3.5.47 · D3 · Physics › Guidance, Navigation & Control (GNC) › Attitude control modes — spin stabilization, 3-axis active
Intuition Yeh page kya hai
Parent note Attitude control modes — spin stabilization, 3-axis active ne tumhe machinery di thi: Euler's equations , gyroscopic precession, reaction wheels, aur PID control. Yahan hum ise drill karte hain. Hum ek scenario matrix banate hain — har tarah ke problem ki ek checklist — phir har cell ka ek example solve karte hain taaki koi bhi exam question surprise na kar sake.
Shuru karne se pehle, symbols ko samajhna zaroori hai taaki neeche ki har line clean lage.
Definition Core quantities jo hum baar baar use karte hain
I (moment of inertia, units kg⋅m 2 ) — "rotational heaviness": body ki spin change karna kitna mushkil hai. Bada I = modnay mein sluggish. Dekho Moment of inertia tensor & principal axes .
I 1 , I 2 , I 3 (principal moments of inertia ) — jab hum measuring axes ko body ke natural symmetry axes (principal axes ) ke saath align karte hain, to inertia "tensor" sirf teen numbers mein collapse ho jata hai: I 1 pehli body axis ke baare mein, I 2 doosri ke baare mein, I 3 teesri ke baare mein. Inhe teen alag "rotational heavinesses" socho — ek har axis ke liye. Dekho Moment of inertia tensor & principal axes .
ω (angular velocity, units rad/s ) — koi cheez kitni tez ghoom rahi hai. Ek poora chakkar 2 π ≈ 6.283 radians hota hai.
L = I ω (angular momentum, units kg⋅m 2 / s ) — poori body ka "spin momentum." Seedhi-line motion mein ordinary momentum m v jo role play karta hai, wahi role yahan L play karta hai.
h (wheel angular momentum, units kg⋅m 2 / s = N⋅m⋅s ) — spin momentum jo reaction wheel ke andar stored hota hai — ek alag spinning flywheel jo body se judi hoti hai. h (wheel) aur L (poori body) ko alag rakho: wheel spin up hoti hai (h badhta hai) aur body counter-rotate karti hai taaki total conserved rahe. Dekho Reaction wheels and Control-Moment Gyros (CMG) .
Definition Rotation angle, uski rates, aur
α ka matlab
θ (rotation angle, units rad ) — body apne target orientation se kitna door ghoom gayi hai kisi axis ke baare mein. Compass ki needle ka sochte hain: θ wo angle hai jo usne sweep kiya.
θ ˙ (dot matlab "time mein rate of change") exactly us axis ke baare mein angular velocity hai, rad/s mein.
θ ¨ (do dots = rate of change of the rate) angular acceleration hai, rad/s 2 mein — spin kitni tezi se speed up ya slow down hoti hai. Hum ise apna single-letter naam α dete hain: α ≡ θ ¨ . To "α " aur "θ ¨ " ek hi cheez hain; α tab use karte hain jab likhna zyada clean ho. Yeh ordinary acceleration ka rotational cousin hai.
Definition Torque, precession
vector , aur cross product
τ (torque, units N⋅m ) — ek twist : push ka rotational version.
Newton ka rotational law: τ = d t d L — ek twist spin momentum ko change karta hai.
Ω (precession rate, ek vector , units rad/s ) — us axis ki taraf point karta hai jiske around spin momentum L dheere dheere cone karta hai. Upar ka chhota arrow ( ) matlab "iska space mein ek direction hai," sirf ek size nahi.
Ω × L cross product hai: yeh Ω aur L dono ke perpendicular ek naya vector produce karta hai, jiska size Ω L sin ( angle between them ) hai. Uski direction right-hand rule se milti hai (fingers Ω se L ki taraf, thumb result deta hai).
Precession law hai τ = Ω × L : ek sideways torque L ko spin up nahi karta, balki use cone karta hai. Sizes lene par (jab Ω ⊥ L ) scalar shortcut milta hai ==Ω = τ / L ==. Ek jhukti hui spinning top ka socho jo girne ki jagah dheere dheere circle trace karti hai: Gyroscopic precession .
Definition Control-law symbols (PD examples mein use hote hain)
Jab ek spacecraft attitude error θ ko actively correct karta hai, to uska onboard computer ek corrective torque ek control law se compute karta hai. Sabse simple hai proportional–derivative (PD): τ = − K p θ − K d θ ˙ .
K p (proportional gain , units N⋅m/rad ) — target ki taraf spring jaisi pull: jitna bada error angle θ , utna zyada push back.
K d (derivative gain , units N⋅m⋅s/rad ) — speed θ ˙ ke against shock-absorber jaisi resistance: wobble dampen karta hai taaki craft forever overshoot na kare.
ω n = K p / I (natural frequency , rad/s ) — corrected axis kitni tezi se oscillate karna chahta hai.
ζ = 2 K p I K d (damping ratio , dimensionless) — response kitna "cushioned" hai. ζ < 1 ring karta hai (under-damped), ζ = 1 critically damped hai, ζ > 1 dheere dheere crawl in karta hai (over-damped). Sweet spot hai ζ ≈ 0.7 . Dekho PID control & second-order systems (damping ratio) .
Is chapter ka har problem in cells mein se ek (ya blend) hai:
#
Cell class
Kya change hota hai / "sign or edge"
Example
A
Gyroscopic stiffness, normal case
positive spin, sideways torque
Ex 1
B
Zero-spin limit
ω s → 0 : stiffness vanish ho jaati hai
Ex 2
C
Spin-axis choice (max vs min I )
kaunsa principal axis stable hai
Ex 3
D
Degenerate inertia (I 1 = I 2 )
symmetric body, coupling term die ho jata hai
Ex 4
E
Reaction-wheel slew (bang-bang)
accelerate phir brake, sign flip
Ex 5
F
Wheel saturation / desaturation
wheel max RPM hit karta hai, external dump
Ex 6
G
PD control tuning (damping ratio)
choose ζ : under/critical/over
Ex 7
H
Real-world word problem
disturbance budget over an orbit
Ex 8
I
Exam twist (sign / quadrant trap)
precession direction by right-hand rule
Ex 9
Hum saare nine solve karte hain taaki har cell filled ho.
Worked example Ex 1 — Ek spinner kitni tezi se drift karta hai?
Ek drum satellite ka I = 200 kg⋅m 2 hai aur wo ω s = 10 rad/s par spin karta hai. Ek steady sideways disturbance torque τ = 0.02 N⋅m act karta hai. Precession (drift) rate nikalo aur comment karo.
Forecast: guess karo — kya wo seconds mein tumble karega, ya ghanton mein drift karega? Apna gut number likh lo.
Spin momentum L = I ω s = 200 × 10 = 2000 kg⋅m 2 / s .
Yeh step kyun? L "stiffness reservoir" hai; jitna bada, utna zyada sideways twist use thoda sa deflect karta hai.
Precession rate Ω = L τ = 2000 0.02 = 1 × 1 0 − 5 rad/s .
Yeh step kyun? Precession law τ = Ω × L se (upar define kiya gaya), sideways τ L ke perpendicular hai, isliye τ = Ω L aur Ω = τ / L — yeh L ko cone karta hai, body ko spin up nahi karta.
Convert karo: 1 × 1 0 − 5 rad/s × π 180 × 3600 ≈ 2.06 deg/hour .
Verify karo: units kg⋅m 2 / s N⋅m = kg⋅m 2 / s kg⋅m 2 / s 2 = 1/ s ✓ (ek rate). Lagbhag 2 deg/hr — extremely stable, yehi wajah hai ki spinners saste, robust pointing ke liye use hote hain.
Worked example Ex 2 — Agar spin zero ho jaye to kya hoga?
Same drum, same torque τ = 0.02 N⋅m , lekin ab ω s = 0 (ek de-spun, non-rotating body). Same disturbance ke under t = 60 s mein kya hoga?
Forecast: stiffness Ω = τ / L mein L = 0 denominator mein hai — precession blow up karta hai. To "dheere cone karna" picture toot jaani chahiye. Guess karo: kya wo infinitely tezi se cone karta hai, ya kuch bilkul alag hota hai?
ω s = 0 ke saath, L = 0 , to Ω = τ / L → ∞ . Precession formula undefined hai — yeh ab motion describe nahi karta.
Yeh step kyun? Precession assume karta hai ek bada existing L jo cone ho. Agar kuch nahi hai, to cone karne ke liye kuch hai hi nahi.
Iske bajay raw law use karo τ = I θ ¨ (koi gyroscopic term nahi kyunki ω = 0 ), jahan θ upar define kiya gaya turn angle hai. Angular acceleration θ ¨ = τ / I = 0.02/200 = 1 × 1 0 − 4 rad/s 2 .
Yeh step kyun? Zero spin par Euler equation cross-terms ( … ) ω j ω k vanish ho jaate hain; sirf I ω ˙ = I θ ¨ bachta hai.
Yeh simply accelerate aur tumble karta hai: θ = 2 1 θ ¨ t 2 = 2 1 ( 1 × 1 0 − 4 ) ( 60 ) 2 = 0.18 rad ≈ 10.3° ek minute mein, aur badhta raha.
Verify karo: dono regimes spirit mein agree karte hain — bada L ⇒ tiny drift (Ex 1); zero L ⇒ runaway tumble (yahan). "Stiffness" sach mein L ke proportional hai. Yehi wajah hai ki non-spinning craft ko zaroor 3-axis active control chalana padhta hai.
Worked example Ex 3 — Kaunsa axis dissipation survive karta hai?
Ek cylinder ka I a = 50 kg⋅m 2 apne long (symmetry) axis ke baare mein aur I t = 120 kg⋅m 2 ek transverse axis ke baare mein hai. Isme fuel slosh (energy dissipation) hai. Kaunsa axis stable spin axis hai?
Figure caption: Cylinder ka ek rectangular side-view. Body ki length ke along padi black dashed double-arrow long (symmetry) axis hai, jiska moment of inertia chhota value I a = 50 kg⋅m 2 hai — "MIN, unstable" label kiya. Beech se guzarta single red vertical arrow transverse axis hai, jiska moment of inertia bada value I t = 120 kg⋅m 2 hai — "MAX, STABLE" label kiya. Red arrow us ek axis ko mark karta hai jis par body energy dissipate hone par settle karegi; black dashed axis tempting-but-wrong choice hai.
Forecast: intuition kehta hai "lamba pencil-jaisa axis sabse zyada stable lagta hai." Likh lo — phir dekho wo galat nikalta hai.
Figure par do candidate axes identify karo: black dashed arrow long (symmetry) axis I a = 50 hai, aur red vertical arrow transverse axis I t = 120 hai. Fixed L par, kinetic energy T = 2 I L 2 . Same L par dono axes compare karo (maano L = 100 ).
Yeh step kyun? Dissipation (slosh, flexing) energy remove karta hai lekin L nahi change kar sakta (koi external torque nahi). Body minimum T par drift karta hai, isliye hum axes ko unki energy ke hisaab se rank karte hain.
Long axis ke baare mein: T a = 2 ( 50 ) 10 0 2 = 100 J . Transverse ke baare mein: T t = 2 ( 120 ) 10 0 2 = 41.7 J .
Yeh step kyun? Denominator mein bada I ⇒ chhoti energy. Max-I axis (figure mein red arrow) minimum-energy state hai.
Minimum energy = T t , transverse (max-I ) axis — red arrow. Wahi ek dissipation-stable choice hai; black dashed long axis unstable hai.
Verify karo: T t = 41.7 < T a = 100 ✓. Yeh Explorer 1 flat-spin anomaly ek line mein hai: Explorer 1 apne minimum -I (pencil) axis ke baare mein spin kiya gaya, energy dissipate ki, aur flat-spin ho gaya. Fix rule: I m a x ke baare mein spin karo.
Worked example Ex 4 — Symmetric body, torque-free
Ek perfectly symmetric spinner ke principal moments I 1 = I 2 = 100 kg⋅m 2 (equal transverse) aur I 3 = 60 (spin axis) hain, upar define ki gayi I 1 , I 2 , I 3 principal-axis notation use karte hue. Koi external torque nahi. Euler equations use karke, spin-axis rate ω 3 ka kya hota hai?
Forecast: kya ω 3 constant rehta hai, drift karta hai, ya oscillate karta hai?
Teesra Euler equation (torque-free): I 3 ω ˙ 3 = ( I 1 − I 2 ) ω 1 ω 2 .
Yeh step kyun? Axis 3 par gyroscopic coupling difference I 1 − I 2 se drive hota hai.
Degeneracy: I 1 = I 2 ⇒ I 1 − I 2 = 0 , to ω ˙ 3 = 0 .
Yeh step kyun? Jab do moments of inertia equal hain, to wo coupling term exactly zero ho jata hai — case class "collapse" ho jaata hai.
Isliye ω 3 = const : symmetry axis ke baare mein spin rate exactly conserved hai, chahe ω 1 , ω 2 kuch bhi karein.
Verify karo: ( I 1 − I 2 ) ω 1 ω 2 = 0 ⋅ ω 1 ω 2 = 0 ✓. Yehi wajah hai ki real spinners axially symmetric banaye jaate hain — ek guaranteed-constant spin rate milti hai aur clean precession behaviour.
Worked example Ex 5 — 30° turn karne mein kitna time?
I = 500 kg⋅m 2 wale ek body ko 30° rotate karna hai. Ek reaction wheel constant torque magnitude τ = 0.1 N⋅m deta hai: pehle half ke liye accelerate, phir brake (sign reverse) doosre half ke liye taaki wo rest par pahunche. Total slew time aur peak wheel momentum nikalo. Dekho Reaction wheels and Control-Moment Gyros (CMG) .
Forecast: total time nearest 30 s tak guess karo.
Angular acceleration α = θ ¨ = τ / I = 0.1/500 = 2 × 1 0 − 4 rad/s 2 (yaad karo α ≡ θ ¨ , upar define kiya).
Yeh step kyun? Wheel torque body par momentum exchange ke zariye act karta hai: τ body = − h ˙ , jahan h upar define kiya gaya wheel momentum hai.
Half-angle = 15° = 0.262 rad . θ = 2 1 α t 2 use karke: t = 2 θ / α = 2 ( 0.262 ) / ( 2 × 1 0 − 4 ) ≈ 51.2 s .
Yeh step kyun? Constant-torque kinematics turn angle θ ke saath; accelerate phase exactly aadha angle cover karta hai.
Total time = 2 × 51.2 ≈ 102.4 s . Midpoint par sign flip "bang-bang" hai: maximum push, phir maximum brake.
Peak wheel momentum h = I α t = 500 ( 2 × 1 0 − 4 ) ( 51.2 ) ≈ 5.12 N⋅m⋅s .
Yeh step kyun? Yeh mid-slew angular momentum hai: body kuch rate tak spin up ho gayi hai, aur conservation se wheel equal-and-opposite h hold karta hai — jo wheel ki rating se neeche hona chahiye.
Verify karo: midpoint par body rate θ ˙ = α t = 2 × 1 0 − 4 × 51.2 = 0.01024 rad/s ; body momentum I θ ˙ = 500 × 0.01024 = 5.12 N⋅m⋅s ✓ step 4 se match karta hai.
Worked example Ex 6 — Momentum dump kab karna hoga?
Ex 5 ka wheel maximum storable momentum h m a x = 4 N⋅m⋅s rakhta hai. Ek steady disturbance torque τ d = 5 × 1 0 − 4 N⋅m wheel ke zariye absorb ho raha hai (attitude hold karne ke liye use spin up kar raha hai). Saturation tak kitna time, aur kya dump karta hai? Dekho Momentum dumping / desaturation with magnetorquers .
Forecast: ghante? din? ek orbit?
Wheel momentum rate h ˙ = τ d = 5 × 1 0 − 4 N⋅m⋅s per s par absorb karta hai, jahan h wheel momentum hai.
Yeh step kyun? Constant external twist ke against attitude hold karne ke liye, wheel ko wo momentum lete rehna hai (total L = L body + h internally conserved hai).
Bharne ka time: t sat = h m a x / h ˙ = 4/ ( 5 × 1 0 − 4 ) = 8000 s ≈ 2.22 hours .
Yeh step kyun? Constant absorption ⇒ h linearly apni ceiling h m a x tak bharta hai.
Dhyan do Ex 5 ka peak slew momentum 5.12 N⋅m⋅s > 4 N⋅m⋅s tha — akela wo slew is wheel ko already saturate kar deta. Wheel ek one-shot 30° slew ke liye undersized hai.
Yeh step kyun? Hamesha slew demand ko rating ke against cross-check karo.
Desaturation: magnetorquers/thrusters fire karo taaki ek external torque apply ho jo wheel ko wapas spin karae (h gire) jabki ek equal external torque body ko hold kare — stored momentum Earth ke field / propellant mein dump karo.
Yeh step kyun? Sirf ek external torque total L change kar sakta hai; wheels akele khud ko un-saturate nahi kar sakte.
Verify karo: units N⋅m N⋅m⋅s = s ✓. 8000 s /3600 = 2.22 hr ✓.
Worked example Ex 7 — Achhi settling ke liye derivative gain choose karo
Ek single axis I θ ¨ = τ ko obey karta hai jahan I = 500 kg⋅m 2 , aur θ upar define kiya gaya attitude error angle hai. PD law τ = − K p θ − K d θ ˙ use ki ja rahi hai K p = 20 N⋅m/rad ke saath. Damping ratio ζ = 0.7 (fast, no ringing) ke liye K d choose karo. K p , K d , ω n , ζ sab upar control-law definition mein define hain. Dekho PID control & second-order systems (damping ratio) .
Forecast: K d nearest 50 tak guess karo.
PD law ko I θ ¨ = τ mein substitute karo: I θ ¨ = − K p θ − K d θ ˙ , yani I θ ¨ + K d θ ˙ + K p θ = 0 — angle θ mein ek damped oscillator , jahan ω n = K p / I aur ζ = 2 K p I K d .
Yeh step kyun? Small-angle linearization gyroscopic terms drop kar deta hai; jo bachta hai wo classic second-order system hai jiske do knobs K p , K d exactly ω n aur ζ set karte hain.
Natural frequency ω n = K p / I = 20/500 = 0.04 = 0.2 rad/s .
Yeh step kyun? ω n set karta hai ki axis kitni tezi se respond karna chahta hai; hum ise pehle compute karte hain kyunki ζ same K p , I par depend karta hai.
ζ = 2 K p I K d ko K d ke liye solve karo: K d = 2 ζ K p I = 2 ( 0.7 ) 20 × 500 = 1.4 10000 = 1.4 × 100 = 140 N⋅m⋅s/rad .
Yeh step kyun? ζ = 0.7 sweet spot hai: near-fastest settling with sirf ek chhoti, brief overshoot — standard aerospace pick.
Recommendation: K d = 140 N⋅m⋅s/rad set karo (with K p = 20 ). Isse ω n = 0.2 rad/s aur ζ = 0.7 milta hai: quick, well-damped pointing.
Verify karo: back-substitute: ζ = 2 20 × 500 140 = 2 × 100 140 = 0.7 ✓. Agar K d chhota hota (ζ < 1 ) to ring karta; bada (ζ > 1 ) to dheere dheere crawl in karta — 0.7 dono ko balance karta hai.
Worked example Ex 8 — Ek orbit mein solar-pressure drift
Ek 3-axis craft ko ek full orbit T orb = 5400 s (90 min) ke liye ek steady solar-radiation-pressure torque τ SRP = 3 × 1 0 − 5 N⋅m ek axis ke baare mein suffer hota hai. Uske wheel mein h m a x = 0.2 N⋅m⋅s store ho sakta hai. Kya dump ki zaroorat se pehle wo ek orbit survive kar sakta hai? Dekho Gravity-gradient & solar-radiation disturbance torques .
Forecast: kya ek orbit ka push wheel ke andar fit hota hai?
Orbit ke dauran accumulated momentum: Δ h = τ SRP T orb = 3 × 1 0 − 5 × 5400 = 0.162 N⋅m⋅s , wheel momentum h mein jodega.
Yeh step kyun? Constant torque ⇒ momentum linearly badhta hai; body ko still rakhne ke liye wheel sab kuch absorb karta hai.
Capacity se compare karo: 0.162 < 0.2 , to wheel hold karta hai — bas barely (81% full).
Yeh step kyun? Mission planners is margin ko dhyan mein rakhte hue har orbit mein ek desaturation schedule karte hain.
Margin: 0.2 − 0.162 = 0.038 N⋅m⋅s spare, yani 0.038/ ( 3 × 1 0 − 5 ) ≈ 1267 s (≈ 21 min ) extra runway saturate hone se pehle.
Verify karo: 0.162/0.2 = 0.81 ✓. Units N⋅m × s = N⋅m⋅s ✓. Realistic: ∼ 1 0 − 5 N⋅m ke SRP torques chote craft par lagbhag har orbit mein ek dump force karte hain.
L kis taraf cone karta hai?
Ek spinner ka spin momentum L "upar" + z ^ ki taraf point karta hai (spin axis vertical). Ek disturbance torque sideways + x ^ ki taraf point karta hai. τ = Ω × L use karke, precession vector Ω ki direction nikalo (trap: students expect karte hain ki L torque ki taraf move karega — wo nahi karta). Figure dekho.
Figure caption: Do black arrows givens dikhate hain — L upar + z ^ ki taraf (spin axis) aur τ daayein + x ^ ki taraf (disturbance torque). Single red arrow answer hai: precession vector Ω + y ^ ki taraf point karta hai (upar-left ki taraf), exactly 90° torque se. Picture ka poora point yeh hai ki red response arrow black torque arrow ke saath aligned nahi hai.
Forecast: guess karo: kya Ω + y ^ , − y ^ , ya torque + x ^ ki taraf point karta hai? Likh lo.
Figure se known vectors likho: L = L z ^ (black up-arrow) aur τ = τ x ^ (black right-arrow). Hume Ω dhundha hai jo τ = Ω × L satisfy kare.
Yeh step kyun? Precession vector is cross-product relation se define hota hai, isliye direction — sirf size nahi — solve karna hoga.
Ek direction guess karo aur right-hand rule se test karo. Try Ω = Ω y ^ : to Ω × L = Ω L ( y ^ × z ^ ) = Ω L x ^ .
Yeh step kyun? Right-hand rule deta hai y ^ × z ^ = x ^ ; hum check karte hain ki yeh given torque direction + x ^ reproduce karta hai ya nahi.
Yeh exactly + x ^ se match karta hai jab Ω > 0 . Isliye Ω + y ^ ki taraf point karta hai — figure mein red arrow . Spin axis + y ^ ki taraf cone karta hai, torque se 90° par, torque ki direction + x ^ mein nahi .
Yeh step kyun? Yeh perpendicular response poora gyroscopic surprise hai: spinning top ko sideways push karo aur wo apni push ke right angles par move karta hai.
Magnitude, as always: Ω = τ / L (Ω ⊥ L hone se sizes lete hue). Direction + y ^ , size τ / L — poora vector answer.
Yeh step kyun? Exam dono chahta hai — number aur correct axis; sirf Ω = ( τ / L ) y ^ complete hai.
Verify karo: cross-product check y ^ × z ^ = x ^ ✓ torque axis reproduce karta hai. Magnitude Ω = τ / L Ex 1 se agree karta hai. Gyroscopes hamesha applied twist se 90° door respond karte hain — Gyroscopic precession ka counter-intuitive heart.
Recall Self-test — answers cover karo
Zero-spin limit: kya precession formula Ω = τ / L abhi bhi apply hota hai? ::: Nahi — L = 0 ise undefined banata hai; τ = I θ ¨ use karo aur body tumble karti hai.
Spin ke liye kaunsa axis dissipation-stable hai, min-I ya max-I ? ::: Max-I (fixed L par minimum energy).
Symmetric body I 1 = I 2 ke liye, bina torque ke kya conserved hai? ::: Spin-axis rate ω 3 (coupling term I 1 − I 2 = 0 ).
L aur h mein kya fark hai? ::: L poori body ka angular momentum hai; h ek reaction wheel ke andar stored momentum hai — dono trade karte hain taaki total conserved rahe.
Symbol α ka yahan kya matlab hai? ::: Angular acceleration, α ≡ θ ¨ , rad/s 2 mein.
Wheel saturate kyun hota hai, aur ise kaise fix kiya jaata hai? ::: Yeh disturbance momentum absorb karta hai max RPM tak; ek external torque (magnetorquer/thruster) ise dump karta hai.
ζ = K d / ( 2 K p I ) mein, kaunsa ζ ringing ke bina fast settling deta hai? ::: Lagbhag 0.7 .
Applied torque ke relative precession direction? ::: Perpendicular (90° door), τ = Ω × L se.
Mnemonic Matrix yaad rakhne ki ek line
"Spin stiff, zero tumbles, max-I lives, symmetric holds, bang-bang slews, wheels saturate, ζ = 0.7 , budget the orbit, cone sideways."
Back to parent: Attitude control modes — spin stabilization, 3-axis active .