Laplace transform Study Guide

📖 Core Concepts Laplace Transform – converts a time‑domain function \(f(t)\) (defined for \(t\ge0\)) into a complex‑frequency function \(F(s)=\displaystyle\int{0}^{\infty}f(t)e^{-st}\,dt\). Variable \(s\) – complex number \(s=\sigma+j\omega\); \(\sigma=\operatorname{Re}(s)\) determines convergence, \(\omega=\operatorname{Im}(s)\) relates to oscillation. Unilateral vs. Bilateral – Unilateral integrates from \(0\) to \(\infty\) (used for causal systems). Bilateral integrates over \((-\infty,\infty)\) and reduces to unilateral when the function is multiplied by the Heaviside step \(u(t)\). Region of Convergence (ROC) – half‑plane \(\operatorname{Re}(s)>a\) where the integral converges; its location tells you about growth/decay and system stability. Key Property – differentiation in time ↔ multiplication by \(s\) (plus initial‑value terms); integration in time ↔ division by \(s\). This turns ODEs into algebraic equations. --- 📌 Must Remember Linearity: \(\mathcal{L}\{a f+b g\}=aF(s)+bG(s)\). Differentiation: \(\mathcal{L}\{f'(t)\}=sF(s)-f(0)\); \(\mathcal{L}\{f^{(n)}(t)\}=s^{n}F(s)-s^{n-1}f(0)-\dots-f^{(n-1)}(0)\). Integration: \(\mathcal{L}\{\int{0}^{t}f(\tau)d\tau\}=F(s)/s\). Convolution: \(\mathcal{L}\{(fg)(t)\}=F(s)G(s)\). Initial‑Value Theorem: \(\displaystyle\lim{t\to0^{+}}f(t)=\lim{s\to\infty}sF(s)\) (if limits exist). Final‑Value Theorem: \(\displaystyle\lim{t\to\infty}f(t)=\lim{s\to0}sF(s)\) provided all poles of \(sF(s)\) lie in the left half‑plane. Stability Criterion: A linear time‑invariant (LTI) system is BIBO‑stable iff the ROC includes the line \(\operatorname{Re}(s)=0\) → all poles have \(\operatorname{Re}(s)<0\). Common Pairs (remember the form, not just the table): \(1 \;\leftrightarrow\; 1/s\) \(t^{n} \;\leftrightarrow\; n!/s^{n+1}\) \(e^{a t} \;\leftrightarrow\; 1/(s-a)\) \(\sin(\omega t) \;\leftrightarrow\; \omega/(s^{2}+\omega^{2})\) \(\cos(\omega t) \;\leftrightarrow\; s/(s^{2}+\omega^{2})\) --- 🔄 Key Processes Solve an ODE with Laplace: Take \(\mathcal{L}\) of every term → replace \(f^{(n)}\) with \(s^{n}F(s)-\)initial terms. Algebraically solve for \(F(s)\). Perform partial‑fraction expansion (if rational). Inverse‑transform each term using the table or known pairs. Partial‑Fraction Expansion (Rational \(F(s)\)): Factor denominator into linear/quadratic factors. Write \(F(s)=\sum \frac{Ai}{s-pi}+\sum\frac{Bj s + Cj}{(s-qj)^2+\omegaj^2}\). Solve for coefficients \(Ai,Bj,Cj\) (cover‑up method or system of equations). Inverse via Bromwich Integral (conceptual): \(f(t)=\frac{1}{2\pi i}\int{\gamma-i\infty}^{\gamma+i\infty}F(s)e^{st}\,ds\). In practice, use residues: each pole \(p\) contributes \(\operatorname{Res}[F(s)e^{st},p]\). --- 🔍 Key Comparisons Unilateral vs. Bilateral Unilateral: integrates from \(0\) → assumes causality; used for physical systems. Bilateral: integrates over \((-\infty,\infty)\) → useful for two‑sided signals, Fourier connection. Laplace vs. Fourier Laplace: \(s=\sigma+j\omega\); includes exponential damping/growth (\(\sigma\)). Fourier: evaluates Laplace at \(s=j\omega\) only when ROC contains the imaginary axis. Laplace vs. Z‑Transform Laplace: continuous‑time, variable \(s\). Z‑Transform: discrete‑time, variable \(z=e^{sT}\) (with sampling period \(T\)). --- ⚠️ Common Misunderstandings Final‑Value Theorem misuse: applying it when any pole of \(sF(s)\) lies in the right half‑plane or on the imaginary axis gives wrong limits. ROC = Stability? ROC must include the \(j\omega\) axis for stability; merely having a half‑plane does not guarantee all poles are left‑half. Differentiation property omission: forgetting the subtraction of initial conditions \(f(0), f'(0),\dots\) leads to incorrect algebraic equations. Convolution vs. multiplication: convolution in time ↔ multiplication in \(s\); the reverse is also true—do not multiply time‑domain signals when you need convolution. --- 🧠 Mental Models / Intuition “s‑Domain is algebraic” – think of \(s\) as a placeholder that turns calculus (derivatives/integrals) into simple algebra (multiply/divide). Poles = system behavior: each pole \(p\) contributes a term \(e^{pt}\). Left‑half poles → decaying exponentials → stable; right‑half → growing → unstable. ROC as a “visibility window”: only the part of the complex plane where the transform “sees” the signal; moving the window right (larger \(\sigma\)) forces faster decay, ensuring convergence. --- 🚩 Exceptions & Edge Cases Impulse \(\delta(t)\): \(\mathcal{L}\{\delta(t)\}=1\) (no \(s\) dependence). Step \(u(t)\): also transforms to \(1/s\) – same as constant 1, but note the Heaviside function ensures causality. Conditional convergence on the boundary: some functions converge at \(\operatorname{Re}(s)=a\) (e.g., \(e^{at}u(t)\) at \(\sigma=a\)) – treat with care when applying theorems. Repeated poles: produce terms of the form \(t^k e^{pt}\) after inverse transform; require partial fractions with polynomial numerators. --- 📍 When to Use Which Algebraic solution vs. direct integration: Use Laplace when ODE has constant coefficients and/or non‑zero initial conditions; avoid for variable‑coefficient ODEs. Partial‑fraction vs. Table lookup: Partial‑fraction: whenever \(F(s)\) is a rational function not directly listed in the table. Table lookup: for standard forms (exponential, sinusoid, polynomial, step, delta). Bilateral Laplace: only when you need to relate to Fourier (e.g., analyzing signals that exist for \(t<0\)). Final‑value theorem: use to find steady‑state values after confirming all poles of \(sF(s)\) are in \(\operatorname{Re}(s)<0\). --- 👀 Patterns to Recognize Denominator of the form \((s-a)(s-b)\): expect sum of two exponentials \(C1 e^{at}+C2 e^{bt}\). Quadratic denominator \(s^{2}+ \omega^{2}\): signals are sinusoidal (sine/cosine) possibly multiplied by exponentials if a shift \(s-a\) is present. Repeated pole \( (s-p)^n\): leads to polynomial‑times‑exponential term \(t^{n-1}e^{pt}\). Factor \(s\) in denominator: indicates integration or step response; numerator constant → unit step; numerator \(1\) → ramp (if \(1/s^{2}\)). --- 🗂️ Exam Traps Choosing the wrong ROC: picking \(\operatorname{Re}(s)>a\) when the problem actually requires \(\operatorname{Re}(s)<a\) (e.g., left‑hand ROC for anti‑causal signals). Neglecting initial conditions: forgetting the \(-f(0)\) term in \(\mathcal{L}\{f'\}\) leads to missing constants in the solution. Final‑value theorem applied to unstable systems: yields a finite limit incorrectly; exam will test pole location first. Mixing unilateral and bilateral transforms: using bilateral formulas for a causal circuit problem will produce extra terms. Misreading transform pairs: \(\mathcal{L}\{t^{n}\}=n!/s^{n+1}\) – a common slip is to write \(n!/s^{n}\) (off by one power). ---