Engineering Applications of Math: PDEs, Transforms & Signals

TAKEAWAYS:

• PDEs model physical systems: Heat, waves, and diffusion in engineering (e.g., NTC’s power grid stability, Daraz’s server cooling).
• Laplace/Z-transforms simplify circuits: Convert differential equations into algebraic equations for control systems (e.g., Ncell’s automated antenna tuning).
• Fourier transforms analyze signals: Decompose audio (WhatsApp calls), images (Khalti QR codes), and sensor data (NEPSE stock trends).
• Residue calculus speeds up integrals: Used in antenna design (Ncell) and fluid dynamics (NTC pipelines).
• Boundary conditions matter: Real-world constraints (e.g., fixed vs. free ends in bridges) change solutions entirely.
• Numerical methods bridge theory and practice: Finite difference/elements approximate PDEs for simulations (e.g., Kathmandu traffic rerouting).

1. Partial Differential Equations (PDEs) in Engineering

PDEs describe how quantities change in space and time. Three key types dominate engineering:

A. Heat Equation (Diffusion)

Definition: $$\frac{\partial u}{\partial t}=\alpha \left(\frac{{\partial }^{2}u}{\partial x^2}+\frac{{\partial }^{2}u}{\partial y^2}+\frac{{\partial }^{2}u}{\partial z^2}\right)$$

• $u(x,y,z,t)$: Temperature (or concentration) at point $(x,y,z)$ and time $t$.
• $\alpha$: Thermal diffusivity (material property).

Real-World Example 1: NTC Power Grid Stability Nepal’s power grid (managed by NTC) uses heat-equation models to predict cable overheating during peak demand (e.g., winter). Engineers solve for $u(x,t)$ to place cooling stations optimally. Worked Example 1: Steady-state heat in a cable Assume a long NTC cable of radius $R$ with uniform heat generation. The steady-state PDE reduces to: $$\frac{1}{r}\frac{d}{dr}\left(r\frac{du}{dr}\right)+\frac{Q}{\alpha }=0$$ Solution steps:

1. Boundary conditions:
   • $u(R)=T_{\text{ambient}}$ (surface temperature).
   • $\frac{du}{dr}(0)=0$ (symmetry at center).
2. Integrate twice: $$u(r)=-\frac{Q}{4\alpha }{r}^2+C_1\ln r+C_2$$ (The $\ln r$ term vanishes for physical solutions.)
3. Apply BCs to find $C_1=0$ and $C_2=T_{\text{ambient}}+\frac{Q{R}^2}{4\alpha }$. Final temperature profile: $$u(r)=T_{\text{ambient}}+\frac{Q}{4\alpha }(R^2-r^2)$$

Key Insight: The parabolic profile shows why thicker cables ($R$) can handle more heat without exceeding safe limits.

B. Wave Equation (Vibrations)

Definition: $$\frac{{\partial }^{2}u}{\partial t^2}=c^2\left(\frac{{\partial }^{2}u}{\partial x^2}+\frac{{\partial }^{2}u}{\partial y^2}\right)$$

• $c$: Wave speed (e.g., sound in air, signals in fiber optics).
• Applications: String vibrations (guitars), electromagnetic waves (Google Fiber), seismic waves.

Real-World Example 2: Kathmandu Traffic Light Optimization Traffic signals at busy intersections (e.g., Thapathali) model pedestrian/wheelchair flow using wave equations. The goal: minimize "stop-and-go" waves. Worked Example 2: Standing waves on a road segment Assume a 100m road with fixed ends (buildings). The general solution is: $$u(x,t)=\left(A\cos (kx)+B\sin (kx)\right)\cos (\omega t)$$ where $k=\frac{n\pi }{L}$ and $\omega =ck$. Boundary conditions:

• $u(0,t)=0$ (wall at $x=0$) → $A=0$.
• $u(L,t)=0$ (wall at $x=L=100m$) → $B\sin (n\pi )=0$ → $n=1,2,3,\dots$. First mode ($n=1$): $$u(x,t)=B\sin \left(\frac{\pi x}{100}\right)\cos \left(\frac{c\pi t}{100}\right)$$

Key Insight: The first mode shows why staggered traffic lights (phase shifts) reduce congestion.

C. Laplace’s Equation (Electrostatics/Fluid Flow)

Definition: $${\nabla }^{2}u=\frac{{\partial }^{2}u}{\partial x^2}+\frac{{\partial }^{2}u}{\partial y^2}+\frac{{\partial }^{2}u}{\partial z^2}=0$$

• $u$: Electric potential (V) or fluid velocity potential.
• Applications: Antenna design (Ncell), groundwater flow (Kathmandu Valley), capacitor geometry.

Real-World Example 3: Ncell Antenna Efficiency Ncell’s 4G antennas use Laplace’s equation to design shapes that minimize signal loss. The potential $u$ around an antenna satisfies: $${\nabla }^{2}u=0\text{with}u=V_0\text{ on antenna surface.}$$ Worked Example 3: 2D potential around a circular antenna In polar coordinates $(r,\theta )$: $$\frac{1}{r}\frac{\partial }{\partial r}\left(r\frac{\partial u}{\partial r}\right)+\frac{1}{r^2}\frac{{\partial }^{2}u}{\partial {\theta }^2}=0$$ Solution: Separation of variables gives: $$u(r,\theta )=\left(A_0+B_0\ln r\right)+\sum _{n=1}^{\infty }\left(A_n{r}^n+B_n{r}^{-n}\right)\left(C_n\cos (n\theta )+D_n\sin (n\theta )\right)$$ For a grounded plane at $r=R$ and potential $V_0$ on the antenna ($r=a$):

• $B_0=0$ (no $\ln r$ singularity).
• $B_n=0$ for $n\ge 1$ (bounded at infinity). Simplified solution (dipole approximation): $$u(r,\theta )=\frac{V_{0}a}{r}\cos \theta \text{(for }r\gg a\text{)}$$

Key Insight: The $1/r$ decay explains why tall antennas (larger $a$) improve signal range.

2. Transforms: From Theory to Engineering Tools

Transforms convert differential equations into algebraic equations, simplifying analysis.

A. Laplace Transform for Dynamic Systems

Definition: $$\mathcal{L}\{f(t)\}=F(s)={\int }_0^{\infty }f(t)e^{-st}dt$$ Key Properties:

Real-World Example 4: Ncell Automated Antenna Tuning Antenna systems use PID controllers (proportional-integral-derivative) to adjust gain dynamically. The Laplace transform converts the PID equation: $$G(s)=K_p+\frac{K_i}{s}+K_{d}s$$ into an algebraic form for stability analysis.

Worked Example 4: Step response of a PID-controlled antenna Given a second-order system: $$\frac{d^{2}y}{d{t}^2}+4\frac{dy}{dt}+3y=2u(t)$$ with PID control $u(t)=K_{p}e(t)+K_i\int e(t)dt+K_d\frac{de}{dt}$, where $e(t)=r(t)-y(t)$. Laplace transform steps:

1. Transform the system: $$s^{2}Y(s)+4sY(s)+3Y(s)=2U(s)$$
2. Express $U(s)$ in terms of $R(s)$ and $Y(s)$: $$U(s)=\left(K_p+\frac{K_i}{s}+K_{d}s\right)(R(s)-Y(s))$$
3. Substitute and solve for the closed-loop transfer function: $$\frac{Y(s)}{R(s)}=\frac{2\left(K_p+\frac{K_i}{s}+K_{d}s\right)}{s^2+(4-2K_d)s+(3-2K_p)+\frac{2K_i}{s}}$$ Stability condition: The denominator’s roots must have negative real parts. For $K_p=1$, $K_i=2$, $K_d=1$: $$s^2+2s+5+\frac{4}{s}=0⟹s^3+2s^2+5s+4=0$$

Key Insight: The phase margin (difference between $-180^{\circ }$ and phase at gain crossover) determines how "aggressive" the tuning is.

B. Fourier Transform for Signal Processing

Definition: $$\mathcal{F}\{f(t)\}=F(\omega )={\int }_{-\infty }^{\infty }f(t)e^{-i\omega t}dt$$ Applications:

• Audio compression: WhatsApp’s voice messages use Fourier transforms to remove redundant frequencies.
• Image compression: Khalti’s QR codes rely on Fourier-based algorithms to encode data efficiently.
• Stock analysis: NEPSE uses Fourier transforms to detect cyclical trends in share prices.

Worked Example 5: Bandwidth of a WhatsApp voice call A 3-second voice clip sampled at 8 kHz (standard for telephony) has:

• Time-domain signal: $f(t)$ (amplitude vs. time).
• Fourier transform: $F(\omega )$ shows frequency components up to 4 kHz (Nyquist limit). Steps:

1. Compute the Fourier transform of $f(t)$: $$F(\omega )={\int }_{-1.5}^{1.5}f(t)e^{-i\omega t}dt$$
2. Identify dominant frequencies (e.g., 300 Hz to 3.4 kHz for human speech).
3. Filter out frequencies above 3.4 kHz to reduce bandwidth.

Key Insight: The sharp drop at 4 kHz explains why higher sampling rates (e.g., 16 kHz) improve call quality but increase data usage.

C. Z-Transform for Discrete Systems

Definition: $$\mathcal{Z}\{f[n]\}=F(z)=\sum _{n=0}^{\infty }f[n]z^{-n}$$ Applications:

• Digital filters: Pathao’s navigation app uses Z-transforms to smooth GPS data.
• Control systems: NTC’s smart grids apply Z-transforms to discrete-time controllers.

Worked Example 6: Pathao’s GPS noise reduction A moving vehicle’s position is sampled every 0.1s. The noisy signal $y[n]$ is filtered using: $$y[n]=x[n]+0.5x[n-1]-0.1x[n-2]+\text{noise}$$ Z-transform steps:

1. Transform the recurrence: $$Y(z)=X(z)+0.5z^{-1}X(z)-0.1z^{-2}X(z)$$
2. Solve for the transfer function: $$H(z)=\frac{Y(z)}{X(z)}=1+0.5z^{-1}-0.1z^{-2}$$
3. Inverse Z-transform to find the impulse response: $$h[n]=\delta [n]+0.5\delta [n-1]-0.1\delta [n-2]$$

flowchart TD
    A["Noisy GPS Input\nx[n]"] --> B["Z-Transform\nX(z)"]
    B --> C["Filter Transfer Function\nH(z) = 1 + 0.5z⁻¹ - 0.1z⁻²"]
    C --> D["Inverse Z-Transform\nh[n]"]
    D --> E["Convolve with Input\nY(z) = H(z)X(z)"]
    E --> F["Smooth Output\ny[n]"]

Key Insight: The negative coefficient ($-0.1$) cancels high-frequency noise (e.g., sudden jumps in GPS data).

3. Residue Calculus: Engineering Shortcuts

Residue calculus evaluates complex integrals efficiently, critical for:

• Transient analysis: NTC’s fault current calculations.
• Antenna radiation patterns: Ncell’s 5G beamforming.
• Fluid dynamics: Daraz’s warehouse robot path planning.

Real-World Example 7: NTC Fault Current Calculation When a short circuit occurs, engineers compute the fault current using residues. The current $I(t)$ in an RL circuit is: $$I(t)={\mathcal{L}}^{-1}\left\{\frac{V_0}{L(s+R/L)}\right\}$$ Worked Example 7: Residue method for $I(t)$

1. Poles of $F(s)=\frac{V_0}{L(s+R/L)}$ are at $s=-R/L$.
2. Residue at $s=-R/L$: $$\text{Res}=\underset{s\to -R/L}{\lim }(s+R/L)\frac{V_0}{L(s+R/L)}=\frac{V_0}{L}$$
3. Inverse Laplace transform: $$I(t)=\frac{V_0}{L}{e}^{-(R/L)t}$$

Key Insight: The exponential decay explains why circuit breakers must act within $t\approx L/R$ to prevent damage.

4. Numerical Methods: Bridging Theory and Practice

Analytical solutions often require approximations. Common methods:

Worked Example 8: Finite difference for heat equation Discretize the 1D heat equation: $$\frac{u_i^{n+1}-u_i^n}{\Delta t}=\alpha \frac{u_{i+1}^n-2u_i^n+u_{i-1}^n}{(\Delta x{)}^2}$$ Stability condition: $\Delta t\le \frac{(\Delta x{)}^2}{2\alpha }$ (Courant criterion). Example: For $\alpha =0.01$, $\Delta x=0.1$, $\Delta t=0.005$: $$u_i^{n+1}=u_i^n+0.05(u_{i+1}^n-2u_i^n+u_{i-1}^n)$$

flowchart LR
    A["Initial Temp.\nu_i^n"] --> B["Compute\nu_{i+1}^n, u_{i-1}^n"]
    B --> C["Update\nu_i^{n+1} = u_i^n + 0.05*(u_{i+1}^n - 2u_i^n + u_{i-1}^n)"]
    C --> D["Check\nStability (Δt ≤ (Δx)²/2α)"]
    D -->|"Yes"| E["Next Time Step"]
    D -->|"No"| F["Reduce Δt"]

Key Insight: The stability condition ensures the simulation doesn’t "blow up" (unphysical oscillations).

In the Real World

1. Ncell’s 5G Networks

   • Idea: Fourier transforms decompose signals to allocate bandwidth efficiently.
   • How: Ncell’s beamforming uses Fourier series to direct signals toward users, reducing interference. The transform identifies dominant frequencies in each user’s channel, allowing simultaneous multi-user communication.

2. Khalti’s QR Code Payments

   • Idea: Fourier-based error correction (Reed-Solomon codes) ensures QR codes work even if partially damaged.
   • How: The code divides data into blocks, computes parity symbols using polynomial arithmetic (a Fourier-like transform), and reconstructs missing data during scanning.

3. NTC’s Smart Grid Optimization

   • Idea: Laplace transforms model the grid’s dynamic response to load changes.
   • How: Engineers use transfer functions to predict voltage drops during peak hours (e.g., winter in Kathmandu). The transform converts differential equations of generator dynamics into algebraic equations, simplifying controller design.

4. Pathao’s Traffic Routing

   • Idea: PDEs model pedestrian/vehicle flow to optimize routes.
   • How: Pathao’s algorithm solves a reaction-diffusion PDE to predict congestion. The solution guides drivers away from clogged areas (e.g., Thapathali during festivals).

5. NEPSE’s Stock Trend Analysis

   • Idea: Fourier transforms reveal cyclical patterns in share prices.
   • How: Analysts apply the transform to historical data to identify 4-year cycles in NEPSE’s index. This helps predict bull/bear markets (e.g., the 2015 crash’s precursors).

Exam Tip

What Examiners Look For

1. Physical Interpretation

   • Always state what $u(x,t)$, $F(\omega )$, or $Y(s)$ represents in the context (e.g., "temperature distribution," "frequency spectrum").
   • Example: For a PDE, write: "This solution describes how heat diffuses in an NTC cable of radius 0.3m with uniform heat generation."

2. Boundary Conditions

   • Explicitly list BCs and justify them (e.g., "insulated end" → $\frac{\partial u}{\partial x}=0$).
   • Common Pitfalls:
     • Forgetting initial conditions for time-dependent problems.
     • Misapplying BCs (e.g., using Dirichlet instead of Neumann).

3. Transform Properties

   • Show all steps when applying Laplace/Fourier/Z-transforms. Partial credit is given for intermediate results.
   • Example: For $\mathcal{L}\{t{e}^{at}\}$, write: $$\mathcal{L}\{t{e}^{at}\}=-\frac{d}{ds}\mathcal{L}\{e^{at}\}=-\frac{d}{ds}\left(\frac{1}{s-a}\right)=\frac{1}{(s-a{)}^2}$$

4. Numerical Methods

   • Derive stability conditions (e.g., Courant criterion) and state them clearly.
   • Example: For finite differences, write: "The scheme is stable if $\Delta t\le \frac{(\Delta x{)}^2}{2\alpha }$. Here, $\Delta t=0.005$ satisfies this for $\alpha =0.01$ and $\Delta x=0.1$."

5. Real-World Connections

   • Examiners love links to Nepalese contexts. Even if the question is abstract:
     • Relate PDEs to NTC’s power grids or Khalti’s servers.
     • Connect transforms to Ncell’s signal processing or Pathao’s algorithms.
   • Example: For a Laplace transform question, add: "This transfer function could model Ncell’s automatic gain control in a 4G base station, where $K_p$ adjusts for signal strength variations."

6. Graphs and Visuals

   • Sketch the solution profile (e.g., temperature vs. radius) or frequency spectrum.
   • Label axes with units (e.g., "Temperature (°C)" not just "u").
   • Example: For a wave equation solution, draw:
     • A snapshot of $u(x,t)$ at $t=0$ (initial shape).
     • A space-time plot showing wave propagation.

Common Mistakes to Avoid

• Ignoring units: Always include units in final answers (e.g., "A/m²" for heat flux).
• Assuming symmetry: Check if the problem is 1D, 2D, or 3D. A 2D heat equation needs $\frac{{\partial }^{2}u}{\partial y^2}$.
• Skipping validation: For numerical methods, verify stability or convergence (e.g., "For $\Delta x=0.05$, the error reduces by half when halving $\Delta x$.").
• Overcomplicating: Use the simplest form of the solution unless asked for a series expansion.

Sample Exam Question Breakdown

Question: "A long cable carries a uniform heat source $Q$ W/m³. The surface is maintained at 20°C. Derive the steady-state temperature distribution if the cable radius is 0.2m and thermal diffusivity is $10^{-5}$ m²/s. Discuss how NTC might use this model to prevent overheating."

Expected Answer Structure:

1. PDE Setup: $$\frac{1}{r}\frac{d}{dr}\left(r\frac{du}{dr}\right)+\frac{Q}{\alpha }=0$$ "This is the steady-state heat equation in cylindrical coordinates for uniform heat generation."

2. Boundary Conditions:

   • $u(0.2)=20$°C (surface temperature).
   • $\frac{du}{dr}(0)=0$ (symmetry at center).

3. Solution: $$u(r)=20+\frac{Q}{4\alpha }(0.2^2-r^2)$$ "The parabolic profile shows the center is hottest ($r=0$)."

4. NTC Application: "NTC could use this to set maximum safe current limits. For example, if $Q$ exceeds $4\alpha \times (20-u_{\text{max}})/0.04$, the cable risks melting."

Visual Summary for Quick Revision

mindmap
  root((Applications in Engineering))
    PDEs
      Heat Equation: NTC Cables
      Wave Equation: Traffic Lights
      Laplace: Antenna Design
    Transforms
      Laplace: PID Controllers (Ncell)
      Fourier: Signal Processing (WhatsApp)
      Z-Transform: Digital Filters (Pathao)
    Residues
      Fault Currents: NTC Grids
      Antenna Patterns: Ncell 5G
    Numerical Methods
      Finite Difference: Heat Simulation
      Finite Element: Robot Paths (Daraz)