Dimensional analysis — Buckingham π theorem: Visual walkthrough

2.2.26 · D2 · Physics › Fluid Mechanics › Dimensional analysis — Buckingham π theorem

Step 1 — "Dimensions" actually kya hote hain (teen counters ka ek bag)

KYA HAI. Hum speed jaisi quantity lete hain aur uska number hataa dete hain, sirf uska $M, L, T$ ka recipe rakhte hain.

KYUN. Poora theorem units ke cancel hone ke baare mein hai. Cancellation dekhne ke liye pehle units ko aise cheez ke roop mein dekhna zaroori hai jinke exponents aap add aur subtract kar sako — ek bookkeeping jise hum picture par rakh sakte hain.

PICTURE. Figure mein, har quantity teen chote counters hai. Speed $= L T^{- 1}$ mein $L$ -counter $+ 1$ par hai, $T$ -counter $- 1$ par hai, $M$ -counter $0$ par hai. Force $F = M L T^{- 2}$ teenon ko light up karta hai.

Step 2 — Ek quantity exponents ka ek arrow hai

KYA HAI. Hum har quantity ko $M$ – $L$ – $T$ space mein ek point/arrow ki tarah plot karte hain.

KYUN. Jab quantities arrows ban jaati hain, toh "quantities combine karna" = "arrows add karna," aur "dimensionless" = "arrow origin par land karta hai." Geometry ab hamare liye reasoning karti hai. Yeh wohi akela trick hai jo poore theorem ko visual banata hai.

PICTURE. Density $ρ = M L^{- 3}$ $(1, - 3, 0)$ ka arrow hai; viscosity $μ = M L^{- 1} T^{- 1}$ hai $(1, - 1, - 1)$ . Jab tum quantities multiply karte ho toh unke arrows add hote hain; jab tum power raise karte ho toh arrow stretch hota hai.

Step 3 — "Dimensionless" ka matlab hai arrows origin par cancel ho jaayein

KYA HAI. Hum $q_{1}^{a_{1}} q_{2}^{a_{2}} \dots$ ka aisa combination dhundhte hain jiske exponent-arrows sum karke zero ho jaayein.

KYUN. Yahi woh object hai jise theorem count karta hai. Ek dimensionless group units ki kisi bhi change mein — metres se miles, grams se pounds — unchanged rehta hai. Woh invariance (Dimensional homogeneity) hi reason hai ki physics unse likhi ja sakti hai.

PICTURE. $ρ, v, d$ aur $F$ lein unknown powers $a, b, c, 1$ ke saath. Har ek ek arrow hai; hum unhe scale aur add karte hain. Hum pooch rahe hain: kaunse scalings se chaaron arrows nose-to-tail wapas origin par return karenge?

Step 4 — Arrows stack karo: dimension matrix

KYA HAI. Hum saare $n$ arrows ko side by side ek $k \times n$ table of exponents ke columns ke roop mein line up karte hain.

KYUN. Step 3 ki zero-arrow condition, saare variables ke liye ek saath, ek simple statement ban jaati hai: $D a = 0$ , jahan $a$ unknown powers ko hold karta hai. Ab ek matrix har possible dimensionless group encode karti hai. Ab hum unhe count kar sakte hain.

PICTURE. Sphere-drag variables $F, d, v, ρ, μ$ ke liye grid $3 \times 5$ hai. Har column teen numbers ka ek stalk hai; poori cheez problem ki ek numeric fingerprint hai.

Step 5 — Rank: arrows actually kitni independent directions fill karte hain

KYA HAI. Hum poochte hain ki $M, L, T$ rows mein se kitni nayi information carry karti hain versus kisi doosre ki copy/combination hain.

KYUN. Har independent row $a$ par ek hard constraint hai — ek equation jise powers ko maanna hi padega. Real constraints ki sankhya rank hai, na sirf "kitni rows humne likhi." (Yeh parent mein Mistake D ka fix hai: $k = rank (D)$ use karo, na ki "3 kyunki universe mein 3 hain.")

PICTURE. Teen constraint equations powers $a$ ke space mein origin se guzarne waale teen planes hain. Agar teenon planes sach mein alag hain (rank 3), toh woh space ko hard slice karte hain. Agar do coincide ho jaayein (ek dependent dimension), toh tumhare paas kam real planes hain.

Step 6 — Survivors count karna: $p = n - rank (D)$

KYA HAI. Hum count karte hain ki rank-many planes ke slicing ke baad kitni free directions bachi hain.

KYUN. Yahi punchline hai. Hum $n$ free powers se shuru karte hain (ek har variable ke liye). Har independent constraint ( $rank (D)$ hain) ek power ko baaki ke terms mein pin kar deta hai. Jo free rehta hai woh exactly hai: $p = knobs jisse shuru kiya n - independent constraints rank (D) = n - k .$ Yeh rank–nullity fact hai: (unknowns ki sankhya) − (rank) = (free parameters). Har free parameter ek independent $π$ group hai.

PICTURE. Ek kamra imagine karo. Powers ka full space $n$ -dimensional hai. Constraints ek wall hain dimension $k$ ki jiske against tumhe dhakela jaata hai. Jo freely move karne ke liye bachta hai — floor space jis par tum chal sakte ho — uski dimension $p = n - k$ hai. Sphere drag ke liye: $5 - 3 = 2$ .

Step 7 — Edge cases: jab ek dimension balance nahi ho sakta (aur rank drop ho jaata hai)

Case A — ek akela dimension (pendulum ka mass). Pendulum problem mein variables hain $T, L, g, m$ , sirf $m$ hi $M$ carry karta hai. Dimension matrix mein $M$ -row hai $(0, 0, 0, 1)$ — mass akela nonzero entry hai. Group ke dimensionless hone ke liye humein $a_{m} = 0$ chahiye: mass appear nahi kar sakta. Isliye $T = 2 π L / g$ mein koi $m$ nahi hai. (Parent Mistake A.)

Case B — fake repeaters (rank collapse). Agar tum aisi repeating variables choose karo jo secretly dependent hain — jaise speed $v$ , diameter $d$ , aur time $t$ jahan $t = d / v$ — toh unke columns independent nahi hain, isliye rank tumhari chosen columns ki sankhya se neeche drop ho jaati hai. Do constraint-planes coincide ho jaate hain, null space barta hai, aur tumhare "π groups" independent nahi honge. (Parent Mistake B.) Fix: repeaters ki apni sub-matrix invertible honi chahiye.

KYA / KYUN. Hum un do tareekon ko inspect karte hain jismein picture degenerate hoti hai: ek aisi dimension jisme sirf ek carrier ho (ek power ko zero force karti hai) aur ek linearly dependent variable set (rank drop karti hai).

PICTURE. Left panel: $M$ -row $m$ par ek single spike hai → uski power $0$ hone ke liye forced hai, ek padlocked axis ke roop mein drawn. Right panel: teen arrows jo coplanar hain (dependent) — woh 3D fill nahi kar sakte, isliye extra solutions ka ek poora plane appear hota hai.

Ek picture mein poora summary

Recall Feynman retelling — walkthrough simple words mein

Har quantity secretly ek recipe hai: mass ki itni pinches, length ki itni, time ki itni. Uss pinch-count ko ek arrow likhao. Quantities multiply karna bas arrows add karna hai; ek unitless combo woh hai jiske arrows perfectly cancel hokar starting point par wapas aa jaayein — har cheez ki zero pinches. Ab apne saare variable-arrows ko ek grid ke columns ke roop mein line up karo. "Kaunse combos cancel hote hain?" poochna grid se ek clean question poochna hai, aur linear algebra uska jawab deta hai: tumne $n$ dials se shuru kiya tha, har independent dimension ek dial jam kar deti hai, aur $k$ dimensions sach mein bite karti hain, isliye $n - k$ dials free ghoomne ke liye rehti hain. Woh free dials tumhare $π$ groups hain. Agar ek dimension sirf ek variable mein show up kare, toh us variable ki dial zero par jam jaati hai — woh appear nahi ho sakta (alvida, pendulum mass). Agar tumhari chosen "building block" variables mein se do secretly same recipe hain, toh ek dial kabhi independent thi hi nahi, grid ek bite khoti hai, aur tumhare groups double up ho jaate hain. Rank se honestly count karo, aur universe ne tumhare liye jinne knobs chhoode hain woh exactly $p = n - k$ hain.

Recall Quick self-check

Har independent dimension exactly ek free power kyun remove karti hai? ::: Yeh ek independent equation $D a = 0$ row hai; rank–nullity kehta hai free powers $= n - (independent rows) = n - k$ . Sphere-drag matrix mein $rank (D)$ aur $p$ kya hai? ::: $rank (D) = 3$ , isliye $p = 5 - 3 = 2$ (drag coefficient aur Reynolds number). "Dimensionless group" hona geometrically kya hai? ::: Scaled exponent-arrows nose-to-tail sum karke origin $(0, 0, 0)$ par wapas aa jaana. Pendulum ka mass kyun appear nahi kar sakta? ::: $M$ -row sirf $m$ column mein nonzero hai, isliye woh power $0$ hone ke liye forced hai.

Connections

2.2.26 Dimensional analysis — Buckingham π theorem (Hinglish) — parent topic jise yeh walkthrough derive karta hai
Reynolds number — Step 3–4 se nikla dusra $π$ group
Drag force and drag coefficient — $π_{1} = F / ρ v^{2} d^{2}$ exactly $C_{D}$ hai
Dimensional homogeneity — isliye zero-arrow condition hi poora game hai
Fundamental and derived units — jahan se $M, L, T$ axes aate hain
Model testing and similarity — model aur prototype ke beech $p$ groups match karna
Navier–Stokes equations — unhe non-dimensionalise karne par rigorously $R e$ milta hai