Skip to content

Conversation

@ChrisRackauckas-Claude
Copy link

Summary

  • Implements Smooth Orthogonal Decomposition (SOD) algorithm as proposed by Chelidze and Zhou (2006)
  • SOD extracts linear normal modes and natural frequencies from time-series data
  • Uses generalized eigenvalue problem with displacement and velocity covariance matrices

Key Features

  • SOD type following the same pattern as existing POD
  • Automatic velocity computation from displacement data using finite differences
  • Support for user-provided velocity data
  • Energy-based truncation for automatic mode selection
  • Natural frequency estimation from smooth orthogonal values (SOVs)
  • Support for both matrix and vector-of-vectors input formats

Test plan

  • Unit tests for basic SOD functionality with Lorenz system
  • Tests for energy-based truncation
  • Tests with user-provided velocities
  • Tests with harmonic oscillator for frequency recovery
  • Tests for show method
  • All tests pass locally

References

Closes #9

cc @ChrisRackauckas

🤖 Generated with Claude Code

Implements the Smooth Orthogonal Decomposition (SOD) algorithm for extracting
linear normal modes and natural frequencies from time-series data, as proposed
by Chelidze and Zhou (2006).

SOD extends POD by considering temporal smoothness in addition to spatial variance.
It uses the covariance matrices of displacement and velocity responses to form a
generalized eigenvalue problem (Σ_xx Ψ = λ Σ_vv Ψ), which yields smooth orthogonal
modes and values without requiring knowledge of the system's mass matrix.

Key features:
- SOD type following the same pattern as POD
- Automatic velocity computation from displacement data using finite differences
- Support for user-provided velocity data
- Energy-based truncation for automatic mode selection
- Natural frequency estimation from smooth orthogonal values (SOVs)
- Support for both matrix and vector-of-vectors input formats

References:
- Chelidze D, Zhou W. Smooth orthogonal decomposition-based vibration mode
  identification. J Sound Vib. 2006; 292(3-5): 461-473.

Closes SciML#9

Co-Authored-By: Claude Opus 4.5 <[email protected]>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

Smooth Orthogonal Decomposition

3 participants