Skip to content

Latest commit

 

History

History
56 lines (46 loc) · 2.09 KB

File metadata and controls

56 lines (46 loc) · 2.09 KB

Project Tasks and Milestones

Core question

How can one use a single scalar control input to bring a vibrating square membrane as close to rest as possible?

Milestone 1: PDE to modal ODEs

  • Write down the controlled wave equation on the unit square.
  • State the Dirichlet eigenfunctions and eigenvalues.
  • Derive the modal equations
$$\ddot{q}_{mn} + \omega_{mn}^2 q_{mn} = \beta_{mn} b(t).$$
  • Explain the meaning of the coupling coefficient $\beta_{mn}$.

Milestone 2: Build the finite-dimensional model

  • Choose a truncation level $M$.
  • Flatten the double indices $(m,n)$ into a vector of modal amplitudes.
  • Construct the matrices $A$ and $B$ for the first-order system.
  • Verify dimensions carefully.

Milestone 3: Design the LQR controller

  • Choose $Q$ and $R$.
  • Solve the algebraic Riccati equation.
  • Compute the feedback gain $K$.
  • Explain how the choice of $R$ affects the controller.

Milestone 4: Run simulations

  • Choose at least one nontrivial initial condition.
  • Simulate the open-loop response.
  • Simulate the closed-loop response.
  • Plot modal energy and control input.

Milestone 5: Reconstruct the membrane

  • Reconstruct $u(x,y,t)$ on a spatial grid from the modal coefficients.
  • Plot snapshots of the membrane at several times.
  • Create an animation or sequence of frames.

Milestone 6: Investigate actuator placement

  • Compare at least two actuator locations.
  • Include the center $\left( \frac{1}{2}, \frac{1}{2} \right)$ as one test case.
  • Identify modes with zero or very small coupling.
  • Discuss how this affects stabilization.

Minimum deliverables

  • Short derivation of the modal model
  • Clear definition of $A$, $B$, $Q$, $R$, $K$
  • At least two plots from the simulation
  • At least one membrane reconstruction figure
  • Discussion of actuator placement

Stretch ideas

  • Replace the point actuator with a smooth Gaussian patch.
  • Add light viscous damping and compare results.
  • Track how performance changes as the truncation size grows.
  • Estimate which low modes dominate the transient response.