Partial Differential Equations: Types, Solutions & Engineering Applications

TAKEAWAYS:

• PDEs model dynamic systems: They describe how quantities like temperature, vibration, or fluid flow change over space and time (e.g., heat in a rod, sound waves in air).
• Classification matters: PDEs are categorized as elliptic (steady-state, e.g., Laplace’s equation), parabolic (diffusion, e.g., heat equation), or hyperbolic (wave propagation, e.g., wave equation) based on their discriminant $B^2-4AC$.
• Separation of variables is a powerful method: Assume $u(x,t)=X(x)T(t)$ to reduce PDEs into ODEs, solvable via boundary conditions.
• Boundary/initial conditions are critical: Real-world problems (e.g., a cooling metal bar) require specifying $u(x,0)$, $u(0,t)$, and $u(L,t)$ to uniquely determine solutions.
• Fourier series and transforms often appear: Solutions to PDEs frequently involve trigonometric series (e.g., sine/cosine expansions for heat flow).
• Engineering applications dominate exams: Expect questions on heat conduction, wave equations (e.g., vibrating strings), and Laplace’s equation (e.g., electrostatics).

1. Introduction to Partial Differential Equations (PDEs)

Partial differential equations relate a function of multiple variables to its partial derivatives. Unlike ordinary differential equations (ODEs), PDEs involve two or more independent variables (e.g., $x$, $t$) and describe dynamic systems where change depends on both space and time.

Why PDEs?

• Model real-world phenomena: Heat flow, sound waves, fluid dynamics, and electromagnetic fields are governed by PDEs.
• General form: $$F\left(x,y,u,\frac{\partial u}{\partial x},\frac{\partial u}{\partial y},\frac{{\partial }^{2}u}{\partial x^2},\frac{{\partial }^{2}u}{\partial y^2},\dots \right)=0$$ where $u=u(x,y)$ is the dependent variable.

Example: Heat Equation

The 1D heat equation models temperature $u(x,t)$ in a rod: $$\frac{\partial u}{\partial t}={\alpha }^2\frac{{\partial }^{2}u}{\partial x^2}$$

• ${\alpha }^2$: Thermal diffusivity (material property).
• Physical interpretation: The rate of temperature change at a point $x$ depends on the second spatial derivative (curvature of the temperature profile).

2. Classification of Second-Order PDEs

All second-order linear PDEs of the form: $$A\frac{{\partial }^{2}u}{\partial x^2}+2B\frac{{\partial }^{2}u}{\partial x\partial y}+C\frac{{\partial }^{2}u}{\partial y^2}+D\frac{\partial u}{\partial x}+E\frac{\partial u}{\partial y}+Fu=G$$ are classified based on the discriminant $D=B^2-4AC$:

3. Key PDEs in Engineering

A. Heat Equation (Parabolic)

Models temperature distribution in a medium: $$\frac{\partial u}{\partial t}={\alpha }^2\frac{{\partial }^{2}u}{\partial x^2}$$ Example: Cooling of a metal rod.

• Boundary conditions:
  • $u(0,t)=100°C$ (left end kept hot),
  • $u(L,t)=0°C$ (right end insulated).
• Initial condition: $u(x,0)=f(x)$.

Solution via Separation of Variables: Assume $u(x,t)=X(x)T(t)$. Substituting into the PDE: $$X(x)T^{\prime }(t)={\alpha }^2{X}^{\prime \prime }(x)T(t)⟹\frac{T^{\prime }(t)}{T(t)}={\alpha }^2\frac{X^{\prime \prime }(x)}{X(x)}=-\lambda$$ This yields two ODEs:

1. $X^{\prime \prime }(x)+\lambda X(x)=0$,
2. $T^{\prime }(t)+{\alpha }^2\lambda T(t)=0$.

Boundary conditions lead to eigenvalues ${\lambda }_n={\left(\frac{n\pi }{L}\right)}^2$ and eigenfunctions $X_n(x)=\sin \left(\frac{n\pi x}{L}\right)$.

General solution: $$u(x,t)=\sum _{n=1}^{\infty }{B}_n\sin \left(\frac{n\pi x}{L}\right)e^{-{\alpha }^2{\lambda }_{n}t}$$ where $B_n$ are Fourier coefficients determined by $f(x)$.

B. Wave Equation (Hyperbolic)

Models vibrations in strings, sound waves, or electromagnetic waves: $$\frac{{\partial }^{2}u}{\partial t^2}=c^2\frac{{\partial }^{2}u}{\partial x^2}$$ Example: Vibrating guitar string.

• Boundary conditions:
  • $u(0,t)=0$ (fixed end),
  • $u(L,t)=0$ (fixed end).
• Initial conditions:
  • $u(x,0)=f(x)$ (initial displacement),
  • $u_t(x,0)=g(x)$ (initial velocity).

D’Alembert’s Solution: $$u(x,t)=\frac{1}{2}[f(x+ct)+f(x-ct)]+\frac{1}{2c}{\int }_{x-ct}^{x+ct}g(s)ds$$ This represents two waves traveling left and right with speed $c$.

C. Laplace’s Equation (Elliptic)

Models steady-state phenomena (e.g., electrostatics, fluid flow): $${\nabla }^{2}u=\frac{{\partial }^{2}u}{\partial x^2}+\frac{{\partial }^{2}u}{\partial y^2}=0$$ Example: Electric potential in a 2D region.

• Boundary conditions: Dirichlet (fixed potential) or Neumann (fixed flux).

Solution in Polar Coordinates: For a circular membrane with radius $a$ and boundary $u(a,\theta )=f(\theta )$: $$u(r,\theta )=\frac{a^2-r^2}{2a}f(\theta )+\frac{r^2}{2a}\cdot \frac{1}{2\pi }{\int }_0^{2\pi }\frac{f(\varphi )}{\cos (\theta -\varphi )}d\varphi$$ (Poisson integral formula.)

4. Separation of Variables: Step-by-Step

Problem: Solve the heat equation $u_t={\alpha }^2{u}_{xx}$ on $0<x<L$ with:

• $u(0,t)=0$, $u(L,t)=0$ (Dirichlet BCs),
• $u(x,0)=\sin \left(\frac{\pi x}{L}\right)$.

Steps:

1. Assume separation: $u(x,t)=X(x)T(t)$.
2. Substitute into PDE: $$X(x)T^{\prime }(t)={\alpha }^2{X}^{\prime \prime }(x)T(t)⟹\frac{T^{\prime }(t)}{T(t)}={\alpha }^2\frac{X^{\prime \prime }(x)}{X(x)}=-\lambda$$
3. Solve ODEs:
   • $X^{\prime \prime }(x)+\lambda X(x)=0$ with $X(0)=X(L)=0$ → ${\lambda }_n={\left(\frac{n\pi }{L}\right)}^2$, $X_n(x)=\sin \left(\frac{n\pi x}{L}\right)$.
   • $T^{\prime }(t)+{\alpha }^2{\lambda }_{n}T(t)=0$ → $T_n(t)=e^{-{\alpha }^2{\lambda }_{n}t}$.
4. General solution: $$u(x,t)=\sum _{n=1}^{\infty }{B}_n\sin \left(\frac{n\pi x}{L}\right)e^{-{\alpha }^2{\lambda }_{n}t}$$
5. Apply initial condition: $$u(x,0)=\sin \left(\frac{\pi x}{L}\right)=B_1\sin \left(\frac{\pi x}{L}\right)⟹B_1=1,B_n=0\text{ for }n\ne 1.$$
6. Final solution: $$u(x,t)=\sin \left(\frac{\pi x}{L}\right)e^{-{\alpha }^2{\left(\frac{\pi }{L}\right)}^{2}t}$$

flowchart TD
    A["1. Assume \(u(x,t) = X(x)T(t)\)"] --> B["2. Substitute into PDE\n\(\frac{T'}{T} = \alpha^2 \frac{X''}{X} = -\lambda\)"]
    B --> C["3. Solve \(X'' + \lambda X = 0\)\nwith BCs \(X(0) = X(L) = 0\)\nEigenvalues: \(\lambda_n = \left(\frac{n\pi}{L}\right)^2\)"]
    C --> D["4. Solve \(T' + \alpha^2 \lambda_n T = 0\)\n\(T_n(t) = e^{-\alpha^2 \lambda_n t}\)"]
    D --> E["5. General solution:\n\(u(x,t) = \sum B_n \sin\left(\frac{n\pi x}{L}\right) e^{-\alpha^2 \lambda_n t}\)"]
    E --> F["6. Apply initial condition\n\(u(x,0) = f(x)\) to find \(B_n\)"]
    F --> G["7. Final solution:\n\(u(x,t) = \sin\left(\frac{\pi x}{L}\right) e^{-\alpha^2 \left(\frac{\pi}{L}\right)^2 t}\)"]

5. Boundary and Initial Conditions

PDEs require auxiliary conditions to ensure unique solutions:

• Initial conditions (ICs): Specify the state at $t=0$ (for time-dependent PDEs).
  • Example: $u(x,0)=f(x)$ (heat equation), $u_t(x,0)=g(x)$ (wave equation).
• Boundary conditions (BCs): Specify behavior at spatial boundaries.
  • Dirichlet: $u(a,t)=g(t)$ (fixed value).
  • Neumann: $u_x(a,t)=h(t)$ (fixed flux/derivative).
  • Mixed: Combination of Dirichlet and Neumann.

Example: Heat flow in a rod with:

• $u(0,t)=100°C$ (Dirichlet),
• $u_x(1,t)=0$ (Neumann, insulated end),
• $u(x,0)=50°C$ (IC).

6. Applications in Engineering

A. Heat Conduction (Parabolic PDE)

• Example: Designing a cooler for electronic components (e.g., CPUs in servers).
  • The heat equation models how heat dissipates through a heat sink.
  • Boundary conditions: Fixed temperature at the base (CPU) and ambient air at the fins.
  • Solution: Fourier series to find temperature distribution and optimize fin design.

B. Wave Propagation (Hyperbolic PDE)

• Example: Vibration analysis in bridges (e.g., the Suspension Bridge in Pokhara).
  • The wave equation models how seismic waves or traffic vibrations propagate.
  • Boundary conditions: Fixed supports at pillars.
  • Solution: D’Alembert’s solution to predict resonance frequencies and avoid structural failure.

C. Electrostatics (Elliptic PDE)

• Example: Design of antennas (e.g., Ncell base stations).
  • Laplace’s equation governs the electric potential around an antenna.
  • Boundary conditions: Fixed potential on the antenna surface.
  • Solution: Separation of variables in polar coordinates to optimize radiation patterns.

D. Fluid Dynamics (Navier-Stokes Equations)

• Example: Traffic flow modeling (e.g., Kathmandu’s busy rings).
  • The Burgers’ equation (a simplified PDE) models car density $\rho (x,t)$: $$\frac{\partial \rho }{\partial t}+\rho \frac{\partial \rho }{\partial x}=\nu \frac{{\partial }^2\rho }{\partial x^2}$$
  • Solution: Predicts jam formation and optimal traffic light timing.

7. Fourier Series and PDE Solutions

Many PDE solutions involve Fourier series to satisfy boundary conditions. For example:

• Heat equation: Eigenfunctions $\sin \left(\frac{n\pi x}{L}\right)$ form a basis.
• Wave equation: Eigenfunctions $\sin \left(\frac{n\pi x}{L}\right)$ or $\cos \left(\frac{n\pi x}{L}\right)$.

Fourier Sine Series Example: Given $f(x)=x$ on $0<x<\pi$, find the sine series: $$f(x)=\sum _{n=1}^{\infty }{B}_n\sin (nx),B_n=\frac{2}{\pi }{\int }_0^{\pi }x\sin (nx)dx=\frac{2(-1{)}^{n+1}}{n}.$$ Thus: $$x=2\left(\sin x-\frac{\sin (2x)}{2}+\frac{\sin (3x)}{3}-\dots \right),0<x<\pi .$$

8. Numerical Methods (Brief Overview)

Analytical solutions are not always possible. Numerical methods approximate PDE solutions:

1. Finite Difference Method (FDM):
   • Discretize space/time into grids.
   • Approximate derivatives (e.g., $u_{xx}\approx \frac{u_{i+1}-2u_i+u_{i-1}}{h^2}$).
2. Finite Element Method (FEM):
   • Divide domain into elements (e.g., triangles).
   • Solve weak form of PDE.
3. Finite Volume Method (FVM):
   • Conserves quantities (e.g., mass, energy) over control volumes.

Example: Heat equation discretization: $$\frac{u_i^{n+1}-u_i^n}{\Delta t}={\alpha }^2\frac{u_{i+1}^n-2u_i^n+u_{i-1}^n}{(\Delta x{)}^2}$$ Rearranged for explicit scheme: $$u_i^{n+1}=u_i^n+\frac{{\alpha }^2\Delta t}{(\Delta x{)}^2}(u_{i+1}^n-2u_i^n+u_{i-1}^n)$$

flowchart TD
    A["1. Discretize domain\nSpace: \(x_i = i \Delta x\), Time: \(t^n = n \Delta t\)"] --> B["2. Approximate derivatives\n\(u_{xx} \approx \frac{u_{i+1} - 2u_i + u_{i-1}}{h^2}\)"]
    B --> C["3. Substitute into PDE\nHeat equation: \(u_t = \alpha^2 u_{xx}\)"]
    C --> D["4. Solve iteratively\nExplicit: \(u_i^{n+1} = u_i^n + r(u_{i+1}^n - 2u_i^n + u_{i-1}^n)\)\nwhere \(r = \frac{\alpha^2 \Delta t}{h^2}\)"]
    D --> E["5. Apply boundary/initial conditions\nUpdate grid values step-by-step"]

In the Real World

1. eSewa and Khalti (Digital Payments):

   • PDE Idea: Diffusion models (parabolic PDEs) simulate how payment transactions propagate through the network.
   • How: The heat equation approximates how a "payment wave" spreads from users to merchants, accounting for delays and network congestion.
   • Example: During Dashain, when millions transact simultaneously, PDEs help predict server load and optimize response times.

2. Pathao (Ride-Hailing):

   • PDE Idea: Traffic flow equations (Burgers’ equation or Navier-Stokes) model driver density and ride demand.
   • How: Pathao uses PDEs to:
     • Predict surge pricing zones (like a "traffic jam PDE").
     • Optimize driver dispatch to balance supply/demand dynamically.
   • Real Example: During monsoon, when roads flood, PDEs simulate reduced vehicle speeds and adjust ETA calculations.

3. NTC (Telecom Infrastructure):

   • PDE Idea: Wave equation for signal propagation in fiber optics.
   • How: The hyperbolic PDE describes how light pulses travel through optical fibers, accounting for dispersion and attenuation.
   • Example: NTC uses PDE-based models to design fiber routes that minimize signal loss, ensuring stable internet in remote areas like Dolpa.

4. NEPSE (Stock Market):

   • PDE Idea: Black-Scholes PDE (elliptic) for option pricing.
   • How: The equation: $$\frac{\partial V}{\partial t}+\frac{1}{2}{\sigma }^2{S}^2\frac{{\partial }^{2}V}{\partial S^2}+rS\frac{\partial V}{\partial S}-rV=0$$ models how stock option prices change with time and volatility.
   • Example: NEPSE traders use PDE solvers to price derivatives on shares like NABIL or NMB.

5. Ncell Tower Design:

   • PDE Idea: Laplace’s equation for antenna radiation patterns.
   • How: Engineers solve ${\nabla }^{2}u=0$ in 3D to design antenna shapes that maximize coverage in hilly terrains (e.g., Kathmandu Valley).
   • Real Example: Ncell’s 5G rollout uses PDE simulations to place towers where signal strength is optimal, reducing dead zones.

Exam Tip

What Examiners Look For

1. Classification First:

   • Always start by classifying the PDE (elliptic/parabolic/hyperbolic) using $B^2-4AC$.
   • Example: For $u_{xx}+3u_{xy}+2u_{yy}=0$, $D=9-16=-7<0$ → elliptic.

2. Boundary Conditions Are Half the Marks:

   • Common mistakes:
     • Forgetting to apply BCs (e.g., $u(0,t)=0$).
     • Misapplying Neumann conditions (e.g., $u_x=0$ vs. $u=0$).
   • Tip: Write BCs explicitly before solving.

3. Separation of Variables Steps:

   • Show every step of assuming $u(x,t)=X(x)T(t)$.
   • Clearly label eigenvalues ${\lambda }_n$ and eigenfunctions $X_n(x)$.

4. Fourier Series Coefficients:

   • For non-zero initial conditions, compute $B_n$ using: $$B_n=\frac{2}{L}{\int }_0^{L}f(x)\sin \left(\frac{n\pi x}{L}\right)dx$$
   • Shortcut: Memorize integrals like $\int x\sin (nx)dx=-\frac{x\cos (nx)}{n}+\frac{\sin (nx)}{n^2}$.

5. Physical Interpretation:

   • Relate solutions to real-world scenarios. For example:
     • Heat equation: "Temperature decays exponentially over time."
     • Wave equation: "The string vibrates with speed $c$."

6. Numerical Methods (if asked):

   • For FDM, show the grid setup and stencil (e.g., 3-point stencil for $u_{xx}$).
   • State the stability condition (e.g., for heat equation: $r=\frac{{\alpha }^2\Delta t}{h^2}\le \frac{1}{2}$).

Common Pitfalls

• Ignoring initial conditions: Always check if $u(x,0)$ or $u_t(x,0)$ is given.
• Incorrect eigenvalues: For $X^{\prime \prime }+\lambda X=0$ with $X(0)=X(L)=0$, ${\lambda }_n={\left(\frac{n\pi }{L}\right)}^2$ (not $\frac{n^2{\pi }^2}{L}$).
• Mixing BC types: Dirichlet for fixed value, Neumann for fixed derivative.
• Forgetting units: In real-world problems, ensure ${\alpha }^2$ has units of $m^2/s$.

Model Answer Structure

For a question like: "Solve the wave equation $u_{tt}=c^2{u}_{xx}$ on $0<x<L$ with $u(0,t)=u(L,t)=0$, $u(x,0)=\sin (\pi x/L)$, $u_t(x,0)=0$."

Step-by-Step Marks Distribution:

1. Classify PDE (hyperbolic, $D=0-4(1)(0)=0$).
2. Assume separation: $u(x,t)=X(x)T(t)$.
3. Derive ODEs:
   • $X^{\prime \prime }(x)+\lambda X(x)=0$ → ${\lambda }_n={\left(\frac{n\pi }{L}\right)}^2$, $X_n(x)=\sin \left(\frac{n\pi x}{L}\right)$.
   • $T^{\prime \prime }(t)+c^2{\lambda }_{n}T(t)=0$ → $T_n(t)=A_n\cos (c{\lambda }_{n}t)+B_n\sin (c{\lambda }_{n}t)$.
4. Apply ICs:
   • $u(x,0)=\sum A_n\sin \left(\frac{n\pi x}{L}\right)=\sin \left(\frac{\pi x}{L}\right)$ → $A_1=1$, $A_n=0$ for $n\ne 1$.
   • $u_t(x,0)=0$ → $B_n=0$ for all $n$.
5. Final solution: $$u(x,t)=\cos \left(\frac{c\pi t}{L}\right)\sin \left(\frac{\pi x}{L}\right)$$

flowchart LR
    A["1. Classify PDE\nHyperbolic: \(D = 0\)"] --> B["2. Separation of variables\n\(u(x,t) = X(x)T(t)\)"]
    B --> C["3. Solve \(X'' + \lambda X = 0\)\nBCs: \(X(0) = X(L) = 0\)\n\(\lambda_n = \left(\frac{n\pi}{L}\right)^2\)"]
    C --> D["4. Solve \(T'' + c^2 \lambda_n T = 0\)\n\(T_n(t) = A_n \cos(c \lambda_n t) + B_n \sin(c \lambda_n t)\)"]
    D --> E["5. Apply ICs\n\(u(x,0) = \sin(\pi x/L)\) → \(A_1 = 1\), \(B_n = 0\)"]
    E --> F["6. Final solution\n\(u(x,t) = \cos\left(\frac{c\pi t}{L}\right) \sin\left(\frac{\pi x}{L}\right)\)"]