Laplace Transforms — Question 8

PDF ↗

Question 8

For ε≥0\varepsilon\ge 0, let x′=−εx−y+1,y′=x−εy,x(0)=y(0)=0,t≥0.x'=-\varepsilon x-y+1,\qquad y'=x-\varepsilon y, \qquad x(0)=y(0)=0,\quad t\ge 0. A student evaluates lim⁡s→0+sℒ{x}\lim_{s\to 0^+}s\mathcal L\{x\} and lim⁡s→0+sℒ{y}\lim_{s\to 0^+}s\mathcal L\{y\} and declares these the limiting state for every ε\varepsilon. For rational transforms, the final-value theorem requires all poles of the simplified sℒ{f}s\mathcal L\{f\} to have negative real part.

Tasks

  1. Find both transformed components for general ε\varepsilon and calculate the student’s two proposed limits.

  2. Solve the case ε=0\varepsilon=0. Check the theorem’s hypothesis and determine whether either component has a limit as t→∞t\to\infty.

  3. For ε>0\varepsilon>0, invert the transforms and establish the true limiting state. Explain why the theorem now applies.

  4. Show that yε→y0y_\varepsilon\to y_0 at each fixed time as ε→0+\varepsilon\to 0^+ but not uniformly on [0,∞)[0,\infty). Use times 2πn2\pi n to obtain a lower bound on the uniform error. Sketch y0y_0 and y1/2y_{1/2}.

Original worksheet page 1: question and worked solution for 5-11-008
Show solutionHide solution

Question 8 – Solution

Strategy. Check pole locations before applying the final-value theorem, and distinguish fixed-time convergence from an infinite-time assertion.

Step 1: Solve the transformed system. With D=(s+ε)2+1D=(s+\varepsilon)^2+1, the equations give F=s+εsD,G=1sD.F=\frac{s+\varepsilon}{sD},\qquad G=\frac 1{sD}. A sufficient convergence half-plane is Re⁡s>0\operatorname{Re}s>0. The proposed final values are ε/(1+ε2)\varepsilon/(1+\varepsilon^2) and 1/(1+ε2)1/(1+\varepsilon^2).

Step 2: Examine the undamped case. At ε=0\varepsilon=0, inversion yields x=sin⁡t,y=1−cos⁡t\boxed{x=\sin t,\ y=1-\cos t}. Neither has a limit: the values along successive maxima and minima differ. Both sFsF and sGsG have poles at ±i\pm i, so the stated theorem’s hypothesis fails. The finite algebraic values 00 and 11 do not prove convergence in time.

Step 3: Resolve the damped response. For ε>0\varepsilon>0, decompose, for example, G=11+ε2(1s−s+εD−εD).G=\frac 1{1+\varepsilon^2}\left(\frac 1s-\frac{s+\varepsilon}{D}-\frac{\varepsilon}{D}\right). Inverting this and the analogous decomposition of FF gives xε=ε−e−εt(εcos⁡t−sin⁡t)1+ε2,yε=1−e−εt(cos⁡t+εsin⁡t)1+ε2.x_\varepsilon=\frac{\varepsilon-e^{-\varepsilon t}(\varepsilon\cos t-\sin t)}{1+\varepsilon^2}, \qquad y_\varepsilon=\frac{1-e^{-\varepsilon t}(\cos t+\varepsilon\sin t)}{1+\varepsilon^2}. Both start at zero. The exponentially decaying terms leave the limit (ε,1)T/(1+ε2)\boxed{(\varepsilon,1)^T/(1+\varepsilon^2)}. Now the poles of sFsF and sGsG are −ε±i-\varepsilon\pm i, strictly in the left half-plane, as required. The limit also solves the equilibrium equations.

Step 4: Test uniformity on the whole half-line. At fixed tt, the formula gives yε(t)→1−cos⁡t=y0(t)y_\varepsilon(t)\to 1-\cos t=y_0(t). But y0(2πn)=0y_0(2\pi n)=0, whereas yε(2πn)=(1−e−2πnε)/(1+ε2)y_\varepsilon(2\pi n)=(1-e^{-2\pi n\varepsilon})/(1+\varepsilon^2). Consequently supt≥0|yε(t)−y0(t)|≥11+ε2.\boxed{\sup_{t\ge 0}|y_\varepsilon(t)-y_0(t)|\ge\frac 1{1+\varepsilon^2}.} This lower bound tends to 11, not 00. Fixed-time convergence therefore cannot be used to pass a limiting-state conclusion to the undamped system. The dotted guide is the damped limit 4/54/5 for ε=1/2\varepsilon=1/2.

See the diagram in the original worksheet below.

Original worksheet page 2: question and worked solution for 5-11-008

Original worksheet layout. Use Enlarge or open the PDF for a closer view.