Dynamic data-driven model reduction: adapting reduced models from incomplete data

Benjamin Peherstorfer, Karen Willcox

Research output: Contribution to journalArticle

Abstract

This work presents a data-driven online adaptive model reduction approach for systems that undergo dynamic changes. Classical model reduction constructs a reduced model of a large-scale system in an offline phase and then keeps the reduced model unchanged during the evaluations in an online phase; however, if the system changes online, the reduced model may fail to predict the behavior of the changed system. Rebuilding the reduced model from scratch is often too expensive in time-critical and real-time environments. We introduce a dynamic data-driven adaptation approach that adapts the reduced model from incomplete sensor data obtained from the system during the online computations. The updates to the reduced models are derived directly from the incomplete data, without recourse to the full model. Our adaptivity approach approximates the missing values in the incomplete sensor data with gappy proper orthogonal decomposition. These approximate data are then used to derive low-rank updates to the reduced basis and the reduced operators. In our numerical examples, incomplete data with 30–40 % known values are sufficient to recover the reduced model that would be obtained via rebuilding from scratch.

Original languageEnglish (US)
Article number11
JournalAdvanced Modeling and Simulation in Engineering Sciences
Volume3
Issue number1
DOIs
StatePublished - Dec 1 2016

Fingerprint

Reduced Model
Incomplete Data
Model Reduction
Data-driven
Update
Sensor
Orthogonal Decomposition
Missing Values
Adaptivity
Large-scale Systems
Sensors
Sufficient
Real-time
Predict
Numerical Examples
Large scale systems
Evaluation
Operator
Decomposition

Keywords

  • Dynamic data-driven application systems
  • Dynamic data-driven reduced models
  • Gappy proper orthogonal decomposition
  • Incomplete sensor data
  • Model reduction
  • Online adaptivity

ASJC Scopus subject areas

  • Modeling and Simulation
  • Applied Mathematics
  • Computer Science Applications
  • Engineering (miscellaneous)

Cite this

Dynamic data-driven model reduction : adapting reduced models from incomplete data. / Peherstorfer, Benjamin; Willcox, Karen.

In: Advanced Modeling and Simulation in Engineering Sciences, Vol. 3, No. 1, 11, 01.12.2016.

Research output: Contribution to journalArticle

@article{a4a3a2cffa39496a87aa08d5f9d3f053,
title = "Dynamic data-driven model reduction: adapting reduced models from incomplete data",
abstract = "This work presents a data-driven online adaptive model reduction approach for systems that undergo dynamic changes. Classical model reduction constructs a reduced model of a large-scale system in an offline phase and then keeps the reduced model unchanged during the evaluations in an online phase; however, if the system changes online, the reduced model may fail to predict the behavior of the changed system. Rebuilding the reduced model from scratch is often too expensive in time-critical and real-time environments. We introduce a dynamic data-driven adaptation approach that adapts the reduced model from incomplete sensor data obtained from the system during the online computations. The updates to the reduced models are derived directly from the incomplete data, without recourse to the full model. Our adaptivity approach approximates the missing values in the incomplete sensor data with gappy proper orthogonal decomposition. These approximate data are then used to derive low-rank updates to the reduced basis and the reduced operators. In our numerical examples, incomplete data with 30–40 {\%} known values are sufficient to recover the reduced model that would be obtained via rebuilding from scratch.",
keywords = "Dynamic data-driven application systems, Dynamic data-driven reduced models, Gappy proper orthogonal decomposition, Incomplete sensor data, Model reduction, Online adaptivity",
author = "Benjamin Peherstorfer and Karen Willcox",
year = "2016",
month = "12",
day = "1",
doi = "10.1186/s40323-016-0064-x",
language = "English (US)",
volume = "3",
journal = "Advanced Modeling and Simulation in Engineering Sciences",
issn = "2213-7467",
publisher = "Springer Open",
number = "1",

}

TY - JOUR

T1 - Dynamic data-driven model reduction

T2 - adapting reduced models from incomplete data

AU - Peherstorfer, Benjamin

AU - Willcox, Karen

PY - 2016/12/1

Y1 - 2016/12/1

N2 - This work presents a data-driven online adaptive model reduction approach for systems that undergo dynamic changes. Classical model reduction constructs a reduced model of a large-scale system in an offline phase and then keeps the reduced model unchanged during the evaluations in an online phase; however, if the system changes online, the reduced model may fail to predict the behavior of the changed system. Rebuilding the reduced model from scratch is often too expensive in time-critical and real-time environments. We introduce a dynamic data-driven adaptation approach that adapts the reduced model from incomplete sensor data obtained from the system during the online computations. The updates to the reduced models are derived directly from the incomplete data, without recourse to the full model. Our adaptivity approach approximates the missing values in the incomplete sensor data with gappy proper orthogonal decomposition. These approximate data are then used to derive low-rank updates to the reduced basis and the reduced operators. In our numerical examples, incomplete data with 30–40 % known values are sufficient to recover the reduced model that would be obtained via rebuilding from scratch.

AB - This work presents a data-driven online adaptive model reduction approach for systems that undergo dynamic changes. Classical model reduction constructs a reduced model of a large-scale system in an offline phase and then keeps the reduced model unchanged during the evaluations in an online phase; however, if the system changes online, the reduced model may fail to predict the behavior of the changed system. Rebuilding the reduced model from scratch is often too expensive in time-critical and real-time environments. We introduce a dynamic data-driven adaptation approach that adapts the reduced model from incomplete sensor data obtained from the system during the online computations. The updates to the reduced models are derived directly from the incomplete data, without recourse to the full model. Our adaptivity approach approximates the missing values in the incomplete sensor data with gappy proper orthogonal decomposition. These approximate data are then used to derive low-rank updates to the reduced basis and the reduced operators. In our numerical examples, incomplete data with 30–40 % known values are sufficient to recover the reduced model that would be obtained via rebuilding from scratch.

KW - Dynamic data-driven application systems

KW - Dynamic data-driven reduced models

KW - Gappy proper orthogonal decomposition

KW - Incomplete sensor data

KW - Model reduction

KW - Online adaptivity

UR - http://www.scopus.com/inward/record.url?scp=85044298892&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85044298892&partnerID=8YFLogxK

U2 - 10.1186/s40323-016-0064-x

DO - 10.1186/s40323-016-0064-x

M3 - Article

VL - 3

JO - Advanced Modeling and Simulation in Engineering Sciences

JF - Advanced Modeling and Simulation in Engineering Sciences

SN - 2213-7467

IS - 1

M1 - 11

ER -