"Plug-and-Play" Tiny AI-Empowered Output Filter Parameter Extraction Framework with Single RNN Cell for Digital Power

Kelvin Yi-Wen Hong; Henry Shu-Hung Chung; Alan Wai-Lun Lo; Huai Wang

doi:10.1109/ECCE47101.2021.9595594

"Plug-and-Play" Tiny AI-Empowered Output Filter Parameter Extraction Framework with Single RNN Cell for Digital Power

Kelvin Yi-Wen Hong
, Henry Shu-Hung Chung
, Alan Wai-Lun Lo
, Huai Wang

Research output: Chapters, Conference Papers, Creative and Literary Works › RGC 32 - Refereed conference paper (with host publication) › peer-review

Abstract

A "plug-and-play" tiny artificial intelligence (AI)-empowered output filter parameter extraction framework for digital power is presented. It can be integrated into existing digital power readily without introducing extra sensors. The framework aims to provide digital power with the output filter parameter information to enhance or maintain system performance and to conduct system diagnostics. The concept is based on firstly transferring the control from the default controller to the framework with the predetermined control law for a few switching cycles via typical priority-based interrupt programming, then perturbing and observing the response of the control action, and finally determining the filter parameters with a recurrent neural network (RNN) having a single cell. The proposed framework has been applied and evaluated on a 150W, 180V / 120V buck DC/DC converter. Results show that the root-mean-square (RMS) errors of determining the values of the output inductor and capacitor are 2.28% and 2.86%, respectively.

Original language	English
Title of host publication	2021 IEEE Energy Conversion Congress and Exposition (ECCE 2021) - PROCEEDINGS
Publisher	IEEE
Pages	2755-2761
ISBN (Electronic)	9781728151359
ISBN (Print)	9781728161280
DOIs	https://doi.org/10.1109/ECCE47101.2021.9595594
Publication status	Published - 2021
Event	13th IEEE Energy Conversion Congress and Exposition (ECCE 2021) - Virtual, Canada Duration: 10 Oct 2021 → 14 Oct 2021 https://ieeexplore.ieee.org/xpl/conhome/9594898/proceeding

Publication series

Name	IEEE Energy Conversion Congress and Exposition, ECCE - Proceedings
ISSN (Print)	2329-3721
ISSN (Electronic)	2329-3748

Conference

Conference	13th IEEE Energy Conversion Congress and Exposition (ECCE 2021)
Abbreviated title	ECCE2021
Place	Canada
Period	10/10/21 → 14/10/21
Internet address	https://ieeexplore.ieee.org/xpl/conhome/9594898/proceeding

Research Keywords

Artificial intelligence
control
digital power
neural network
power converters
system identification

Access Output

10.1109/ECCE47101.2021.9595594

Other Access Options

Check CityUHK Library

Permanent Link

Right click to copy link

Cite this

Hong, K. Y.-W., Chung, H. S.-H., Lo, A. W.-L., & Wang, H. (2021). "Plug-and-Play" Tiny AI-Empowered Output Filter Parameter Extraction Framework with Single RNN Cell for Digital Power. In 2021 IEEE Energy Conversion Congress and Exposition (ECCE 2021) - PROCEEDINGS (pp. 2755-2761). (IEEE Energy Conversion Congress and Exposition, ECCE - Proceedings). IEEE. https://doi.org/10.1109/ECCE47101.2021.9595594

@inproceedings{72bad2256bb64c669a0d25370680a760,

title = "{"}Plug-and-Play{"} Tiny AI-Empowered Output Filter Parameter Extraction Framework with Single RNN Cell for Digital Power",

abstract = "A {"}plug-and-play{"} tiny artificial intelligence (AI)-empowered output filter parameter extraction framework for digital power is presented. It can be integrated into existing digital power readily without introducing extra sensors. The framework aims to provide digital power with the output filter parameter information to enhance or maintain system performance and to conduct system diagnostics. The concept is based on firstly transferring the control from the default controller to the framework with the predetermined control law for a few switching cycles via typical priority-based interrupt programming, then perturbing and observing the response of the control action, and finally determining the filter parameters with a recurrent neural network (RNN) having a single cell. The proposed framework has been applied and evaluated on a 150W, 180V / 120V buck DC/DC converter. Results show that the root-mean-square (RMS) errors of determining the values of the output inductor and capacitor are 2.28\% and 2.86\%, respectively.",

keywords = "Artificial intelligence, control, digital power, neural network, power converters, system identification",

author = "Hong, \{Kelvin Yi-Wen\} and Chung, \{Henry Shu-Hung\} and Lo, \{Alan Wai-Lun\} and Huai Wang",

year = "2021",

doi = "10.1109/ECCE47101.2021.9595594",

language = "English",

isbn = "9781728161280",

series = "IEEE Energy Conversion Congress and Exposition, ECCE - Proceedings",

publisher = "IEEE",

pages = "2755--2761",

booktitle = "2021 IEEE Energy Conversion Congress and Exposition (ECCE 2021) - PROCEEDINGS",

address = "United States",

note = "13th IEEE Energy Conversion Congress and Exposition (ECCE 2021), ECCE2021 ; Conference date: 10-10-2021 Through 14-10-2021",

url = "https://ieeexplore.ieee.org/xpl/conhome/9594898/proceeding",

}

Hong, KY-W, Chung, HS-H, Lo, AW-L & Wang, H 2021, "Plug-and-Play" Tiny AI-Empowered Output Filter Parameter Extraction Framework with Single RNN Cell for Digital Power. in 2021 IEEE Energy Conversion Congress and Exposition (ECCE 2021) - PROCEEDINGS. IEEE Energy Conversion Congress and Exposition, ECCE - Proceedings, IEEE, pp. 2755-2761, 13th IEEE Energy Conversion Congress and Exposition (ECCE 2021), Canada, 10/10/21. https://doi.org/10.1109/ECCE47101.2021.9595594

"Plug-and-Play" Tiny AI-Empowered Output Filter Parameter Extraction Framework with Single RNN Cell for Digital Power. / Hong, Kelvin Yi-Wen; Chung, Henry Shu-Hung; Lo, Alan Wai-Lun et al.
2021 IEEE Energy Conversion Congress and Exposition (ECCE 2021) - PROCEEDINGS. IEEE, 2021. p. 2755-2761 (IEEE Energy Conversion Congress and Exposition, ECCE - Proceedings).

Research output: Chapters, Conference Papers, Creative and Literary Works › RGC 32 - Refereed conference paper (with host publication) › peer-review

TY - GEN

T1 - "Plug-and-Play" Tiny AI-Empowered Output Filter Parameter Extraction Framework with Single RNN Cell for Digital Power

AU - Hong, Kelvin Yi-Wen

AU - Chung, Henry Shu-Hung

AU - Lo, Alan Wai-Lun

AU - Wang, Huai

PY - 2021

Y1 - 2021

N2 - A "plug-and-play" tiny artificial intelligence (AI)-empowered output filter parameter extraction framework for digital power is presented. It can be integrated into existing digital power readily without introducing extra sensors. The framework aims to provide digital power with the output filter parameter information to enhance or maintain system performance and to conduct system diagnostics. The concept is based on firstly transferring the control from the default controller to the framework with the predetermined control law for a few switching cycles via typical priority-based interrupt programming, then perturbing and observing the response of the control action, and finally determining the filter parameters with a recurrent neural network (RNN) having a single cell. The proposed framework has been applied and evaluated on a 150W, 180V / 120V buck DC/DC converter. Results show that the root-mean-square (RMS) errors of determining the values of the output inductor and capacitor are 2.28% and 2.86%, respectively.

AB - A "plug-and-play" tiny artificial intelligence (AI)-empowered output filter parameter extraction framework for digital power is presented. It can be integrated into existing digital power readily without introducing extra sensors. The framework aims to provide digital power with the output filter parameter information to enhance or maintain system performance and to conduct system diagnostics. The concept is based on firstly transferring the control from the default controller to the framework with the predetermined control law for a few switching cycles via typical priority-based interrupt programming, then perturbing and observing the response of the control action, and finally determining the filter parameters with a recurrent neural network (RNN) having a single cell. The proposed framework has been applied and evaluated on a 150W, 180V / 120V buck DC/DC converter. Results show that the root-mean-square (RMS) errors of determining the values of the output inductor and capacitor are 2.28% and 2.86%, respectively.

KW - Artificial intelligence

KW - control

KW - digital power

KW - neural network

KW - power converters

KW - system identification

UR - https://www.scopus.com/pages/publications/85123349176

UR - https://www.scopus.com/pages/publications/85123349176#tab=impact

U2 - 10.1109/ECCE47101.2021.9595594

DO - 10.1109/ECCE47101.2021.9595594

M3 - RGC 32 - Refereed conference paper (with host publication)

SN - 9781728161280

T3 - IEEE Energy Conversion Congress and Exposition, ECCE - Proceedings

SP - 2755

EP - 2761

BT - 2021 IEEE Energy Conversion Congress and Exposition (ECCE 2021) - PROCEEDINGS

PB - IEEE

T2 - 13th IEEE Energy Conversion Congress and Exposition (ECCE 2021)

Y2 - 10 October 2021 through 14 October 2021

ER -

Hong KYW, Chung HSH, Lo AWL, Wang H. "Plug-and-Play" Tiny AI-Empowered Output Filter Parameter Extraction Framework with Single RNN Cell for Digital Power. In 2021 IEEE Energy Conversion Congress and Exposition (ECCE 2021) - PROCEEDINGS. IEEE. 2021. p. 2755-2761. (IEEE Energy Conversion Congress and Exposition, ECCE - Proceedings). Epub 2021 Nov 16. doi: 10.1109/ECCE47101.2021.9595594

"Plug-and-Play" Tiny AI-Empowered Output Filter Parameter Extraction Framework with Single RNN Cell for Digital Power

Abstract

Publication series

Conference

Research Keywords

Access Output

Other Access Options

Permanent Link

Other Files and Links

Fingerprint

Cite this