Identification of MIMO Wiener-type Koopman Models for Data-Driven Model Reduction using Deep Learning

TL;DR

This paper introduces a novel deep learning approach for data-driven nonlinear model reduction of MIMO systems using Wiener-type Koopman models, demonstrating superior accuracy and reduction capabilities in case studies.

Contribution

It develops a Wiener-type Koopman modeling framework combined with deep learning, extending Koopman theory to better suit model reduction of complex MIMO systems.

Findings

Wiener-type Koopman models outperform linear and bilinear models in prediction accuracy.

The proposed models achieve significant reduction in model order.

The approach is validated on systems including a chemical reactor and distillation column.

Abstract

We use Koopman theory to develop a data-driven nonlinear model reduction and identification strategy for multiple-input multiple-output (MIMO) input-affine dynamical systems. While the present literature has focused on linear and bilinear Koopman models, we derive and use a Wiener-type Koopman formulation. We discuss that the Wiener structure is particularly suitable for model reduction, and can be naturally derived from Koopman theory. Moreover, the Wiener block-structure unifies the mathematical simplicity of linear dynamical blocks and the accuracy of bilinear dynamics. We present a Koopman deep-learning strategy combining autoencoders and linear dynamics that generates low-order surrogate models of MIMO Wiener type. In three case studies, we apply our framework for identification and reduction of a system with input multiplicity, a chemical reactor and a high-purity distillation column. We compare the prediction performance of the identified Wiener models to linear and bilinear Koopman models. We observe the highest accuracy and strongest model reduction capabilities of low-order Wiener-type Koopman models, making them promising for control.