Home / Courses / GATE Physics / Practice / Partial differential equations and separation of variables

Partial differential equations and separation of variables

Mathematical methods · 84 questions and templates for GATE Physics, in order from warm-up to top rank. Check each answer, then read the full solution.

New to this topic? Read the lesson first: Partial differential equations and separation of variables.

Loading questions… (they need JavaScript; nothing you type is sent anywhere)

Written questions with solutions

The same questions as a list, so you can print them or read without the practice box. Templates (fresh numbers) appear only in the practice box above.

  1. MCQWarm-up

    Laplace's equation uxx+uyy=0 is
    1. elliptic
    2. parabolic
    3. hyperbolic
    4. of mixed type
    Answer and solution

    Answer: A

    For Auxx+Buxy+Cuyy+…=0 the type is set by B2−4AC. Here A=C=1, B=0: B2−4AC=−4<0, elliptic. Elliptic equations describe equilibria (no time direction); data must be given on a closed boundary.

    Why people get this wrong: The type depends only on the second-derivative terms.

  2. MCQWarm-up

    The heat equation ut=κuxx (with κ>0) is
    1. elliptic
    2. hyperbolic
    3. parabolic
    4. not classifiable because it has only one second derivative
    Answer and solution

    Answer: C

    In the variables (x,t): A=κ, B=0, C=0 (there is no utt). B2−4AC=0: parabolic. Parabolic equations smooth things out and have a direction of time.

    Why people get this wrong: Missing a second derivative does not stop classification; it is exactly what makes B2−4AC=0.

  3. MCQWarm-up

    A string of length 0.5 m is fixed at both ends, and transverse waves travel on it at 200 m/s. Its fundamental frequency is
    1. 100 Hz
    2. 400 Hz
    3. 200 Hz
    4. 50 Hz
    Answer and solution

    Answer: C

    Separation gives modes sin⁡(nπx/L) with ωn=nπc/L, so fn=nc2L. For n=1: 2002×0.5=200 Hz.

    Why people get this wrong: Using c/L (400 Hz): the fundamental fits half a wavelength, so λ=2L.

  4. MCQWarm-up

    Separating u(x,t)=X(x)T(t) in ut=κuxx with u(0,t)=u(L,t)=0, the allowed spatial functions are
    1. cos⁡(nπx/L)
    2. enπx/L
    3. sinh⁡(nπx/L)
    4. sin⁡(nπx/L)
    Answer and solution

    Answer: D

    X′′=−k2X with X(0)=X(L)=0. X=sin⁡kx with sin⁡kL=0, so k=nπ/L, n=1,2,… Each mode decays as e−κk2t.

    Why people get this wrong: Cosines satisfy zero slope at the ends (insulated ends), not zero value.

  5. MSQWarm-up

    Which functions satisfy the two-dimensional Laplace equation uxx+uyy=0?
    1. x2−y2
    2. xy
    3. excos⁡y
    4. x2+y2
    Answer and solution

    Answer: A, B, C

    x2−y2: 2−2=0. xy: 0+0=0. excos⁡y: excos⁡y−excos⁡y=0. x2+y2: 2+2=4≠0. (The first three are real parts of analytic functions z2, −i2z2 and ez.)

    Why people get this wrong: x2+y2 looks symmetric, but symmetry is not what Laplace's equation asks for.

  6. MCQWarm-up

    The general solution of the one-dimensional wave equation utt=c2uxx is
    1. f(x−ct)+g(x+ct)
    2. f(x−ct)g(x+ct)
    3. f(x)e−ct
    4. f(x2−c2t2)
    Answer and solution

    Answer: A

    With ξ=x−ct and η=x+ct, the equation becomes uξη=0, so u=f(ξ)+g(η): one wave moving right and one moving left, each keeping its shape.

    Why people get this wrong: A product of the two is not a solution in general.

  7. MCQWarm-up

    The equation ut+uux=0 is
    1. first order and linear
    2. first order and non-linear
    3. second order and linear
    4. second order and non-linear
    Answer and solution

    Answer: B

    The highest derivative is first order. The term uux multiplies the unknown by its own derivative, so the equation is non-linear: the sum of two solutions is not a solution. (This is the inviscid Burgers equation, whose waves steepen into shocks.)

    Why people get this wrong: Having only first derivatives does not make an equation linear; look at how u enters the coefficients.

  8. NATStandard

    A rod of length 1 m with thermal diffusivity κ=1.0×10−4 m²/s has both ends held at 0 °C. The slowest-decaying temperature mode has a decay time (the time for its amplitude to fall by a factor e) of ______ s (to the nearest second).
    Answer and solution

    Answer: 1013 s (±2)

    Mode n is sin⁡(nπx/L)e−κ(nπ/L)2t. The slowest is n=1, with τ=L2π2κ=19.87×10−4≈1013 s, about 17 minutes.

    Why people get this wrong: Forgetting the π2 gives L2/κ=104 s, ten times too long.

  9. MCQStandard

    On a rod with both ends at zero temperature, how much faster does the n=3 mode decay than the n=1 mode?
    1. 3 times
    2. 3 times
    3. 9 times
    4. 27 times
    Answer and solution

    Answer: C

    The decay rate is κ(nπ/L)2∝n2, so the ratio is 32=9. Sharp features (high n) in a temperature profile vanish first; that is why the profile quickly becomes a smooth half-sine.

    Why people get this wrong: Taking the rate proportional to n, as wave frequencies are.

  10. NATStandard

    For k>0, the equation uxx+kuxy+4uyy=0 is parabolic when k = ______.
    Answer and solution

    Answer: 4

    Parabolic means B2−4AC=0: k2−4(1)(4)=0, so k=4. For k<4 it is elliptic, for k>4 hyperbolic.

    Why people get this wrong: Using B2−AC, the form for Auxx+2Buxy+Cuyy (it would give k=2 with a different B). Here the coefficient of uxy is k itself.

  11. MCQStandard

    The equation uxx+4uxy+3uyy=0 is
    1. elliptic
    2. hyperbolic
    3. parabolic
    4. elliptic for x>0 and hyperbolic for x<0
    Answer and solution

    Answer: B

    B2−4AC=16−12=4>0: hyperbolic. It factors as (∂x+∂y)(∂x+3∂y)u=0, with two families of real characteristics.

    Why people get this wrong: The coefficients are constant, so the type cannot change from place to place.

  12. NATStandard

    A string 0.8 m long, fixed at both ends, carries waves at 240 m/s. The frequency of its third harmonic is ______ Hz.
    Answer and solution

    Answer: 450 Hz

    fn=nc2L=3×2401.6=450 Hz.

    Why people get this wrong: The third harmonic is n=3, not the third overtone (n=4).

  13. MCQStandard

    Separating Laplace's equation in plane polar coordinates, the solutions that are finite at r=0 and single-valued in θ are combinations of
    1. r−ncos⁡nθ and r−nsin⁡nθ
    2. rncos⁡nθ and rnsin⁡nθ (plus a constant)
    3. enrcos⁡nθ
    4. Jn(r)cos⁡nθ
    Answer and solution

    Answer: B

    With u=R(r)Θ(θ): Θ′′=−n2Θ and single-valuedness forces integer n. Then r2R′′+rR′−n2R=0 gives R=rn or r−n (and 1, ln⁡r for n=0). Finite at the origin keeps rn and the constant. Bessel functions appear for the Helmholtz equation, not Laplace's.

    Why people get this wrong: r−n and ln⁡r belong to the region outside a circle.

  14. MSQStandard

    A rod has insulated ends (ux=0 at x=0 and x=L). Which statements about solutions of ut=κuxx are true?
    1. The spatial modes are cos⁡(nπx/L), n=0,1,2,…
    2. The n=0 mode is a constant that does not decay
    3. The average temperature along the rod stays constant in time
    4. The temperature tends to zero everywhere
    Answer and solution

    Answer: A, B, C

    Zero slope at both ends picks cosines, including n=0 (a constant), whose decay rate κ(0)2 is zero. ddt∫0Ludx=κ[ux]0L=0: no heat leaves, so the mean is conserved and the rod ends up uniform at its initial mean temperature, not at zero.

    Why people get this wrong: Zero temperature is the end state only when the ends are held at zero.

  15. NATStandard

    Heat spreads in still water (κ=1.4×10−7 m²/s). Over 100 s, the diffusion length κt is ______ mm (to two decimal places).
    Answer and solution

    Answer: 3.74 mm (±0.02)

    1.4×10−7×100=1.4×10−5=3.74×10−3 m =3.74 mm. Diffusion distance grows only as t: four times the distance takes sixteen times as long.

    Why people get this wrong: Treating diffusion like travel at a constant speed (distance proportional to t).

  16. MCQStandard

    A rod has one end held at 0 °C and the other at 100 °C. Long after it is set up, the temperature at its midpoint is
    1. 100/e °C
    2. 25 °C
    3. 50 °C
    4. 0 °C, since the transients decay to zero
    Answer and solution

    Answer: C

    In the steady state ut=0, so uxx=0 and the profile is a straight line: u=100x/L, giving 50 °C at the middle. The transients decay; the steady part stays.

    Why people get this wrong: Only the deviation from the straight line decays to zero.

  17. MCQStandard

    A non-constant solution of Laplace's equation in a bounded region takes its maximum value
    1. at the centre of the region
    2. at a point where the boundary is most curved
    3. anywhere; there is no rule
    4. only on the boundary
    Answer and solution

    Answer: D

    The mean-value property says u at any point equals its average over any circle (sphere) round that point inside the region. A strict interior maximum would be larger than its own average, which is impossible. So maxima and minima of harmonic functions lie on the boundary. That is also why there is no stable electrostatic equilibrium in free space (Earnshaw).

    Why people get this wrong: This holds for Laplace's equation only; with a source (Poisson) an interior maximum is possible.

  18. MCQStandard

    Putting u(x,y)=X(x)Y(y) into Laplace's equation uxx+uyy=0 leads to
    1. X′′X=Y′′Y=constant
    2. X′′=Y′′
    3. X′′Y′′=constant
    4. X′′X=−Y′′Y=constant
    Answer and solution

    Answer: D

    Substituting gives X′′Y+XY′′=0. Dividing by XY: X′′X=−Y′′Y. The left side depends only on x and the right only on y, so both equal the same constant. One factor then oscillates and the other grows or decays (e.g. sin⁡kxsinh⁡ky).

    Why people get this wrong: Dropping the minus sign would make both factors oscillate (or both grow), which does not satisfy Laplace's equation.

  19. NATStandard

    The function u=e−4tsin⁡kx satisfies the heat equation ut=uxx for a positive value of k equal to ______.
    Answer and solution

    Answer: 2

    ut=−4u and uxx=−k2u. Equal when k2=4, so k=2. Shorter waves (larger k) decay faster, as e−k2t.

    Why people get this wrong: Matching 4 with k instead of k2.

  20. NATExam level

    A rod 0≤x≤L starts at a uniform 100 °C, and both ends are then held at 0 °C. In u=∑bnsin⁡(nπx/L)e−κ(nπ/L)2t, the coefficient b1 is ______ °C (to one decimal place).
    Answer and solution

    Answer: 127.3 °C (±0.2)

    bn=2L∫0L100sin⁡nπxLdx=200nπ(1−cos⁡nπ), which is 400nπ for odd n and 0 for even n. So b1=400/π≈127.3 °C. It exceeds 100 because the sine series must overshoot to fit a flat profile with zeros at the ends.

    Why people get this wrong: Taking b1=100: the first sine term alone is not the initial profile.

  21. NATExam level

    For the rod in the previous setting (uniform 100 °C start, ends at 0 °C, b1=400/π), keep only the first term. At the moment when κt/L2=0.1, the midpoint temperature is ______ °C (to one decimal place).
    Answer and solution

    Answer: 47.5 °C (±0.3)

    u(L/2,t)≈400πsin⁡π2e−π2κt/L2=127.3×e−0.987=127.3×0.373≈47.5 °C. The next term, n=3, is −4003πe−9×0.987≈−0.006 °C, so one term is plenty.

    Why people get this wrong: Using e−κt/L2 without the π2.

  22. MCQExam level

    A string fixed at both ends is plucked exactly at its midpoint and released from rest. Which harmonics are missing from its motion?
    1. all odd harmonics
    2. all harmonics above the tenth
    3. none: every harmonic is present
    4. all even harmonics
    Answer and solution

    Answer: D

    The initial shape is symmetric about x=L/2. The even modes sin⁡(2mπx/L) are antisymmetric about the midpoint, so their overlap with the shape is zero. (Equivalently each even mode has a node at the plucking point.) Only odd harmonics sound, with amplitudes falling as 1/n2.

    Why people get this wrong: A mode with a node at the plucking point cannot be excited by plucking there.

  23. MSQExam level

    Solve uxx+uyy=0 in 0<x<a, 0<y<b with u=0 on x=0, x=a and y=0, and u(x,b)=f(x). Which statements are true?
    1. The y-dependence of each term is cosh⁡(nπy/a)
    2. The x-dependence of each term is sin⁡(nπx/a)
    3. The y-dependence of each term is sinh⁡(nπy/a)
    4. The coefficient of term n is fn/sinh⁡(nπb/a), where fn is the nth sine coefficient of f
    Answer and solution

    Answer: B, C, D

    Zero at x=0,a forces X=sin⁡(nπx/a), so Y′′=(nπ/a)2Y. Zero at y=0 picks sinh⁡, not cosh⁡. At y=b: ∑cnsinh⁡(nπb/a)sin⁡(nπx/a)=f(x), so cn=fn/sinh⁡(nπb/a).

    Why people get this wrong: The oscillating direction is the one with two zero conditions; the other gets hyperbolic functions.

  24. MCQExam level

    In a square, one side is held at potential V0 and the other three at 0. The potential at the centre of the square is
    1. V0/2
    2. V0/4
    3. V0/π
    4. 0
    Answer and solution

    Answer: B

    Superposition and symmetry: if each of the four sides in turn were at V0, the sum of the four solutions would be u=V0 everywhere (all sides at V0). By symmetry each contributes equally at the centre, so each gives V0/4. No series needed.

    Why people get this wrong: Averaging the two nearest sides gives V0/2, which ignores the other two grounded sides.

  25. MCQExam level

    In separating ut=κuxx on 0<x<L with u=0 at both ends, why must the separation constant be negative (X′′/X=−k2)?
    1. Because temperatures cannot be negative
    2. A zero or positive constant gives only X≡0 with these end conditions
    3. Because κ is positive
    4. It need not be: any sign gives non-trivial solutions
    Answer and solution

    Answer: B

    Constant 0: X=A+Bx, and X(0)=X(L)=0 give A=B=0. Positive k2: X=Acosh⁡kx+Bsinh⁡kx, and the two zeros again force A=B=0. Only X′′/X=−k2 gives sin⁡(nπx/L). The sign of κ then makes the time factor decay.

    Why people get this wrong: The boundary conditions, not physics intuition about temperature, decide the sign.

  26. NATExam level

    A string fixed at x=0 and x=L (wave speed c) is released from rest in the shape u(x,0)=3sin⁡(2πx/L) mm. At x=L/4 and t=L/(6c) its displacement is ______ mm.
    Answer and solution

    Answer: 1.5 mm (±0.01)

    The shape is the single mode n=2, which oscillates in place: u=3sin⁡2πxLcos⁡2πctL. At x=L/4: sin⁡π2=1. At t=L/6c: cos⁡π3=12. So u=1.5 mm.

    Why people get this wrong: Released from rest means a cosine in time; a sine would start at zero.

  27. MCQExam level

    The axially symmetric (independent of ϕ) solutions of Laplace's equation in spherical coordinates are sums of
    1. (Aℓrℓ+Bℓr−(ℓ+1))Pℓ(cos⁡θ)
    2. (Aℓrℓ+Bℓr−ℓ)cos⁡ℓθ
    3. jℓ(r)Pℓ(cos⁡θ)
    4. (Aℓeℓr+Bℓe−ℓr)Pℓ(cos⁡θ)
    Answer and solution

    Answer: A

    The angular equation is Legendre's, with regular solutions Pℓ(cos⁡θ). The radial equation ddr(r2R′)=ℓ(ℓ+1)R has R=rℓ or r−(ℓ+1) (try rs: s(s+1)=ℓ(ℓ+1)). Inside a sphere keep rℓ, outside keep r−(ℓ+1).

    Why people get this wrong: The decaying power is r−(ℓ+1) in 3D, not r−ℓ as in 2D polar.

  28. MSQExam level

    For utt=c2uxx on the whole line with data u(x,0) and ut(x,0), which statements are true?
    1. u(x0,t0) depends on the initial data at every point of the line
    2. The characteristics are the lines x±ct= constant
    3. u(x0,t0) depends only on the initial data in [x0−ct0,x0+ct0]
    4. Disturbances travel at exactly speed c
    Answer and solution

    Answer: B, C, D

    d'Alembert: u=12[f(x−ct)+f(x+ct)]+12c∫x−ctx+ctg(s)ds. Only data between the two characteristics through (x0,t0) enters: the domain of dependence. So signals move at speed c, never faster, unlike the heat equation, where a change anywhere is felt everywhere at once.

    Why people get this wrong: Infinite propagation speed belongs to the heat equation.

  29. MCQExam level

    On a very long string, a triangular bump of height 2 cm is held at rest and released. Long after release, the largest displacement on the string is
    1. 2 cm
    2. 1 cm
    3. 2 cm
    4. 0.5 cm
    Answer and solution

    Answer: B

    With zero initial velocity, u=12[f(x−ct)+f(x+ct)]: the bump splits into two half-height copies travelling apart. Once they no longer overlap, each has height 1 cm.

    Why people get this wrong: Energy is shared between two pulses, but the shapes add linearly, so each is exactly half as high.

  30. MCQExam level

    A square plate of side L obeys ut=κ(uxx+uyy) with all edges at zero. The slowest decay rate of any mode is
    1. π2κ/L2
    2. 2π2κ/L2
    3. 4π2κ/L2
    4. 2π2κ/L2
    Answer and solution

    Answer: B

    Modes are sin⁡mπxLsin⁡nπyL with rate κπ2(m2+n2)/L2. The smallest has m=n=1: 2π2κ/L2, twice the rate for a rod of the same length (heat escapes in both directions).

    Why people get this wrong: m=0 or n=0 gives a mode that is identically zero here, not a slower one.

← All topics · Timed mock test →