Model Order Reduction for Large-scale Circuits Using Higher Order Dynamic Mode Decomposition

TL;DR

This paper introduces a high order dynamic mode decomposition (HODMD) technique for large-scale circuit model order reduction, improving efficiency and accuracy over traditional DMD methods without requiring explicit system equations.

Contribution

It proposes a novel HODMD algorithm with delayed embedding for efficient large-scale circuit simulation, overcoming limitations of standard DMD in signal reconstruction.

Findings

HODMD outperforms DMD in computational efficiency.

HODMD accurately reconstructs signals in large-scale circuits.

Validated on three numerical test cases.

Abstract

Model order reduction (MOR) has long been a mainstream strategy to accelerate large-scale transient circuit simulation. Dynamic Mode Decomposition (DMD) represents a novel data-driven characterization method, extracting dominant dynamical modes directly from time-domain simulation data without requiring explicit system equations. This paper first deduces the DMD algorithm and then proposes high order dynamic mode decomposition (HODMD) incorporating delayed embedding technique, specifically targeting computational efficiency in large-scale circuit simulations. Compared with the DMD method, the HODMD method overcomes the problem that the output signal cannot be reconstructed when the spatial resolution is insufficient. The proposed HODMD algorithm is applicable to general circuits and does not impose any constraints on the topology of the pertinent circuit or type of the components. Three representative numerical test cases are presented to systematically validate both the computational efficiency and accuracy of the proposed HODMD method.