Browse papers
A

Section A: Long Answer Questions

Attempt all / any as specified.

4 questions
1long12 marks

(a) State and prove the first shifting theorem for Laplace transforms. Hence evaluate L{e2t(3cos4t2sin4t)}\mathcal{L}\{e^{-2t}(3\cos 4t - 2\sin 4t)\}. [6]

(b) Using the Laplace transform method, solve the initial value problem

d2ydt2+4dydt+4y=e2t,y(0)=0,  y(0)=1.\frac{d^2y}{dt^2} + 4\frac{dy}{dt} + 4y = e^{-2t}, \qquad y(0)=0,\; y'(0)=1.

[6]

(a) First Shifting Theorem [6]

Statement. If L{f(t)}=F(s)\mathcal{L}\{f(t)\} = F(s) for s>as > a, then

L{eatf(t)}=F(sa).\mathcal{L}\{e^{at}f(t)\} = F(s-a).

Proof. By definition,

L{eatf(t)}=0esteatf(t)dt=0e(sa)tf(t)dt=F(sa),\mathcal{L}\{e^{at}f(t)\} = \int_0^\infty e^{-st}\,e^{at}f(t)\,dt = \int_0^\infty e^{-(s-a)t}f(t)\,dt = F(s-a),

which converges for sa>s-a > (abscissa of convergence), i.e. s>as > a. Replacing ss by sas-a in F(s)F(s) shifts the transform along the ss-axis. \blacksquare

Evaluation. Take f(t)=3cos4t2sin4tf(t) = 3\cos 4t - 2\sin 4t. Then

L{3cos4t}=3ss2+16,L{2sin4t}=8s2+16,\mathcal{L}\{3\cos 4t\} = \frac{3s}{s^2+16}, \qquad \mathcal{L}\{2\sin 4t\} = \frac{8}{s^2+16},

so F(s)=3s8s2+16F(s) = \dfrac{3s - 8}{s^2+16}. Here a=2a = -2, hence by the shifting theorem replace ss+2s \to s+2:

L{e2t(3cos4t2sin4t)}=3(s+2)8(s+2)2+16=3s2(s+2)2+16.\mathcal{L}\{e^{-2t}(3\cos 4t - 2\sin 4t)\} = \frac{3(s+2) - 8}{(s+2)^2 + 16} = \boxed{\dfrac{3s - 2}{(s+2)^2 + 16}}.

(b) IVP by Laplace Transform [6]

Given y+4y+4y=e2ty'' + 4y' + 4y = e^{-2t}, y(0)=0y(0)=0, y(0)=1y'(0)=1. Let Y(s)=L{y}Y(s)=\mathcal{L}\{y\}.

Taking transforms with L{y}=s2Ysy(0)y(0)\mathcal{L}\{y''\} = s^2Y - sy(0) - y'(0) and L{y}=sYy(0)\mathcal{L}\{y'\} = sY - y(0):

(s2Y01)+4(sY0)+4Y=1s+2.\big(s^2Y - 0 - 1\big) + 4\big(sY - 0\big) + 4Y = \frac{1}{s+2}. (s2+4s+4)Y1=1s+2    (s+2)2Y=1+1s+2=s+3s+2.\big(s^2 + 4s + 4\big)Y - 1 = \frac{1}{s+2} \;\Rightarrow\; (s+2)^2 Y = 1 + \frac{1}{s+2} = \frac{s+3}{s+2}. Y=s+3(s+2)3=(s+2)+1(s+2)3=1(s+2)2+1(s+2)3.Y = \frac{s+3}{(s+2)^3} = \frac{(s+2)+1}{(s+2)^3} = \frac{1}{(s+2)^2} + \frac{1}{(s+2)^3}.

Inverting term by term (using L1{1/(s+2)n}=tn1(n1)!e2t\mathcal{L}^{-1}\{1/(s+2)^n\} = \dfrac{t^{n-1}}{(n-1)!}e^{-2t}):

y(t)=te2t+12t2e2t=e2t(t+12t2).\boxed{y(t) = t\,e^{-2t} + \tfrac{1}{2}t^2 e^{-2t} = e^{-2t}\left(t + \tfrac{1}{2}t^2\right).}

Check: y(0)=0y(0)=0, y(0)=1y'(0)=1 are satisfied.

laplace-transformsinverse-laplaceode-solution
2long12 marks

(a) Obtain the Fourier series expansion of the function

f(x)={πx,π<x<0πx,0<x<πf(x) = \begin{cases} -\pi - x, & -\pi < x < 0 \\ \pi - x, & 0 < x < \pi \end{cases}

and deduce that π4=113+1517+\dfrac{\pi}{4} = 1 - \dfrac{1}{3} + \dfrac{1}{5} - \dfrac{1}{7} + \cdots [8]

(b) Find the half-range cosine series for f(x)=xf(x) = x in the interval 0<x<20 < x < 2. [4]

(a) Fourier Series and Leibniz Series [8]

For π<x<π-\pi < x < \pi, f(x)=xf(x) = -x plus an odd step part; we compute directly. Write the Fourier series f(x)=a02+n1(ancosnx+bnsinnx)f(x)=\dfrac{a_0}{2}+\sum_{n\ge1}(a_n\cos nx + b_n\sin nx) on (π,π)(-\pi,\pi).

Constant term.

a0=1ππ0(πx)dx+1π0π(πx)dx.a_0 = \frac1\pi\int_{-\pi}^{0}(-\pi - x)\,dx + \frac1\pi\int_{0}^{\pi}(\pi - x)\,dx.

π0(πx)dx=[πxx2/2]π0=(π2π2/2)=π2/2.\int_{-\pi}^0(-\pi-x)dx = [-\pi x - x^2/2]_{-\pi}^0 = -(\pi^2 - \pi^2/2) = -\pi^2/2. 0π(πx)dx=π2π2/2=π2/2.\int_0^\pi(\pi-x)dx = \pi^2 - \pi^2/2 = \pi^2/2. Hence a0=0a_0 = 0.

Cosine coefficients. The function f(x)+xf(x)+x equals π-\pi on (π,0)(-\pi,0) and +π+\pi on (0,π)(0,\pi), which is odd; and x-x is odd. So f(x)f(x) is an odd function, giving an=0a_n = 0 for all nn.

Sine coefficients.

bn=2π0π(πx)sinnxdx=2π[π ⁣0π ⁣sinnxdx0π ⁣xsinnxdx].b_n = \frac2\pi\int_0^\pi (\pi - x)\sin nx\,dx = \frac2\pi\left[\pi\!\int_0^\pi\!\sin nx\,dx - \int_0^\pi\! x\sin nx\,dx\right].

0πsinnxdx=1cosnπn=1(1)nn.\int_0^\pi\sin nx\,dx = \dfrac{1-\cos n\pi}{n}=\dfrac{1-(-1)^n}{n}. Using 0πxsinnxdx=πcosnπn=π(1)nn\int_0^\pi x\sin nx\,dx = -\dfrac{\pi\cos n\pi}{n}=-\dfrac{\pi(-1)^n}{n}:

bn=2π[π(1(1)n)n+π(1)nn]=2n[1(1)n+(1)n]=2n.b_n = \frac2\pi\left[\frac{\pi(1-(-1)^n)}{n} + \frac{\pi(-1)^n}{n}\right] = \frac{2}{n}\big[1-(-1)^n+(-1)^n\big] = \frac{2}{n}.

Therefore

f(x)=n=12nsinnx=2(sinx+12sin2x+13sin3x+).\boxed{f(x) = \sum_{n=1}^\infty \frac{2}{n}\sin nx = 2\left(\sin x + \tfrac12\sin 2x + \tfrac13\sin 3x + \cdots\right).}

Deduction. Put x=π2x = \dfrac{\pi}{2}. Then f(π/2)=ππ/2=π/2f(\pi/2) = \pi - \pi/2 = \pi/2, and sin(nπ/2)\sin(n\pi/2) is 00 for even nn and (1)(n1)/2(-1)^{(n-1)/2} for odd nn:

π2=2(113+1517+)    π4=113+1517+\frac{\pi}{2} = 2\left(1 - \frac13 + \frac15 - \frac17 + \cdots\right) \;\Rightarrow\; \boxed{\frac{\pi}{4} = 1 - \frac13 + \frac15 - \frac17 + \cdots}

(b) Half-Range Cosine Series of f(x)=xf(x)=x on (0,2)(0,2) [4]

Here L=2L = 2. Cosine series: f(x)=a02+n1ancosnπx2f(x) = \dfrac{a_0}{2} + \sum_{n\ge1} a_n\cos\dfrac{n\pi x}{2}.

a0=2202xdx=11x2202=2.a_0 = \frac{2}{2}\int_0^2 x\,dx = \frac{1}{1}\cdot\frac{x^2}{2}\Big|_0^2 = 2. an=2202xcosnπx2dx=[2xnπsinnπx2+4n2π2cosnπx2]02=4n2π2(cosnπ1)=4((1)n1)n2π2.a_n = \frac{2}{2}\int_0^2 x\cos\frac{n\pi x}{2}\,dx = \left[\frac{2x}{n\pi}\sin\frac{n\pi x}{2} + \frac{4}{n^2\pi^2}\cos\frac{n\pi x}{2}\right]_0^2 = \frac{4}{n^2\pi^2}\big(\cos n\pi - 1\big) = \frac{4((-1)^n - 1)}{n^2\pi^2}.

Thus an=0a_n = 0 for even nn, and an=8n2π2a_n = -\dfrac{8}{n^2\pi^2} for odd nn. Hence

f(x)=18π2nodd1n2cosnπx2=18π2(cosπx2+19cos3πx2+).\boxed{f(x) = 1 - \frac{8}{\pi^2}\sum_{n\,\text{odd}}\frac{1}{n^2}\cos\frac{n\pi x}{2} = 1 - \frac{8}{\pi^2}\left(\cos\frac{\pi x}{2} + \frac{1}{9}\cos\frac{3\pi x}{2} + \cdots\right).}
fourier-serieshalf-range-expansion
3long16 marks

A laterally insulated metal rod of length LL has its ends maintained at zero temperature. The temperature u(x,t)u(x,t) satisfies the one-dimensional heat equation

ut=c22ux2.\frac{\partial u}{\partial t} = c^2 \frac{\partial^2 u}{\partial x^2}.

(a) Using the method of separation of variables, derive the general solution satisfying the boundary conditions u(0,t)=0u(0,t)=0 and u(L,t)=0u(L,t)=0. [10]

(b) Hence find the temperature distribution if the initial temperature is u(x,0)=sinπxL+3sin3πxLu(x,0) = \sin\dfrac{\pi x}{L} + 3\sin\dfrac{3\pi x}{L}. [6]

(a) Separation of Variables for the Heat Equation [10]

Seek u(x,t)=X(x)T(t)u(x,t) = X(x)T(t). Substituting into ut=c2uxxu_t = c^2 u_{xx}:

XT=c2XT    Tc2T=XX=λ  (constant).X T' = c^2 X'' T \;\Rightarrow\; \frac{T'}{c^2 T} = \frac{X''}{X} = -\lambda \;(\text{constant}).

This gives two ODEs:

X+λX=0,T+c2λT=0.X'' + \lambda X = 0, \qquad T' + c^2\lambda T = 0.

Boundary conditions. u(0,t)=0X(0)=0u(0,t)=0 \Rightarrow X(0)=0 and u(L,t)=0X(L)=0u(L,t)=0 \Rightarrow X(L)=0 (non-trivial TT).

Eigenvalue analysis. For λ0\lambda \le 0 the only solution satisfying both BCs is X0X\equiv0. For λ=p2>0\lambda = p^2 > 0:

X(x)=Acospx+Bsinpx.X(x) = A\cos px + B\sin px.

X(0)=0A=0X(0)=0 \Rightarrow A=0. X(L)=BsinpL=0X(L)=B\sin pL = 0 with B0sinpL=0p=nπLB\ne0 \Rightarrow \sin pL = 0 \Rightarrow p = \dfrac{n\pi}{L}, n=1,2,3,n=1,2,3,\dots

Thus eigenvalues λn=(nπL)2\lambda_n = \left(\dfrac{n\pi}{L}\right)^2 and eigenfunctions Xn(x)=sinnπxLX_n(x) = \sin\dfrac{n\pi x}{L}.

Time part. Tn=c2(nπL)2TnTn(t)=ec2n2π2t/L2.T_n' = -c^2\left(\dfrac{n\pi}{L}\right)^2 T_n \Rightarrow T_n(t) = e^{-c^2 n^2\pi^2 t/L^2}.

By superposition, the general solution satisfying the BCs is

u(x,t)=n=1bnsinnπxL  ec2n2π2t/L2.\boxed{u(x,t) = \sum_{n=1}^\infty b_n \sin\frac{n\pi x}{L}\;e^{-c^2 n^2\pi^2 t/L^2}.}

(b) Temperature Distribution [6]

Apply the initial condition u(x,0)=sinπxL+3sin3πxLu(x,0) = \sin\dfrac{\pi x}{L} + 3\sin\dfrac{3\pi x}{L}:

n=1bnsinnπxL=sinπxL+3sin3πxL.\sum_{n=1}^\infty b_n \sin\frac{n\pi x}{L} = \sin\frac{\pi x}{L} + 3\sin\frac{3\pi x}{L}.

Comparing coefficients of the orthogonal sine modes: b1=1b_1 = 1, b3=3b_3 = 3, all other bn=0b_n = 0. Hence

u(x,t)=sinπxLec2π2t/L2+3sin3πxLe9c2π2t/L2.\boxed{u(x,t) = \sin\frac{\pi x}{L}\,e^{-c^2\pi^2 t/L^2} + 3\sin\frac{3\pi x}{L}\,e^{-9c^2\pi^2 t/L^2}.}

The second mode decays 9×9\times faster than the first.

partial-differential-equationsboundary-value-problemsheat-equation
4long12 marks

(a) State Green's theorem in the plane. Using it, evaluate C(3x28y2)dx+(4y6xy)dy\oint_C (3x^2 - 8y^2)\,dx + (4y - 6xy)\,dy, where CC is the boundary of the region bounded by y=xy = \sqrt{x} and y=x2y = x^2. [7]

(b) Evaluate the surface integral SFn^dS\iint_S \mathbf{F}\cdot \hat{n}\,dS where F=4xi^2y2j^+z2k^\mathbf{F} = 4x\,\hat{i} - 2y^2\,\hat{j} + z^2\,\hat{k} and SS is the surface bounding the region x2+y2=4x^2 + y^2 = 4, z=0z=0, z=3z=3, using the divergence theorem. [5]

(a) Green's Theorem [7]

Statement. If CC is a positively oriented, piecewise-smooth, simple closed curve bounding a region RR, and M,NM,N have continuous partial derivatives on RR, then

CMdx+Ndy=R(NxMy)dA.\oint_C M\,dx + N\,dy = \iint_R \left(\frac{\partial N}{\partial x} - \frac{\partial M}{\partial y}\right) dA.

Application. Here M=3x28y2M = 3x^2 - 8y^2, N=4y6xyN = 4y - 6xy.

Nx=6y,My=16y    NxMy=6y+16y=10y.\frac{\partial N}{\partial x} = -6y, \qquad \frac{\partial M}{\partial y} = -16y \;\Rightarrow\; \frac{\partial N}{\partial x} - \frac{\partial M}{\partial y} = -6y + 16y = 10y.

The region between y=x2y=x^2 and y=xy=\sqrt{x} (intersecting at x=0,1x=0,1) has x2yxx^2 \le y \le \sqrt{x}, 0x10\le x\le1:

R10ydA=01 ⁣ ⁣x2x10ydydx=015[y2]x2xdx=015(xx4)dx.\iint_R 10y\,dA = \int_0^1\!\!\int_{x^2}^{\sqrt{x}} 10y\,dy\,dx = \int_0^1 5\big[y^2\big]_{x^2}^{\sqrt{x}}dx = \int_0^1 5\,(x - x^4)\,dx. =5[x22x55]01=5(1215)=5310=32.= 5\left[\frac{x^2}{2} - \frac{x^5}{5}\right]_0^1 = 5\left(\frac12 - \frac15\right) = 5\cdot\frac{3}{10} = \boxed{\frac{3}{2}}.

(b) Surface Integral via Divergence Theorem [5]

With F=4xi^2y2j^+z2k^\mathbf{F} = 4x\,\hat i - 2y^2\,\hat j + z^2\,\hat k,

F=44y+2z.\nabla\cdot\mathbf{F} = 4 - 4y + 2z.

By the divergence theorem, SFn^dS=V(44y+2z)dV\displaystyle\iint_S \mathbf{F}\cdot\hat n\,dS = \iiint_V (4 - 4y + 2z)\,dV over the cylinder x2+y24x^2+y^2\le4, 0z30\le z\le3.

By symmetry V(4y)dV=0\iiint_V (-4y)\,dV = 0 (odd in yy). Volume V=π(2)2(3)=12πV = \pi(2)^2(3) = 12\pi.

V4dV=4(12π)=48π.\iiint_V 4\,dV = 4(12\pi) = 48\pi. V2zdV=2(π4)03zdz=8π92=36π.\iiint_V 2z\,dV = 2\cdot(\pi\cdot4)\int_0^3 z\,dz = 8\pi\cdot\frac{9}{2} = 36\pi.

Total:

SFn^dS=48π+36π=84π.\boxed{\iint_S \mathbf{F}\cdot\hat n\,dS = 48\pi + 36\pi = 84\pi.}
line-integralsgreens-theoremsurface-integrals
B

Section B: Short Answer Questions

Attempt all / any as specified.

7 questions
5short8 marks

(a) Find the Laplace transform of the periodic square wave function of period 2a2a defined by f(t)=kf(t) = k for 0<t<a0 < t < a and f(t)=kf(t) = -k for a<t<2aa < t < 2a. [4]

(b) Using the convolution theorem, find L1{1s2(s2+1)}\mathcal{L}^{-1}\left\{\dfrac{1}{s^2(s^2+1)}\right\}. [4]

(a) Laplace Transform of a Square Wave [4]

For a periodic function of period pp, L{f}=11eps0pestf(t)dt\mathcal{L}\{f\} = \dfrac{1}{1-e^{-ps}}\displaystyle\int_0^p e^{-st}f(t)\,dt. Here p=2ap = 2a.

02aestf(t)dt=k0aestdtka2aestdt=ks(1eas)ks(ease2as).\int_0^{2a} e^{-st}f(t)\,dt = k\int_0^a e^{-st}dt - k\int_a^{2a} e^{-st}dt = \frac{k}{s}(1 - e^{-as}) - \frac{k}{s}(e^{-as} - e^{-2as}). =ks(12eas+e2as)=ks(1eas)2.= \frac{k}{s}\big(1 - 2e^{-as} + e^{-2as}\big) = \frac{k}{s}(1 - e^{-as})^2.

Hence

L{f}=k(1eas)2s(1e2as)=k(1eas)2s(1eas)(1+eas)=ks1eas1+eas.\mathcal{L}\{f\} = \frac{k(1-e^{-as})^2}{s(1-e^{-2as})} = \frac{k(1-e^{-as})^2}{s(1-e^{-as})(1+e^{-as})} = \frac{k}{s}\cdot\frac{1-e^{-as}}{1+e^{-as}}. L{f}=kstanhas2.\boxed{\mathcal{L}\{f\} = \frac{k}{s}\tanh\frac{as}{2}.}

(b) Inverse Transform by Convolution [4]

Write 1s2(s2+1)=1s21s2+1\dfrac{1}{s^2(s^2+1)} = \dfrac{1}{s^2}\cdot\dfrac{1}{s^2+1} with L1{1/s2}=t\mathcal{L}^{-1}\{1/s^2\}=t and L1{1/(s2+1)}=sint\mathcal{L}^{-1}\{1/(s^2+1)\}=\sin t.

By the convolution theorem L1{F(s)G(s)}=0tf(τ)g(tτ)dτ\mathcal{L}^{-1}\{F(s)G(s)\} = \int_0^t f(\tau)g(t-\tau)\,d\tau:

L1{1s2(s2+1)}=0tτsin(tτ)dτ.\mathcal{L}^{-1}\left\{\frac{1}{s^2(s^2+1)}\right\} = \int_0^t \tau\sin(t-\tau)\,d\tau.

Integrating by parts (or directly): 0tτsin(tτ)dτ=[τcos(tτ)]0t0tcos(tτ)dτ=t[sin(tτ)(1)]...\int_0^t \tau\sin(t-\tau)d\tau = \big[\tau\cos(t-\tau)\big]_0^t - \int_0^t \cos(t-\tau)d\tau = t - \big[\sin(t-\tau)(-1)\big]... evaluating gives

=tsint.= t - \sin t. L1{1s2(s2+1)}=tsint.\boxed{\mathcal{L}^{-1}\left\{\frac{1}{s^2(s^2+1)}\right\} = t - \sin t.}

(Check by partial fractions: 1s2(s2+1)=1s21s2+1\frac{1}{s^2(s^2+1)} = \frac{1}{s^2} - \frac{1}{s^2+1}, inverting gives tsintt - \sin t.)

laplace-transformsconvolution
6short8 marks

Find the series solution of the ordinary differential equation

2xd2ydx2+dydx+y=02x\frac{d^2y}{dx^2} + \frac{dy}{dx} + y = 0

about the regular singular point x=0x=0 using the Frobenius method. Determine the indicial equation and obtain at least one of the two linearly independent solutions.

Frobenius Series Solution of 2xy+y+y=02xy'' + y' + y = 0 [8]

The point x=0x=0 is a regular singular point. Assume y=n=0anxn+ry = \sum_{n=0}^\infty a_n x^{n+r}, a00a_0\ne0. Then

y=(n+r)anxn+r1,y=(n+r)(n+r1)anxn+r2.y' = \sum (n+r)a_n x^{n+r-1}, \qquad y'' = \sum (n+r)(n+r-1)a_n x^{n+r-2}.

Substitute:

2(n+r)(n+r1)anxn+r1+(n+r)anxn+r1+anxn+r=0.2\sum (n+r)(n+r-1)a_n x^{n+r-1} + \sum (n+r)a_n x^{n+r-1} + \sum a_n x^{n+r} = 0.

Combine the first two: (n+r)[2(n+r1)+1]anxn+r1=(n+r)(2n+2r1)anxn+r1\sum (n+r)\big[2(n+r-1)+1\big]a_n x^{n+r-1} = \sum (n+r)(2n+2r-1)a_n x^{n+r-1}.

Indicial equation (coefficient of lowest power xr1x^{r-1}, n=0n=0):

r(2r1)a0=0    r(2r1)=0    r=0 or r=12.r(2r-1)a_0 = 0 \;\Rightarrow\; \boxed{r(2r-1)=0} \;\Rightarrow\; r = 0 \text{ or } r = \tfrac12.

Recurrence relation. Matching xn+rx^{n+r} (i.e. shifting index nn1n\to n-1 in the first sum):

(n+r)(2n+2r1)an+an1=0    an=an1(n+r)(2n+2r1).(n+r)(2n+2r-1)a_n + a_{n-1} = 0 \;\Rightarrow\; a_n = \frac{-a_{n-1}}{(n+r)(2n+2r-1)}.

Solution for r=12r = \tfrac12. Then (n+12)(2n)=n(2n+1)(n+\tfrac12)(2n) = n(2n+1), so an=an1n(2n+1)a_n = \dfrac{-a_{n-1}}{n(2n+1)}.

a1=a013=a03!,a2=a125=a05!,a3=a237=a07!,a_1 = \frac{-a_0}{1\cdot3} = -\frac{a_0}{3!},\quad a_2 = \frac{-a_1}{2\cdot5} = \frac{a_0}{5!},\quad a_3 = \frac{-a_2}{3\cdot7}=-\frac{a_0}{7!},\dots

In general an=(1)na0(2n+1)!a_n = \dfrac{(-1)^n a_0}{(2n+1)!}. With a0=1a_0=1:

y1=x1/2n=0(1)nxn(2n+1)!=1xn=0(1)n(x)2n+1(2n+1)!=sinxx.y_1 = x^{1/2}\sum_{n=0}^\infty \frac{(-1)^n x^n}{(2n+1)!} = \frac{1}{\sqrt{x}}\sum_{n=0}^\infty \frac{(-1)^n (\sqrt{x})^{2n+1}}{(2n+1)!} = \boxed{\frac{\sin\sqrt{x}}{\sqrt{x}}.}

Solution for r=0r = 0. Here an=an1n(2n1)a_n = \dfrac{-a_{n-1}}{n(2n-1)}, giving an=(1)na0(2n)!a_n = \dfrac{(-1)^n a_0}{(2n)!}, so

y2=n=0(1)nxn(2n)!=cosx.y_2 = \sum_{n=0}^\infty \frac{(-1)^n x^n}{(2n)!} = \cos\sqrt{x}.

The two linearly independent solutions are sinx\sin\sqrt x and cosx\cos\sqrt x (up to the 1/x1/\sqrt x factor noted above); general solution y=c1cosx+c2sinxy = c_1\cos\sqrt x + c_2\sin\sqrt x.

series-solutions-odefrobenius-method
7short8 marks

(a) Prove the recurrence relation ddx[xnJn(x)]=xnJn1(x)\dfrac{d}{dx}\left[x^n J_n(x)\right] = x^n J_{n-1}(x) for Bessel functions. [4]

(b) Using Rodrigues' formula, obtain the Legendre polynomials P2(x)P_2(x) and P3(x)P_3(x), and verify the orthogonality relation 11P2(x)P3(x)dx=0\displaystyle\int_{-1}^{1} P_2(x)P_3(x)\,dx = 0. [4]

(a) Bessel Recurrence Relation [4]

Use the series Jn(x)=k=0(1)kk!Γ(n+k+1)(x2)2k+nJ_n(x) = \sum_{k=0}^\infty \dfrac{(-1)^k}{k!\,\Gamma(n+k+1)}\left(\dfrac{x}{2}\right)^{2k+n}.

Then

xnJn(x)=k=0(1)kk!Γ(n+k+1)x2k+2n22k+n.x^n J_n(x) = \sum_{k=0}^\infty \frac{(-1)^k}{k!\,\Gamma(n+k+1)}\frac{x^{2k+2n}}{2^{2k+n}}.

Differentiate term by term:

ddx[xnJn(x)]=k=0(1)k(2k+2n)k!Γ(n+k+1)x2k+2n122k+n.\frac{d}{dx}\big[x^n J_n(x)\big] = \sum_{k=0}^\infty \frac{(-1)^k (2k+2n)}{k!\,\Gamma(n+k+1)}\frac{x^{2k+2n-1}}{2^{2k+n}}.

Since 2k+2n=2(n+k)2k+2n = 2(n+k) and Γ(n+k+1)=(n+k)Γ(n+k)\Gamma(n+k+1) = (n+k)\Gamma(n+k):

=k=0(1)kk!Γ(n+k)x2k+2n122k+n1=xnk=0(1)kk!Γ((n1)+k+1)x2k+n122k+n1=xnJn1(x).= \sum_{k=0}^\infty \frac{(-1)^k}{k!\,\Gamma(n+k)}\frac{x^{2k+2n-1}}{2^{2k+n-1}} = x^n\sum_{k=0}^\infty \frac{(-1)^k}{k!\,\Gamma((n-1)+k+1)}\frac{x^{2k+n-1}}{2^{2k+n-1}} = x^n J_{n-1}(x). ddx[xnJn(x)]=xnJn1(x).\boxed{\frac{d}{dx}\big[x^n J_n(x)\big] = x^n J_{n-1}(x).} \quad\blacksquare

(b) Legendre Polynomials via Rodrigues' Formula [4]

Rodrigues' formula: Pn(x)=12nn!dndxn(x21)nP_n(x) = \dfrac{1}{2^n n!}\dfrac{d^n}{dx^n}\big(x^2-1\big)^n.

P2P_2: (x21)2=x42x2+1(x^2-1)^2 = x^4 - 2x^2 + 1. Second derivative: 12x2412x^2 - 4. So

P2(x)=1222!(12x24)=12x248=12(3x21).P_2(x) = \frac{1}{2^2\cdot2!}(12x^2-4) = \frac{12x^2-4}{8} = \boxed{\tfrac12(3x^2-1)}.

P3P_3: (x21)3=x63x4+3x21(x^2-1)^3 = x^6 - 3x^4 + 3x^2 - 1. Third derivative: 120x372x120x^3 - 72x. So

P3(x)=1233!(120x372x)=120x372x48=12(5x33x).P_3(x) = \frac{1}{2^3\cdot3!}(120x^3 - 72x) = \frac{120x^3-72x}{48} = \boxed{\tfrac12(5x^3-3x)}.

Orthogonality check. The product P2P3=14(3x21)(5x33x)P_2 P_3 = \tfrac14(3x^2-1)(5x^3-3x) is an odd function of xx. Integrating an odd function over the symmetric interval [1,1][-1,1] gives

11P2(x)P3(x)dx=0,\int_{-1}^1 P_2(x)P_3(x)\,dx = 0,

verifying orthogonality. (Directly: 1411(15x59x35x3+3x)dx=0\tfrac14\int_{-1}^1(15x^5 - 9x^3 - 5x^3 + 3x)dx = 0 since each integrand term is odd.)

bessel-functionslegendre-functions
8short8 marks

(a) State the Cauchy-Riemann equations. Show that the function f(z)=zˉf(z) = \bar{z} is nowhere analytic. [3]

(b) Show that u(x,y)=x33xy2+3x23y2+1u(x,y) = x^3 - 3xy^2 + 3x^2 - 3y^2 + 1 is harmonic and find its harmonic conjugate v(x,y)v(x,y) such that f(z)=u+ivf(z) = u + iv is analytic. [5]

(a) Cauchy-Riemann Equations; f(z)=zˉf(z)=\bar z [3]

For f(z)=u(x,y)+iv(x,y)f(z) = u(x,y) + i\,v(x,y) to be analytic, the Cauchy-Riemann (CR) equations must hold and the partials be continuous:

ux=vy,uy=vx.\frac{\partial u}{\partial x} = \frac{\partial v}{\partial y}, \qquad \frac{\partial u}{\partial y} = -\frac{\partial v}{\partial x}.

For f(z)=zˉ=xiyf(z) = \bar z = x - iy: u=xu = x, v=yv = -y. Then

ux=1,vy=1    uxvy.u_x = 1, \quad v_y = -1 \;\Rightarrow\; u_x \ne v_y.

The first CR equation fails at every point, so f(z)=zˉf(z)=\bar z is nowhere analytic. \blacksquare

(b) Harmonic Function and Conjugate [5]

Given u=x33xy2+3x23y2+1u = x^3 - 3xy^2 + 3x^2 - 3y^2 + 1.

ux=3x23y2+6x,uxx=6x+6;uy=6xy6y,uyy=6x6.u_x = 3x^2 - 3y^2 + 6x, \quad u_{xx} = 6x + 6; \qquad u_y = -6xy - 6y, \quad u_{yy} = -6x - 6. uxx+uyy=(6x+6)+(6x6)=0,u_{xx} + u_{yy} = (6x+6) + (-6x-6) = 0,

so uu is harmonic.

Find vv via CR equations. vy=ux=3x23y2+6xv_y = u_x = 3x^2 - 3y^2 + 6x. Integrate w.r.t. yy:

v=3x2yy3+6xy+g(x).v = 3x^2 y - y^3 + 6xy + g(x).

Differentiate w.r.t. xx: vx=6xy+6y+g(x)v_x = 6xy + 6y + g'(x). CR requires vx=uy=6xy+6yv_x = -u_y = 6xy + 6y. Hence g(x)=0g(x)=Cg'(x) = 0 \Rightarrow g(x) = C.

v(x,y)=3x2yy3+6xy+C.\boxed{v(x,y) = 3x^2 y - y^3 + 6xy + C.}

Analytic function. f(z)=u+iv=z3+3z2+1+iCf(z) = u + iv = z^3 + 3z^2 + 1 + iC (one checks z3=(x33xy2)+i(3x2yy3)z^3 = (x^3-3xy^2) + i(3x^2y - y^3) and 3z2=(3x23y2)+i(6xy)3z^2 = (3x^2-3y^2)+i(6xy)).

complex-analysiscauchy-riemannanalytic-functions
9short8 marks

(a) State Cauchy's residue theorem. [2]

(b) Using the residue theorem, evaluate Cz+1z22zdz\displaystyle\oint_C \frac{z+1}{z^2 - 2z}\,dz, where CC is the circle z=3|z| = 3 described in the positive sense. [6]

(a) Cauchy's Residue Theorem [2]

If f(z)f(z) is analytic inside and on a simple closed contour CC except at a finite number of isolated singular points z1,z2,,zkz_1,z_2,\dots,z_k inside CC, then

Cf(z)dz=2πij=1kResz=zjf(z),\oint_C f(z)\,dz = 2\pi i \sum_{j=1}^k \operatorname{Res}_{z=z_j} f(z),

where CC is traversed in the positive (counter-clockwise) sense.

(b) Evaluation [6]

f(z)=z+1z22z=z+1z(z2).f(z) = \frac{z+1}{z^2 - 2z} = \frac{z+1}{z(z-2)}.

Simple poles at z=0z=0 and z=2z=2, both inside z=3|z|=3.

Residue at z=0z=0:

Resz=0=limz0zz+1z(z2)=0+102=12.\operatorname{Res}_{z=0} = \lim_{z\to0} z\cdot\frac{z+1}{z(z-2)} = \frac{0+1}{0-2} = -\frac12.

Residue at z=2z=2:

Resz=2=limz2(z2)z+1z(z2)=2+12=32.\operatorname{Res}_{z=2} = \lim_{z\to2}(z-2)\cdot\frac{z+1}{z(z-2)} = \frac{2+1}{2} = \frac32.

By the residue theorem,

Cf(z)dz=2πi(12+32)=2πi1=2πi.\oint_C f(z)\,dz = 2\pi i\left(-\frac12 + \frac32\right) = 2\pi i \cdot 1 = \boxed{2\pi i}.
complex-analysiscontour-integrationresidue-theorem
10short8 marks

(a) Form the partial differential equation by eliminating the arbitrary function ff from z=f(x2+y2+z2)z = f(x^2 + y^2 + z^2). [3]

(b) Solve the linear partial differential equation x(yz)p+y(zx)q=z(xy)x(y - z)\,p + y(z - x)\,q = z(x - y) using Lagrange's method, where p=z/xp = \partial z/\partial x and q=z/yq = \partial z/\partial y. [5]

(a) Form the PDE from z=f(x2+y2+z2)z = f(x^2+y^2+z^2) [3]

Let w=x2+y2+z2w = x^2+y^2+z^2, so z=f(w)z = f(w). Differentiate:

p=zx=f(w)(2x+2zp)    p=2f(w)(x+zp),p = z_x = f'(w)(2x + 2z\,p) \;\Rightarrow\; p = 2f'(w)(x + zp), q=zy=f(w)(2y+2zq)    q=2f(w)(y+zq).q = z_y = f'(w)(2y + 2z\,q) \;\Rightarrow\; q = 2f'(w)(y + zq).

Dividing to eliminate f(w)f'(w):

pq=x+zpy+zq    p(y+zq)=q(x+zp)    py+pzq=qx+qzp.\frac{p}{q} = \frac{x + zp}{y + zq} \;\Rightarrow\; p(y + zq) = q(x + zp) \;\Rightarrow\; py + pzq = qx + qzp.

The pzqpzq terms cancel:

pyqx=0i.e.ypxq=0.\boxed{py - qx = 0 \quad\text{i.e.}\quad y\,p - x\,q = 0.}

(b) Lagrange's Method [5]

The equation x(yz)p+y(zx)q=z(xy)x(y-z)p + y(z-x)q = z(x-y) has auxiliary (Lagrange) equations

dxx(yz)=dyy(zx)=dzz(xy).\frac{dx}{x(y-z)} = \frac{dy}{y(z-x)} = \frac{dz}{z(x-y)}.

First multipliers (1,1,1)(1,1,1): numerators sum dx+dy+dzdx+dy+dz, denominators sum

x(yz)+y(zx)+z(xy)=xyxz+yzxy+xzyz=0.x(y-z)+y(z-x)+z(x-y) = xy-xz+yz-xy+xz-yz = 0.

So dx+dy+dz=0x+y+z=c1dx+dy+dz = 0 \Rightarrow x+y+z = c_1.

Second multipliers (1/x,1/y,1/z)(1/x, 1/y, 1/z): numerator dxx+dyy+dzz\dfrac{dx}{x}+\dfrac{dy}{y}+\dfrac{dz}{z}, denominators sum

(yz)+(zx)+(xy)=0.(y-z)+(z-x)+(x-y) = 0.

So dxx+dyy+dzz=0ln(xyz)=constxyz=c2\dfrac{dx}{x}+\dfrac{dy}{y}+\dfrac{dz}{z} = 0 \Rightarrow \ln(xyz) = \text{const} \Rightarrow xyz = c_2.

General solution:

Φ(x+y+z,  xyz)=0,\boxed{\Phi(x+y+z,\; xyz) = 0,}

or equivalently x+y+z=F(xyz)x+y+z = F(xyz) for an arbitrary function FF.

partial-differential-equationslagrange-methodfirst-order-pde
11short8 marks

(a) Show that the vector field F=(2xy+z3)i^+x2j^+3xz2k^\mathbf{F} = (2xy + z^3)\hat{i} + x^2\hat{j} + 3xz^2\hat{k} is conservative, and find a scalar potential ϕ\phi such that F=ϕ\mathbf{F} = \nabla\phi. [4]

(b) Verify Stokes' theorem for F=yi^+zj^+xk^\mathbf{F} = y\,\hat{i} + z\,\hat{j} + x\,\hat{k} over the upper half of the surface of the sphere x2+y2+z2=1x^2 + y^2 + z^2 = 1 bounded by its projection on the xyxy-plane. [4]

(a) Conservative Field and Potential [4]

F=(2xy+z3)i^+x2j^+3xz2k^\mathbf{F} = (2xy + z^3)\hat i + x^2\hat j + 3xz^2\hat k. Compute ×F\nabla\times\mathbf{F}:

  • ii: y(3xz2)z(x2)=00=0\partial_y(3xz^2) - \partial_z(x^2) = 0 - 0 = 0
  • jj: z(2xy+z3)x(3xz2)=3z23z2=0\partial_z(2xy+z^3) - \partial_x(3xz^2) = 3z^2 - 3z^2 = 0
  • kk: x(x2)y(2xy+z3)=2x2x=0\partial_x(x^2) - \partial_y(2xy+z^3) = 2x - 2x = 0

Since ×F=0\nabla\times\mathbf{F} = \mathbf{0}, the field is conservative.

Potential. ϕx=2xy+z3ϕ=x2y+xz3+g(y,z)\phi_x = 2xy + z^3 \Rightarrow \phi = x^2 y + xz^3 + g(y,z). Then ϕy=x2+gy=x2gy=0\phi_y = x^2 + g_y = x^2 \Rightarrow g_y = 0. And ϕz=3xz2+gz=3xz2gz=0\phi_z = 3xz^2 + g_z = 3xz^2 \Rightarrow g_z = 0. So g=Cg = C.

ϕ(x,y,z)=x2y+xz3+C.\boxed{\phi(x,y,z) = x^2 y + xz^3 + C.}

(b) Verify Stokes' Theorem [4]

Stokes: CFdr=S(×F)n^dS\displaystyle\oint_C \mathbf{F}\cdot d\mathbf{r} = \iint_S (\nabla\times\mathbf{F})\cdot\hat n\,dS, with F=yi^+zj^+xk^\mathbf{F} = y\hat i + z\hat j + x\hat k and SS the upper hemisphere of x2+y2+z2=1x^2+y^2+z^2=1, boundary C:x2+y2=1,z=0C:\,x^2+y^2=1,\,z=0.

Line integral. Parametrize CC: x=cosθ,y=sinθ,z=0x=\cos\theta,\,y=\sin\theta,\,z=0, 0θ2π0\le\theta\le2\pi. On CC, Fdr=ydx+zdy+xdz=sinθ(sinθdθ)+0+0\mathbf{F}\cdot d\mathbf{r} = y\,dx + z\,dy + x\,dz = \sin\theta(-\sin\theta\,d\theta) + 0 + 0.

C=02πsin2θdθ=π.\oint_C = \int_0^{2\pi} -\sin^2\theta\,d\theta = -\pi.

Surface integral. ×F=(yxzz)...\nabla\times\mathbf{F} = (\partial_y x - \partial_z z)... compute:

  • ii: yxzz=01=1\partial_y x - \partial_z z = 0 - 1 = -1
  • jj: zyxx=01=1\partial_z y - \partial_x x = 0 - 1 = -1
  • kk: xzyy=01=1\partial_x z - \partial_y y = 0 - 1 = -1

So ×F=i^j^k^\nabla\times\mathbf{F} = -\hat i - \hat j - \hat k. By Stokes the flux through the hemisphere equals the flux of the (constant) curl through its planar boundary disk DD in the xyxy-plane with n^=k^\hat n = \hat k (consistent orientation):

S(×F)n^dS=D(1)dA=(area of unit disk)=π.\iint_S (\nabla\times\mathbf{F})\cdot\hat n\,dS = \iint_D (-1)\,dA = -(\text{area of unit disk}) = -\pi.

Conclusion. Both sides equal π-\pi, so Stokes' theorem is verified.

line-integralsconservative-fieldsstokes-theorem

Frequently asked questions

Where can I find the BE Computer Engineering (IOE, TU) Engineering Mathematics III (IOE, SH 501) question paper 2079?
The full BE Computer Engineering (IOE, TU) Engineering Mathematics III (IOE, SH 501) 2079 (regular) question paper is available free on Kekkei. You can read every question online and attempt the paper under timed exam conditions.
Does the Engineering Mathematics III (IOE, SH 501) 2079 paper come with solutions?
Yes. Every question on this Engineering Mathematics III (IOE, SH 501) past paper includes a step-by-step solution, plus instant AI feedback when you attempt it on Kekkei.
How many marks is the BE Computer Engineering (IOE, TU) Engineering Mathematics III (IOE, SH 501) 2079 paper?
The BE Computer Engineering (IOE, TU) Engineering Mathematics III (IOE, SH 501) 2079 paper carries 80 full marks and is meant to be completed in 180 minutes, across 11 questions.
Is practising this Engineering Mathematics III (IOE, SH 501) past paper free?
Yes — reading and attempting this Engineering Mathematics III (IOE, SH 501) past paper on Kekkei is completely free.