega$ se rotate kiya"). Lekin tumne **frame** rotate kiya, toh components transform hote hain inverse se: $R_3(-\omega)$. Yeh galat karo aur tumhara orbit mirror-flipped nikalta hai — periapsis galat side par. --- ## Step 6 — Teen spins ko ek matrix mein assemble karo **KYA.** Teen spin-boxes ko order mein multiply karo — $R_3(-\Omega)$ outermost, phir $R_1(-i)$, phir $R_3(-\omega)$ innermost — aur product ek matrix $\tilde R$ mein collapse ho jaata hai (jahan $c\theta=\cos\theta$, $s\theta=\sin\theta$): $$\tilde R=\begin{bmatrix} c\Omega c\omega - s\Omega s\omega c i & -c\Omega s\omega - s\Omega c\omega c i & s\Omega s i\\ s\Omega c\omega + c\Omega s\omega c i & -s\Omega s\omega + c\Omega c\omega c i & -c\Omega s i\\ s\omega s i & c\omega s i & c i \end{bmatrix}$$ **ENTRIES KAHAN SE AATI HAIN.** Inhe yaad nahi karte — inhe *assemble hote dekho* Step 4 ki teen pictures se: - **Bottom-right $c i$** pure tilt hai: yeh woh akeli jagah hai jo donon $\hat z$-spins ($\Omega$ aur $\omega$ up-axis ko akela chhodte hain) se untouched rehti hai, toh wahan sirf $R_1(-i)$ ka $\cos i$ bachta hai. - **Bottom row** ($s\omega s i,\ c\omega s i,\ c i$) orbit ke $\hat w$ axis ki image hai: yeh sirf innermost spin $\omega$ aur tilt $i$ feel karta hai — koi $\Omega$ nahi — kyunki tilted plane ko $\hat z$ ke baare mein swivel karna yeh nahi badalta ki reference plane se ek vector *kitna bahar* reach karta hai. - **Top-left block** ($c\Omega c\omega - s\Omega s\omega c i$, etc.) donon $\hat z$-spins ($\Omega$ aur $\omega$) hain jo ek saath chained hain, *linked* us $c i$ factor se jo wahan leak karta hai jahan tilt unke beech hai. Jab $i=0$ (koi tilt nahi) yeh plain angle-addition $\cos(\Omega+\omega)$ tak reduce ho jaate hain — donon flat swivels simply add ho jaate hain. Phir inertial state vectors simply hain: $$\vec r = \tilde R\,\vec r_{pqw}, \qquad \vec v = \tilde R\,\vec v_{pqw}$$ **DONON KE LIYE SAME $\tilde R$ KYUN.** Position aur velocity donon arrows hain jo same flat orbit plane par rehte hain. Plane ko spin karna *donon* ko identical rule se spin karta hai. Ek matrix, do multiplications. **PICTURE — worked numbers.** Parent ke Example 2 se: $a=8000$, $e=0.1$, $i=30^\circ$, $\Omega=40^\circ$, $\omega=60^\circ$, $\nu=0^\circ$. - $p = a(1-e^2) = 8000(0.99) = 7920$ km. - $r = p/(1+e) = 7920/1.1 = 7200$ km $= a(1-e)$ ✓ (periapsis distance). - $\vec r_{pqw} = (7200, 0, 0)$; $\vec v_{pqw} = \sqrt{\mu/p}\,(0, 1.1, 0)$, aur $\sqrt{398600/7920} = 7.094$, toh $\vec v_{pqw} = (0, 7.803, 0)$. ![[deepdives/dd-physics-3.2.17-d2-s06.png]] > [!example] Final inertial vectors (Example 2) > $\tilde R$ ko $\Omega=40^\circ, i=30^\circ, \omega=60^\circ$ ke saath $\vec r_{pqw}=(7200,0,0)$ aur $\vec v_{pqw}=(0,7.803,0)$ par apply karne se milta hai (rounded): > $$\vec r \approx (-2748,\ 5670,\ 3118)\ \text{km}, \qquad \vec v \approx (-6.290,\ -2.567,\ 3.379)\ \text{km/s}$$ > Sanity check: $|\vec r| = 7200$ km (rotation length preserve karta hai ✓) aur $\vec r \cdot \vec v = 0$ (periapsis par, motion position ke perpendicular hai ✓). Donon neeche VERIFY block mein machine-checked hain. --- ## Step 7 — Conics ka poora spectrum aur degenerate cases **KYA.** Eccentricity $e$ decide karta hai *kaunsi conic* orbit hai, aur do element *labels* apna meaning kho sakte hain. Donon matter karte hain — kabhi assume mat karo "ellipse, tilted, non-circular." **Shape se ($e$):** - $e = 0$ — **circle**. Koi closest point exist nahi karta, toh $\hat p$ ke paas aim karne ke liye kuch nahi. - $0 < e < 1$ — **ellipse** (woh case jo humne draw kiya). $a>0$, khud par close hoti hai. - $e = 1$ — **parabola**. Orbit kabhi close nahi hoti; $a \to \infty$ aur $p = h^2/\mu$ finite rehta hai, toh hum ise $p$ se describe karte hain, $a$ se nahi. $\vec r_{pqw},\vec v_{pqw}$ ke perifocal formulas $p$ use karke phir bhi kaam karte hain. - $e > 1$ — **hyperbola** (fly-by/escape). Yahan $a<0$ (parent ka $a=-\mu/2\varepsilon$ energy $\varepsilon>0$ ke saath automatically yeh deta hai). Wahi perifocal vectors apply hote hain, sirf us $\nu$ ke angles ke liye valid jahan $1+e\cos\nu>0$. **Orientation se (labels jo gayab ho jaate hain):** - $e = 0$ (circular) — $\omega$ (periapsis *se* measure kiya hua) undefined hai, kyunki measure karne ke liye koi periapsis nahi hai. Iske bajaye ==argument of latitude== $u = \omega + \nu$ use karo jo seedha ascending node se measure hota hai. - $i = 0$ (**prograde equatorial**) — plane flat lie karta hai, donon planes kabhi cross nahi karte, toh node vector $\vec n = \hat z \times \vec h$ zero tak shrink ho jaata hai aur $\Omega$ undefined hai. Iske bajaye true longitude of periapsis use karo. - $i = 180^\circ$ (**retrograde equatorial**) — bhi flat hai, orbit *ulti* chalti hai; phir se $\vec n \to 0$ aur $\Omega$ yahan bhi undefined hai, ek case jise log often bhool jaate hain. - $e = 0$ **aur** $i = 0$ donon saath — *donon* $\omega$ aur $\Omega$ gayab ho jaate hain; sirf true longitude $\ell = \Omega + \omega + \nu$ ek meaningful angle ke roop mein bachti hai. **PICTURES PHIR BHI KYUN HOLD KARTE HAIN.** Har case mein body ki *real position aur velocity* hoti hai — physics theek hai. Sirf chosen *angle labels* apna reference arrow kho dete hain. Ilaaj yeh hai ki kisi aise cheez se measure karo jo abhi bhi exist karta hai: argument of latitude, ya true longitude — yaani [[Orbital singularities and equinoctial elements|equinoctial elements]] par switch karo. **PICTURE.** Left: ek circle — $\omega$ ke liye koi periapsis arrow nahi. Middle: ek flat equatorial ellipse — $\Omega$ ke liye koi crossing line nahi ($i=0$ **aur** $i=180^\circ$ donon ke liye hold karta hai). Right: parabola vs hyperbola — open, non-closing conics. ![[deepdives/dd-physics-3.2.17-d2-s07.png]] > [!mistake] "Yeh ek bug hai!" > $\omega$ ya $\Omega$ mein NaN *lagta hai* broken code. Aisa hai nahi — woh number us geometry ke liye kabhi define hi nahi tha. $e \approx 0$, $i \approx 0$, ya $i \approx 180^\circ$ detect karo aur alternate element par swap karo; arccos ko "patch" mat karo. --- ## Ek-picture summary Sab kuch ek canvas par: ellipse ko perifocal coordinates mein flat draw karo ($\vec r_{pqw}, \vec v_{pqw}$ lo), phir combined matrix $\tilde R$ se $\omega \to i \to \Omega$ spin karo aur inertial frame mein land karo. ![[deepdives/dd-physics-3.2.17-d2-s08.png]] > [!recall]- Feynman: poora walkthrough simple words mein > Socho ki orbit ek sheet of paper par ek squashed ring ki drawing hai, jisme planet ek focus par pin kiya gaya hai aur ring ka near point seedha right ki taraf aim kar raha hai. Is aasaan sheet par main instantly bol sakta hoon spacecraft kahan hai — bas near point se angle $\nu$ walk karo — aur kitna fast, kyunki near point par woh seedha sideways apni top speed par fly kar rahi hai aur mujhe pata hai area-sweep rate kabhi nahi badlti. Toh mujhe paper par do arrows milte hain: position aur velocity. Ab problem: real space meri paper ki tarah flat nahi hai. Toh main paper ko teen baar spin karta hoon — pehle ise apne plane mein turn karo taaki near point sahi taraf aim kare ($\omega$), phir paper ko us line ke baare mein angle par tip karo jahan donon planes milte hain ($i$), phir poori tipped cheez ko lazy-Susan ki tarah sahi compass direction face karne ke liye swing karo ($\Omega$). Donon arrows paper ke saath ride karte hain. Ek matrix $\tilde R$ teeno spins ek saath karta hai. Do warnings: main *frame* spin karta hoon, arrows nahi, toh har spin angle minus sign le leta hai; aur agar ring ek perfect circle hai, perfectly flat lie karti hai (either forward ya backward), ya ek open parabola/hyperbola hai, toh chhe labels mein se kuch simply apply nahi hote — main kisi aur cheez se measure karta hoon jiske paas reference abhi bhi ho.