Flow Lab

Syllabus / Module 9: Classical PDEs: Heat, Waves & Potentials

Lesson 9.8

Higher Dimensions: Membranes & Drums

Separation in 2D rectangles and disks; vibrating membranes; Bessel functions; nodal lines and Chladni patterns.

Written byClaude Sonnet

🌊
Why this matters for simulation: Acoustic modes of a combustion chamber or wind tunnel follow the same mathematics.

By the end of this lesson you can…

  • Reduce the 2D wave and heat equations to $-\nabla^2\varphi = \lambda\varphi$ with $\varphi = 0$ on the edge.
  • Find a rectangle's product modes and explain degeneracy and mixed-mode nodal lines.
  • Derive the drum modes $J_m(j_{mn}r/R)\cos m\theta$, and say why a drum has no pitch.
  • Predict how 2D heat decays on a rectangle.
  • Compute Laplacian eigenvalues with the 5-point stencil and check them on a square and a disk.
On this page
·
  1. From a line to a plane
  2. The rectangular membrane
  3. Degeneracy and the lines that move
  4. Heat on a rectangle
  5. The circular drum
  6. Why a drum has no pitch
  7. What if there is no formula?
  8. Can you hear the shape of a drum?
  9. Check your understanding
  10. Problems
  11. Computation
  12. Key Takeaways
  13. What comes next

9.7 Nonhomogeneous Problems & Eigenfunction Expansions ended with a promise: the machinery of project, solve one ODE per mode, and sum goes through unchanged on a drum, but the modes themselves are no longer sines. This lesson pays that debt. A flight simulator or a wind tunnel lives in two and three space dimensions, so the rods and strings of the last lessons were a rehearsal. The plan is the same: find the shapes that keep their form, find how fast each oscillates or decays, and add them up.

▶ Video A membrane's modes on a rectangle and on a disk, nodal lines, the drum's uneven frequency ladder, and the 5-point Laplacian as a matrix whose eigenvectors are the modescaptions only
Transcript
  1. A drumhead is the string problem in two dimensions. Tension pulls from four sides, so the wave equation becomes u t t equals c squared times the Laplacian, and heat becomes u t equals alpha times the Laplacian. Separate time from space and both reduce to one question: minus the Laplacian of phi equals lambda phi, with phi zero on the edge. The frequency is c root lambda, the heat decay rate alpha lambda. On a grid this is a symmetric matrix: the modes are its eigenvectors.
  2. On a rectangle each factor is a string, so the modes are products of sines, with lambda m n equal to pi squared times m squared over a squared plus n squared over b squared. On the square, the frequency ratios to the lowest are 1, 1.5811, 2, 2.2361 and 2.5495. Not whole numbers, except that the two-two mode is the one-one mode with all lengths halved: an exact octave.
  3. On the square, the one-two and two-one modes share a frequency, a degenerate pair, because swapping x and y turns one into the other. So any mixture is also a mode. The identity, with sine pi x sine pi y times cosine pi y plus or minus cosine pi x, shows that the sum vanishes along x plus y equals 1, and the difference along x equals y. Numerically the sum is below ten to the minus fifteen there: rounding noise.
  4. Degeneracy can also be arithmetic. Count pairs with m squared plus n squared equal to s: for s equals 5 there are 2, for 25 still 2, but 50 gives 3, and 65 gives 4. Symmetry is fragile, though. Stretch the square to 1 by 1.1 and the eigenvalues separate, 42.496 against 47.635, a 12 percent gap, so the nodal line drifts.
  5. Now heat. A plate starts at 1 with cold edges. The lowest eigenvalue on the unit square is 19.739, a time constant of 0.0507. At the centre, all the modes give 0.99837 at time 0.01, while the lowest mode alone says 1.33074, above the starting value. By 0.05 one mode is within 1.3 percent, by 0.1 within 0.02 percent. High modes die first, and the plate relaxes to the lowest eigenvector.
  6. Shape matters. Lambda one one is pi squared times 1 plus 1 over b squared. At b equal 0.5 it is 49.348, and the centre is only 0.13660 at time 0.05. At b equal 2 it is 12.337, and the centre is still 0.76989. The narrowest dimension sets the rate. At equal area, the square has 19.739, the thin rectangle 41.946, and the disk 18.168, the slowest. In a duct, the same eigenvalue sets the viscous spin-down.
  7. A circular drum separates in polar coordinates: cosine m theta around, and Bessel's equation along the radius. The clamped rim needs J m to vanish, so the frequencies are Bessel zeros: 2.404826, 3.831706, 5.135622. Divided by the first they give 1, 1.5933, 2.1355, 2.2954, 2.6531. No integer pattern: a drum gives a thump, not a pitch. Mode m n has m nodal diameters and n minus 1 nodal circles.
  8. With no formula, replace the Laplacian by the five-point stencil. On the square the error in the lowest eigenvalue is second order: 19.4868, 19.6759 and 19.7234 at N of 8, 16 and 32, against 19.7392, and each halving cuts the error by 4. On the disk, the staircase boundary is slower: 109, 437 and 1789 unknowns give errors of 9.9, 3.3 and 2.1 percent. The grid also splits the exact pair 26.3746 into 25.758 and 25.782.
  9. Can you hear the shape of a drum? Weyl's law counts the eigenvalues below lambda as about area times lambda over 4 pi, less a perimeter correction. For the unit square the exact counts are 71, 300 and 764 below 1000, 4000 and 10000. Area alone gives 79.6, 318.3 and 795.8; with the perimeter, 69.5, 298.2 and 763.9. So the spectrum reveals area and perimeter, though in 1992 two different polygons were found with identical eigenvalues. Cavities and tanks use this same eigenproblem.
  10. To recap. Two dimensions change the geometry, not the plan: minus the Laplacian of phi equals lambda phi. Rectangles give product modes and moving nodal lines, disks give Bessel modes with no pitch, and heat relaxes to the lowest mode. Where there is no formula, the five-point matrix computes the eigenvalues. Next: infinite domains and the Fourier transform, where the eigenvalues form a continuum and the sum becomes an integral.

From a line to a plane

A membrane is a drumhead: a thin sheet under tension TT (force per unit edge length) with mass σ\sigma per unit area. On a tiny square of side Δ\Delta, the left and right edges pull like a string, giving the vertical force T uxx Δ2T\,u_{xx}\,\Delta^2, and the top and bottom edges add T uyy Δ2T\,u_{yy}\,\Delta^2. Newton’s law with mass σΔ2\sigma\Delta^2 gives

utt=c2(uxx+uyy)=c2∇2u,c=T/σ,u_{tt} = c^2\left(u_{xx} + u_{yy}\right) = c^2\nabla^2u,\qquad c = \sqrt{T/\sigma},

the 2D wave equation. The symbol ∇2u=uxx+uyy\nabla^2u = u_{xx}+u_{yy} is the Laplacian: the operator of 9.6 Laplace's Equation: Rectangles & Disks. The heat equation of 9.2 The Heat (Diffusion) Equation becomes ut=α∇2uu_t = \alpha\nabla^2u. In both, u=0u = 0 on the edge (a clamped rim, a cold wall).

Separate time from space as in 9.4 The Wave Equation & Vibrating Strings: u=φ(x,y) τ(t)u = \varphi(x,y)\,\tau(t). Then τ′′/(c2τ)=∇2φ/φ\tau''/(c^2\tau) = \nabla^2\varphi/\varphi must be a constant, −λ-\lambda, and

−∇2φ=λφ  in the region,φ=0  on the edge,τ′′=−c2λ τ (wave),τ′=−αλ τ (heat).-\nabla^2\varphi = \lambda\varphi\ \text{ in the region},\qquad \varphi = 0\ \text{ on the edge},\qquad \tau'' = -c^2\lambda\,\tau\ \text{(wave)},\quad \tau' = -\alpha\lambda\,\tau\ \text{(heat)}.

Each eigenvalue λ\lambda gives a wave of angular frequency ω=cλ\omega = c\sqrt\lambda and a heat mode that decays like e−αλte^{-\alpha\lambda t}. All the geometry sits in one question: the eigenfunctions and eigenvalues of −∇2-\nabla^2 on this shape. This is the Sturm-Liouville problem of 8.2 Eigenvalue Problems & Sturm–Liouville Theory with two space variables: the eigenvalues are positive, eigenfunctions with different λ\lambda are perpendicular (∬φjφk dA=0\iint\varphi_j\varphi_k\,dA = 0), and every initial shape is a sum of them.

The rectangular membrane

Take the rectangle 0<x<a0 \lt x \lt a, 0<y<b0 \lt y \lt b and separate again, φ=X(x)Y(y)\varphi = X(x)Y(y). Then X′′/X+Y′′/Y=−λX''/X + Y''/Y = -\lambda, so each term is its own constant and each factor is a string problem with fixed ends:

φmn=sin⁡mπxa sin⁡nπyb,λmn=π2(m2a2+n2b2),ωmn=πcm2a2+n2b2,\varphi_{mn} = \sin\frac{m\pi x}{a}\,\sin\frac{n\pi y}{b},\qquad \lambda_{mn} = \pi^2\left(\frac{m^2}{a^2}+\frac{n^2}{b^2}\right),\qquad \omega_{mn} = \pi c\sqrt{\frac{m^2}{a^2}+\frac{n^2}{b^2}},

with m,n=1,2,3,…m, n = 1, 2, 3, \dots The general solution is u=∑m,n(Amncos⁡ωmnt+Bmnsin⁡ωmnt)φmnu = \sum_{m,n}(A_{mn}\cos\omega_{mn}t + B_{mn}\sin\omega_{mn}t)\varphi_{mn}, with Amn=4ab∬f φmn dAA_{mn} = \frac{4}{ab}\iint f\,\varphi_{mn}\,dA from a double sine series (the series of 8.3 Fourier Series once in xx, once in yy). Mode (m,n)(m,n) has m−1m-1 vertical and n−1n-1 horizontal nodal lines, curves where the sheet does not move; sand on a vibrating plate collects there. (Problem 1 puts numbers on a square drum.)

The ratios to the lowest, (m2+n2)/2\sqrt{(m^2+n^2)/2}, are 1, 1.5811, 2, 2.2361, 2.54951,\ 1.5811,\ 2,\ 2.2361,\ 2.5495 for (1,1),(1,2),(2,2),(1,3),(2,3)(1,1),(1,2),(2,2),(1,3),(2,3) (from the sim): not whole numbers, except that (2,2)(2,2) is (1,1)(1,1) with all lengths halved, an exact octave.

Degeneracy and the lines that move

On a square, ω12=ω21\omega_{12} = \omega_{21}: two different shapes with one frequency, a degenerate pair. The reason is symmetry, since swapping xx and yy turns one mode into the other. The consequence is new in 2D. Any combination cos⁡γ φ12+sin⁡γ φ21\cos\gamma\,\varphi_{12} + \sin\gamma\,\varphi_{21} oscillates at the shared frequency, so it is also a mode, with its own nodal curve. Using sin⁡2θ=2sin⁡θcos⁡θ\sin 2\theta = 2\sin\theta\cos\theta,

sin⁡πxsin⁡2πy±sin⁡2πxsin⁡πy=2sin⁡πx sin⁡πy (cos⁡πy±cos⁡πx).\sin\pi x\sin2\pi y \pm \sin2\pi x\sin\pi y = 2\sin\pi x\,\sin\pi y\,\left(\cos\pi y \pm \cos\pi x\right).

The factor sin⁡πxsin⁡πy\sin\pi x\sin\pi y vanishes on the edge. The bracket vanishes where cos⁡πy=∓cos⁡πx\cos\pi y = \mp\cos\pi x, which is the line x+y=1x + y = 1 for the plus sign and x=yx = y for the minus. The sim checks both: the sum of the two modes along x+y=1x + y = 1 is below 10−1510^{-15} at every sample, which is rounding noise. Intermediate γ\gamma bends the nodal line, giving Chladni patterns (from memory: real Chladni figures are made on plates, which obey a fourth-order equation, so a membrane gives the idea, not the exact patterns).

Degeneracy can also be arithmetic. Counting ordered pairs with m2+n2=sm^2 + n^2 = s (the sim’s counter): s=25s = 25 gives only (3,4),(4,3)(3,4),(4,3), since (0,5)(0,5) is not allowed, while s=50s = 50 gives three, (1,7),(5,5),(7,1)(1,7),(5,5),(7,1), and s=65s = 65 gives four. The coincidence of (5,5)(5,5) with the others is number theory, not symmetry.

Symmetry is fragile. Stretch the square to 1×1.11\times1.1 and the sim gives λ12=π2(1+4/1.21)=42.496\lambda_{12} = \pi^2(1 + 4/1.21) = 42.496 against λ21=π2(4+1/1.21)=47.635\lambda_{21} = \pi^2(4 + 1/1.21) = 47.635: a 12% gap. The pair separates, and the combination φ12+φ21\varphi_{12}+\varphi_{21} is no longer a single mode: it beats, and its nodal line drifts.

◆ Interactive Membranes, drums, finite-difference eigenvalues and 2D heat

Loading simulation…

What to try. In the rectangle view press (1,2)+(2,1)(1,2)+(2,1), then set the aspect to 1.11.1 and animate. Switch to the disk, set m=0m = 0, then m=2m = 2, and compare the frequency ladder with the string’s dashed harmonics.

Heat on a rectangle

Heat uses the same eigenproblem with decay in place of oscillation. A plate 0<x<10\lt x\lt1, 0<y<b0\lt y\lt b starts at u=1u = 1 with its edges held at 00 (α=1\alpha = 1). The coefficient of φmn\varphi_{mn} is 4b∬φmn dA\frac{4}{b}\iint\varphi_{mn}\,dA, which for odd m,nm, n gives 16/(mnπ2)16/(mn\pi^2) (zero if either is even, as in the 1D rod). So

u(x,y,t)=∑m,n odd16mnπ2 e−π2(m2+n2/b2)t sin⁡mπx sin⁡nπyb.u(x,y,t) = \sum_{m,n\ \text{odd}}\frac{16}{mn\pi^2}\,e^{-\pi^2(m^2+n^2/b^2)t}\,\sin m\pi x\,\sin\frac{n\pi y}{b}.

For the unit square λ11=2π2=19.739\lambda_{11} = 2\pi^2 = 19.739, so the time constant of the slowest mode is 1/λ11=0.05071/\lambda_{11} = 0.0507. The centre temperature from the sim’s 59-term sum is

tt0.010.020.050.10
all modes0.998370.950940.596470.22514
lowest mode only1.330741.092360.604210.22519

Early on one mode is badly wrong (even above the starting value 11): the sharp edges of the initial data need many modes. By t=0.05t = 0.05 one mode is within 1.3%1.3\% and by t=0.1t = 0.1 within 0.02%0.02\%: high modes die first and the plate relaxes to φ11\varphi_{11}. Repeated smoothing finds the lowest eigenvector, an idea the numerical method below reuses.

Shape effects. Stretching a side changes λ11=π2(1+1/b2)\lambda_{11} = \pi^2(1 + 1/b^2). At b=0.5b = 0.5 it is 49.34849.348 (time constant 0.02030.0203) and the centre temperature at t=0.05t = 0.05 is 0.136600.13660; at b=2b = 2 it is 12.33712.337 (time constant 0.08110.0811) and the centre is still 0.769890.76989. A thin plate cools through its short side: the narrowest dimension sets the rate. At equal area (1), the square has λ11=19.739\lambda_{11} = 19.739, the 2×0.52\times0.5 rectangle 41.94641.946, and the disk (radius 1/π1/\sqrt\pi) j012/R2=18.168j_{01}^2/R^2 = 18.168, with j01=2.404826j_{01}=2.404826 from the next section. The disk cools slowest, and I recall that this is a theorem, Faber-Krahn: among shapes of equal area the disk has the smallest first eigenvalue (from memory).

The circular drum

For a disk of radius RR use polar coordinates (r,θ)(r,\theta) and the Laplacian ∇2φ=φrr+φr/r+φθθ/r2\nabla^2\varphi = \varphi_{rr} + \varphi_r/r + \varphi_{\theta\theta}/r^2. Separate, φ=ρ(r)Θ(θ)\varphi = \rho(r)\Theta(\theta). The angular equation Θ′′=−m2Θ\Theta'' = -m^2\Theta must be periodic, since θ\theta and θ+2π\theta+2\pi are the same point, so Θ=cos⁡mθ\Theta = \cos m\theta or sin⁡mθ\sin m\theta with m=0,1,2,…m = 0, 1, 2,\dots The radial equation, with λ=k2\lambda = k^2, is

ρ′′+1rρ′+(k2−m2r2)ρ=0,\rho'' + \frac1r\rho' + \left(k^2 - \frac{m^2}{r^2}\right)\rho = 0,

Bessel’s equation of order mm (5.8 Series Solutions). One solution, Jm(kr)J_m(kr), is smooth at the centre; the second blows up there and is discarded because a drumhead does not. The clamped rim requires Jm(kR)=0J_m(kR) = 0, so kRkR is a zero of JmJ_m. Call the nn-th positive zero jmnj_{mn}. Then

φmn=Jm ⁣(jmnrR){cos⁡mθsin⁡mθ},ωmn=c jmnR.\varphi_{mn} = J_m\!\left(\frac{j_{mn}r}{R}\right)\begin{Bmatrix}\cos m\theta\\ \sin m\theta\end{Bmatrix},\qquad \omega_{mn} = \frac{c\,j_{mn}}{R}.

For m≥1m \ge 1 the cosine and sine modes share a frequency: a degenerate pair caused by rotational symmetry. Mode (m,n)(m,n) has mm nodal diameters and n−1n-1 interior nodal circles, so (0,1)(0,1) moves as one piece and (1,1)(1,1) rocks about a diameter.

The sim evaluates Jm(x)=1π∫0πcos⁡(mt−xsin⁡t) dtJ_m(x) = \frac1\pi\int_0^\pi\cos(mt - x\sin t)\,dt (from memory) by the trapezoid rule, scans for sign changes and bisects:

(m,n)(m,n)(0,1)(1,1)(2,1)(0,2)(3,1)(1,2)(4,1)(2,2)(0,3)(0,4)
jmnj_{mn}2.4048263.8317065.1356225.5200786.3801627.0155877.5883428.4172448.65372811.791534
ratio to j01j_{01}11.59332.13552.29542.65312.91733.15553.50013.59854.9033

Why a drum has no pitch

A string’s frequencies are whole multiples of the fundamental, so its sound repeats with one period and the ear hears a pitch. The drum’s lowest ratios are 1, 1.59, 2.14, 2.30, 2.65, 2.92, 3.16, 3.50, 3.601,\ 1.59,\ 2.14,\ 2.30,\ 2.65,\ 2.92,\ 3.16,\ 3.50,\ 3.60, with no integer pattern: Bessel zeros are roughly π\pi apart only far out (from memory). The sound is a cluster of unrelated frequencies, a thump with a vague pitch. I recall that a timpani’s air loading moves (1,1),(2,1),(3,1)(1,1),(2,1),(3,1) toward 3:23{:}2, 2:12{:}1, 5:25{:}2 (the sim has 1.5933, 2.1355, 2.65311.5933,\ 2.1355,\ 2.6531), which is why it can be tuned (from memory). Expanding an initial shape uses a Fourier series in θ\theta and a Fourier-Bessel series in rr with weight rr, as in 8.6 Orthogonal Polynomials & Special Functions.

What if there is no formula?

Rectangles and disks are the exceptions; for a wing section the eigenvalues must be computed. Put a grid of spacing hh on the region and replace ∇2\nabla^2 by the 5-point stencil

−∇2u≈4ui,j−ui+1,j−ui−1,j−ui,j+1−ui,j−1h2,-\nabla^2u \approx \frac{4u_{i,j} - u_{i+1,j} - u_{i-1,j} - u_{i,j+1} - u_{i,j-1}}{h^2},

with u=0u = 0 at nodes outside the region. This symmetric matrix AA has 4/h24/h^2 on the diagonal and −1/h2-1/h^2 for each neighbour; its eigenvectors, laid back on the grid, are the discrete modes.

The square. Here the eigenvalues are known, (4/h2)[sin⁡2(mπh/2)+sin⁡2(nπh/2)](4/h^2)\left[\sin^2(m\pi h/2) + \sin^2(n\pi h/2)\right] (the sim’s values match it to every printed digit). Expanding the sine gives the relative error −π2h2/12=−0.8225 h2-\pi^2h^2/12 = -0.8225\,h^2 for the lowest one: second order. At h=0.1h = 0.1 the formula gives 19.577419.5774 against 19.739219.7392, an error of −0.820%-0.820\%. The sim’s iteration, which never uses the formula, gives

N=1/hN = 1/h81632exact
λ1h\lambda_1^h19.486819.675919.723419.7392
error−0.2524-0.2524−0.0633-0.0633−0.0158-0.0158

Each halving of hh cuts the error by 4.04.0. The pair λ2=λ3\lambda_2 = \lambda_3 stays equal (47.2338, 48.8116, 49.2134 against 49.348049.3480) because the grid has the square’s symmetry.

How it finds them. The sim solves Az=xA\mathbf z = \mathbf x and repeats with x←z\mathbf x\leftarrow\mathbf z (inverse iteration): each solve divides the component along an eigenvector by its λ\lambda, so the smallest λ\lambda wins, much as the heat decay above leaves the lowest mode. It carries ten vectors and re-diagonalizes in that small subspace each round, since one vector alone stalls on a close pair (10.1 Finite Differences: Discretizing Derivatives has the stencil).

The disk, on a square grid. Keep the nodes strictly inside the circle. The boundary is a staircase. The lowest eigenvalue against the exact j012=5.7832j_{01}^2 = 5.7832:

N=2/hN = 2/h122448
unknowns1094371789
λ1h\lambda_1^h5.20855.59515.6634
error−9.9%-9.9\%−3.3%-3.3\%−2.1%-2.1\%

This is far slower than h2h^2: the staircase puts the wall in the wrong place by up to a grid step, so I expect first order at best. Smooth boundaries need cut cells or fitted grids (10.11 2D Grids, Boundary Conditions & Staggered (MAC) Grids). The grid also shows its axes. The m=1m=1 pair stays degenerate (14.363914.3639 twice at N=48N=48, against 14.682014.6820), but the m=2m=2 pair, exactly 26.374626.3746 twice, splits into 25.75825.758 and 25.78225.782 at N=48N = 48 (25.24125.241 and 25.33425.334 at N=24N = 24). A 90∘90^\circ turn only flips the sign of cos⁡2θ\cos2\theta, while a 45∘45^\circ turn gives sin⁡2θ\sin2\theta, and the grid lacks 45∘45^\circ symmetry.

Can you hear the shape of a drum?

Each shape has its own list of λk\lambda_k, so does the list identify the shape? Mark Kac posed this as “can one hear the shape of a drum?” (1966, from memory). Weyl’s law (from memory) says the number of eigenvalues below Λ\Lambda is about AΛ/4πA\Lambda/4\pi for area AA, and with the perimeter PP the next term is −PΛ/4π-P\sqrt\Lambda/4\pi (from memory). For the unit square (A=1A = 1, P=4P = 4) the sim counts lattice points:

Λ\Lambda1000400010000
exact count71300764
area term only79.6318.3795.8
area and perimeter69.5298.2763.9

So a long list of eigenvalues reveals area and perimeter. But I recall that the full answer is no: in 1992 Gordon, Webb and Wolpert found two different polygons with identical eigenvalues (from memory).

Check your understanding

✅ Concept check

On a square membrane, which other mode has the same frequency as the $(2,3)$ mode?

✅ Concept check

Why does a drumhead have no clear pitch while a string does?

✅ Concept check

The frequency of the $(1,1)$ mode of a circular drum, divided by that of the $(0,1)$ mode, equals $j_{11}/j_{01}$. What is it to four decimals?

✅ Concept check

The 5-point Laplacian on the unit square is second order. You halve $h$. The error in the lowest eigenvalue goes down by a factor of about

Problems

✏️ Pen & paperFrequencies of a square drum●○○

A square membrane has side L=0.3L = 0.3 m and c=100c = 100 m/s. Find the three lowest distinct frequencies in hertz and say which is degenerate.

Show worked solution

fmn=c2Lm2+n2=166.67m2+n2f_{mn} = \dfrac{c}{2L}\sqrt{m^2+n^2} = 166.67\sqrt{m^2+n^2} Hz. The values m2+n2=2,5,8m^2+n^2 = 2, 5, 8 belong to (1,1)(1,1), to (1,2)(1,2) and (2,1)(2,1), and to (2,2)(2,2), giving 235.70235.70, 372.68372.68 and 471.40471.40 Hz. The middle one is degenerate.

✏️ Pen & paperThe diagonal nodal line●●○

Show that sin⁡πxsin⁡2πy+sin⁡2πxsin⁡πy\sin\pi x\sin2\pi y + \sin2\pi x\sin\pi y vanishes on x+y=1x + y = 1, and find the nodal curve of the difference. Why are both modes at the same frequency?

Show worked solution

The sum is 2sin⁡πxsin⁡πy (cos⁡πy+cos⁡πx)2\sin\pi x\sin\pi y\,(\cos\pi y + \cos\pi x). On x+y=1x + y = 1, cos⁡πy=−cos⁡πx\cos\pi y = -\cos\pi x, so the bracket is 00. For the difference the bracket is cos⁡πy−cos⁡πx\cos\pi y - \cos\pi x, zero on y=xy = x. Both combine φ12\varphi_{12} and φ21\varphi_{21}, which share ω=πc5/L\omega = \pi c\sqrt5/L, so one time factor cos⁡ωt\cos\omega t serves both.

✏️ Pen & paperA round drum●●○

A drum has radius R=0.25R = 0.25 m and c=100c = 100 m/s. Use j01=2.404826j_{01} = 2.404826, j11=3.831706j_{11} = 3.831706, j21=5.135622j_{21} = 5.135622, j02=5.520078j_{02} = 5.520078 to find the frequencies of modes (0,1),(1,1),(2,1),(0,2)(0,1), (1,1), (2,1), (0,2), and say which have nodal circles and diameters.

Show worked solution

fmn=c jmn/(2πR)=63.662 jmnf_{mn} = c\,j_{mn}/(2\pi R) = 63.662\,j_{mn} Hz: f01=153.10f_{01} = 153.10, f11=243.93f_{11} = 243.93, f21=326.94f_{21} = 326.94, f02=351.42f_{02} = 351.42. Only (0,2)(0,2) has an interior nodal circle (n−1=1n-1 = 1); (1,1)(1,1) and (2,1)(2,1) have one and two nodal diameters; (0,1)(0,1) has no interior nodal line.

✏️ Pen & paperError of the 5-point eigenvalue●●●

(a) Show 4h2sin⁡2kh2=k2(1−k2h212+⋯ )\dfrac{4}{h^2}\sin^2\dfrac{kh}{2} = k^2\left(1 - \dfrac{k^2h^2}{12} + \cdots\right). (b) Estimate the relative error of the lowest 5-point eigenvalue on the unit square. (c) For h=0.1h = 0.1 compare with the exact 8h2sin⁡2πh2\dfrac{8}{h^2}\sin^2\dfrac{\pi h}{2}.

Show worked solution

(a) sin⁡2u=u2−u4/3+⋯\sin^2u = u^2 - u^4/3 + \cdots with u=kh/2u = kh/2 gives 4h2(k2h24−k4h448)=k2−k4h212\dfrac{4}{h^2}\left(\dfrac{k^2h^2}{4} - \dfrac{k^4h^4}{48}\right) = k^2 - \dfrac{k^4h^2}{12}. (b) Two directions with k=πk = \pi: λ11h=2π2(1−π2h2/12)\lambda^h_{11} = 2\pi^2\left(1 - \pi^2h^2/12\right), relative error −0.8225 h2-0.8225\,h^2. (c) At h=0.1h = 0.1 the estimate is −0.8225%-0.8225\%. Directly, sin⁡(0.05π)=0.156434\sin(0.05\pi) = 0.156434 and λh=800×0.024472=19.5774\lambda^h = 800\times0.024472 = 19.5774, so the error is −0.820%-0.820\%; the next series term explains the difference.

Computation

First, the exercise: build the 5-point Laplacian and compute its eigenvalues on the unit square, and compare them with π2(m2+n2)\pi^2(m^2+n^2).

🧪 Coding exerciseExercise: the 5-point Laplacian and its eigenvalues
⌘/Ctrl + Enter

Second, a disk on a square grid. The staircase boundary makes the eigenvalues converge slowly, and splits the m=2m=2 pair.

🐍 PythonLab: the unit disk on a masked grid
⌘/Ctrl + Enter

Try N=24N = 24 and watch the fourth and fifth values separate.

Key Takeaways

  • The 2D wave and heat equations reduce to −∇2φ=λφ-\nabla^2\varphi = \lambda\varphi with φ=0\varphi = 0 on the edge; frequency is cλc\sqrt\lambda and decay rate αλ\alpha\lambda. On a grid this is a symmetric matrix eigenproblem.
  • On a rectangle the modes are products of sines with ωmn=πcm2/a2+n2/b2\omega_{mn} = \pi c\sqrt{m^2/a^2+n^2/b^2}. On the square (1,2)(1,2) and (2,1)(2,1) share a frequency, so sin⁡πxsin⁡2πy±sin⁡2πxsin⁡πy\sin\pi x\sin2\pi y\pm\sin2\pi x\sin\pi y are modes with nodal lines x+y=1x+y=1 and x=yx=y; stretching to 1×1.11\times1.1 splits the pair (42.49642.496 against 47.63547.635).
  • On a disk, φ=Jm(jmnr/R)cos⁡mθ\varphi = J_m(j_{mn}r/R)\cos m\theta and ω=cjmn/R\omega = cj_{mn}/R, with mm nodal diameters and n−1n-1 nodal circles. The ratios 1,1.5933,2.1355,2.2954,2.65311, 1.5933, 2.1355, 2.2954, 2.6531 are not whole numbers, so a drum has no clear pitch.
  • Heat on a 1×b1\times b plate decays at π2(1+1/b2)\pi^2(1+1/b^2): 19.73919.739 for the unit square, 49.34849.348 at b=0.5b=0.5, 12.33712.337 at b=2b=2. The lowest mode takes over after a few time constants.
  • The 5-point stencil is second order on the unit square (relative error −π2h2/12-\pi^2h^2/12; 19.486819.4868, 19.675919.6759, 19.723419.7234 against 19.739219.7392). On a staircase disk the lowest value is off by 9.9%9.9\%, 3.3%3.3\%, 2.1%2.1\% at N=12,24,48N = 12, 24, 48 and the m=2m=2 pair splits.
  • Weyl counting (from memory) gives area and perimeter from the spectrum: 300300 eigenvalues of the unit square below 40004000 against 298.2298.2 predicted.

What comes next

Every region here was bounded, so the spectrum was a discrete list and the solution a sum. A wind-tunnel wake, a pulse on a very long wire or heat from a point in unbounded space has no edge to fix the modes: the eigenvalues form a continuum and the sum becomes an integral. The next lesson, 9.9 Infinite Domains & the Fourier Transform, develops this Fourier transform, with the Gaussian heat kernel and group velocity as payoffs that no bounded region can give.

Finished this lesson?

Mark it complete to track your progress (saved in this browser). Reading every section does this for you.

Where this leads

These later lessons build directly on this one: