Fast and self-learning indoor airflow simulation based on in situ adaptive tabulation

Wei Tian, Thomas Alonso Sevilla, Dan Li, Wangda Zuo, Michael Wetter

Research output: Contribution to journalArticle

3 Citations (Scopus)

Abstract

Fast simulation for stratified indoor airflow distributions is desired for various applications, such as design of advanced indoor environments, emergency management, and coupled annual energy simulation for buildings with stratified air distributions. Reduced order models trained by pre-computed computational fluid dynamics results are fast, but their prediction may be inaccurate when applied for conditions outside the training domain. To overcome this limitation, we propose a fast and self-learning model based on an in situ adaptive tabulation (ISAT) algorithm, which is trained by a fast fluid dynamics (FFD) model as an example. The idea is that the ISAT will retrieve the solutions from an existing data set if the estimated prediction error is within a pre-defined tolerance. Otherwise, the ISAT will execute the FFD simulation, which is accelerated by running in parallel on a graphics processing unit, for a full-scale simulation. This paper systematically investigates the feasibility of the ISAT for indoor airflow simulations by presenting the ISAT-FFD implementation alongside results related to its overall performance. Using a stratified indoor airflow as an example, we evaluated how the training time of ISAT was impacted by four factors (training methods, error tolerances, number of inputs, and number of outputs). Then we demonstrated that a trained ISAT model can predict the key information for inputs both inside and outside the training domain. The ISAT was able to answer query points both inside and close to training domain using retrieve actions within a time less than 0.001 s for each query. Finally, we provided suggestions for using the ISAT for building applications.

Original languageEnglish (US)
Pages (from-to)1-14
Number of pages14
JournalJournal of Building Performance Simulation
DOIs
StateAccepted/In press - Feb 17 2017

Fingerprint

Self-learning
Fluid dynamics
Simulation
Fluid Dynamics
Adaptive algorithms
Dynamic models
Computational fluid dynamics
Tolerance
Computer simulation
Air
Query
Emergency Management
Reduced Order Model
Graphics Processing Unit
Fluid Model
Prediction Error
Inaccurate
Adaptive Algorithm
Computational Fluid Dynamics
Dynamic Simulation

Keywords

  • fast fluid dynamics
  • in situ adaptive tabulation
  • indoor airflow simulation
  • reduced order model
  • self-learning

ASJC Scopus subject areas

  • Architecture
  • Modeling and Simulation
  • Building and Construction
  • Computer Science Applications

Cite this

Fast and self-learning indoor airflow simulation based on in situ adaptive tabulation. / Tian, Wei; Sevilla, Thomas Alonso; Li, Dan; Zuo, Wangda; Wetter, Michael.

In: Journal of Building Performance Simulation, 17.02.2017, p. 1-14.

Research output: Contribution to journalArticle

Tian, Wei ; Sevilla, Thomas Alonso ; Li, Dan ; Zuo, Wangda ; Wetter, Michael. / Fast and self-learning indoor airflow simulation based on in situ adaptive tabulation. In: Journal of Building Performance Simulation. 2017 ; pp. 1-14.
@article{fe17e1ba13b040e18db53399c6d94f3b,
title = "Fast and self-learning indoor airflow simulation based on in situ adaptive tabulation",
abstract = "Fast simulation for stratified indoor airflow distributions is desired for various applications, such as design of advanced indoor environments, emergency management, and coupled annual energy simulation for buildings with stratified air distributions. Reduced order models trained by pre-computed computational fluid dynamics results are fast, but their prediction may be inaccurate when applied for conditions outside the training domain. To overcome this limitation, we propose a fast and self-learning model based on an in situ adaptive tabulation (ISAT) algorithm, which is trained by a fast fluid dynamics (FFD) model as an example. The idea is that the ISAT will retrieve the solutions from an existing data set if the estimated prediction error is within a pre-defined tolerance. Otherwise, the ISAT will execute the FFD simulation, which is accelerated by running in parallel on a graphics processing unit, for a full-scale simulation. This paper systematically investigates the feasibility of the ISAT for indoor airflow simulations by presenting the ISAT-FFD implementation alongside results related to its overall performance. Using a stratified indoor airflow as an example, we evaluated how the training time of ISAT was impacted by four factors (training methods, error tolerances, number of inputs, and number of outputs). Then we demonstrated that a trained ISAT model can predict the key information for inputs both inside and outside the training domain. The ISAT was able to answer query points both inside and close to training domain using retrieve actions within a time less than 0.001 s for each query. Finally, we provided suggestions for using the ISAT for building applications.",
keywords = "fast fluid dynamics, in situ adaptive tabulation, indoor airflow simulation, reduced order model, self-learning",
author = "Wei Tian and Sevilla, {Thomas Alonso} and Dan Li and Wangda Zuo and Michael Wetter",
year = "2017",
month = "2",
day = "17",
doi = "10.1080/19401493.2017.1288761",
language = "English (US)",
pages = "1--14",
journal = "Journal of Building Performance Simulation",
issn = "1940-1493",
publisher = "Taylor and Francis Ltd.",

}

TY - JOUR

T1 - Fast and self-learning indoor airflow simulation based on in situ adaptive tabulation

AU - Tian, Wei

AU - Sevilla, Thomas Alonso

AU - Li, Dan

AU - Zuo, Wangda

AU - Wetter, Michael

PY - 2017/2/17

Y1 - 2017/2/17

N2 - Fast simulation for stratified indoor airflow distributions is desired for various applications, such as design of advanced indoor environments, emergency management, and coupled annual energy simulation for buildings with stratified air distributions. Reduced order models trained by pre-computed computational fluid dynamics results are fast, but their prediction may be inaccurate when applied for conditions outside the training domain. To overcome this limitation, we propose a fast and self-learning model based on an in situ adaptive tabulation (ISAT) algorithm, which is trained by a fast fluid dynamics (FFD) model as an example. The idea is that the ISAT will retrieve the solutions from an existing data set if the estimated prediction error is within a pre-defined tolerance. Otherwise, the ISAT will execute the FFD simulation, which is accelerated by running in parallel on a graphics processing unit, for a full-scale simulation. This paper systematically investigates the feasibility of the ISAT for indoor airflow simulations by presenting the ISAT-FFD implementation alongside results related to its overall performance. Using a stratified indoor airflow as an example, we evaluated how the training time of ISAT was impacted by four factors (training methods, error tolerances, number of inputs, and number of outputs). Then we demonstrated that a trained ISAT model can predict the key information for inputs both inside and outside the training domain. The ISAT was able to answer query points both inside and close to training domain using retrieve actions within a time less than 0.001 s for each query. Finally, we provided suggestions for using the ISAT for building applications.

AB - Fast simulation for stratified indoor airflow distributions is desired for various applications, such as design of advanced indoor environments, emergency management, and coupled annual energy simulation for buildings with stratified air distributions. Reduced order models trained by pre-computed computational fluid dynamics results are fast, but their prediction may be inaccurate when applied for conditions outside the training domain. To overcome this limitation, we propose a fast and self-learning model based on an in situ adaptive tabulation (ISAT) algorithm, which is trained by a fast fluid dynamics (FFD) model as an example. The idea is that the ISAT will retrieve the solutions from an existing data set if the estimated prediction error is within a pre-defined tolerance. Otherwise, the ISAT will execute the FFD simulation, which is accelerated by running in parallel on a graphics processing unit, for a full-scale simulation. This paper systematically investigates the feasibility of the ISAT for indoor airflow simulations by presenting the ISAT-FFD implementation alongside results related to its overall performance. Using a stratified indoor airflow as an example, we evaluated how the training time of ISAT was impacted by four factors (training methods, error tolerances, number of inputs, and number of outputs). Then we demonstrated that a trained ISAT model can predict the key information for inputs both inside and outside the training domain. The ISAT was able to answer query points both inside and close to training domain using retrieve actions within a time less than 0.001 s for each query. Finally, we provided suggestions for using the ISAT for building applications.

KW - fast fluid dynamics

KW - in situ adaptive tabulation

KW - indoor airflow simulation

KW - reduced order model

KW - self-learning

UR - http://www.scopus.com/inward/record.url?scp=85013124231&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85013124231&partnerID=8YFLogxK

U2 - 10.1080/19401493.2017.1288761

DO - 10.1080/19401493.2017.1288761

M3 - Article

AN - SCOPUS:85013124231

SP - 1

EP - 14

JO - Journal of Building Performance Simulation

JF - Journal of Building Performance Simulation

SN - 1940-1493

ER -