Neural Network Models and Transfer Learning for Impedance Modeling of Grid-Tied Inverters

Yufei Li; Yicheng Liao; Xiongfei Wang; Lars Nordstrom; Prateek Mittal; Minjie Chen; H. Vincent Poor

doi:10.1109/PEDG54999.2022.9923064

Neural Network Models and Transfer Learning for Impedance Modeling of Grid-Tied Inverters

Yufei Li, Yicheng Liao, Xiongfei Wang, Lars Nordstrom, Prateek Mittal, Minjie Chen, H. Vincent Poor

Research output: Contribution to book/anthology/report/conference proceeding › Article in proceeding › Research › peer-review

4 Citations (Scopus)

Abstract

The future power grid will be supported by a large number of grid-tied inverters whose dynamics are critical for grid stability and power flow control. The operating conditions of these inverters vary across a wide range, leading to different small-signal impedances and different grid-interface behaviors. Analytical impedance models derived at specific operating points can hardly capture nonlinearities and nonidealities of the physical systems. The applicability of electromagnetic transient (EMT) simulations is often limited by the system complexity and the available computational resources. This paper applies neural network and transfer learning to impedance modeling of grid-tied inverters. It is shown that a neural network (NN) trained by data automatically acquired from EMT simulations outperforms the one trained by traditional analytical models when unknown information exist in simulations. Pre-training the NN with analytically calculated data can greatly reduce the amount of simulation data needed to achieve good modeling results.

Original language	English
Title of host publication	2022 IEEE 13th International Symposium on Power Electronics for Distributed Generation Systems, PEDG 2022
Publisher	IEEE
Publication date	2022
ISBN (Electronic)	9781665466189
DOIs	https://doi.org/10.1109/PEDG54999.2022.9923064
Publication status	Published - 2022
Externally published	Yes
Event	13th IEEE International Symposium on Power Electronics for Distributed Generation Systems, PEDG 2022 - Kiel, Germany Duration: 26 Jun 2022 → 29 Jun 2022

Conference

Conference	13th IEEE International Symposium on Power Electronics for Distributed Generation Systems, PEDG 2022
Country/Territory	Germany
City	Kiel
Period	26/06/2022 → 29/06/2022

Series	2022 IEEE 13th International Symposium on Power Electronics for Distributed Generation Systems, PEDG 2022

Bibliographical note

Funding Information:
This work was supported in part by the C3.ai Digital Transformation Institute as a collaboration between Princeton University and the KTH Royal Institute of Technology Fig. 1. Modeling inverter impedances as neural networks for stability analysis of an active distribution network with many different inverters.

Publisher Copyright:
© 2022 IEEE.

Keywords

grid-tied inverter
impedance
machine learning
Neural network
small-signal model
transfer learning

Access to Document

10.1109/PEDG54999.2022.9923064

AUB Link

Search for the material in Aalborg University Library's search engine

Cite this

Li, Yufei ; Liao, Yicheng ; Wang, Xiongfei et al. / Neural Network Models and Transfer Learning for Impedance Modeling of Grid-Tied Inverters. 2022 IEEE 13th International Symposium on Power Electronics for Distributed Generation Systems, PEDG 2022. IEEE, 2022. (2022 IEEE 13th International Symposium on Power Electronics for Distributed Generation Systems, PEDG 2022).

@inproceedings{afee65dbf65b4978aed879621672439c,

title = "Neural Network Models and Transfer Learning for Impedance Modeling of Grid-Tied Inverters",

abstract = "The future power grid will be supported by a large number of grid-tied inverters whose dynamics are critical for grid stability and power flow control. The operating conditions of these inverters vary across a wide range, leading to different small-signal impedances and different grid-interface behaviors. Analytical impedance models derived at specific operating points can hardly capture nonlinearities and nonidealities of the physical systems. The applicability of electromagnetic transient (EMT) simulations is often limited by the system complexity and the available computational resources. This paper applies neural network and transfer learning to impedance modeling of grid-tied inverters. It is shown that a neural network (NN) trained by data automatically acquired from EMT simulations outperforms the one trained by traditional analytical models when unknown information exist in simulations. Pre-training the NN with analytically calculated data can greatly reduce the amount of simulation data needed to achieve good modeling results.",

keywords = "grid-tied inverter, impedance, machine learning, Neural network, small-signal model, transfer learning",

author = "Yufei Li and Yicheng Liao and Xiongfei Wang and Lars Nordstrom and Prateek Mittal and Minjie Chen and Poor, {H. Vincent}",

note = "Funding Information: This work was supported in part by the C3.ai Digital Transformation Institute as a collaboration between Princeton University and the KTH Royal Institute of Technology Fig. 1. Modeling inverter impedances as neural networks for stability analysis of an active distribution network with many different inverters. Publisher Copyright: {\textcopyright} 2022 IEEE.; 13th IEEE International Symposium on Power Electronics for Distributed Generation Systems, PEDG 2022 ; Conference date: 26-06-2022 Through 29-06-2022",

year = "2022",

doi = "10.1109/PEDG54999.2022.9923064",

language = "English",

series = "2022 IEEE 13th International Symposium on Power Electronics for Distributed Generation Systems, PEDG 2022",

booktitle = "2022 IEEE 13th International Symposium on Power Electronics for Distributed Generation Systems, PEDG 2022",

publisher = "IEEE",

address = "United States",

}

Li, Y, Liao, Y, Wang, X, Nordstrom, L, Mittal, P, Chen, M & Poor, HV 2022, Neural Network Models and Transfer Learning for Impedance Modeling of Grid-Tied Inverters. in 2022 IEEE 13th International Symposium on Power Electronics for Distributed Generation Systems, PEDG 2022. IEEE, 2022 IEEE 13th International Symposium on Power Electronics for Distributed Generation Systems, PEDG 2022, 13th IEEE International Symposium on Power Electronics for Distributed Generation Systems, PEDG 2022, Kiel, Germany, 26/06/2022. https://doi.org/10.1109/PEDG54999.2022.9923064

Neural Network Models and Transfer Learning for Impedance Modeling of Grid-Tied Inverters. / Li, Yufei; Liao, Yicheng; Wang, Xiongfei et al.
2022 IEEE 13th International Symposium on Power Electronics for Distributed Generation Systems, PEDG 2022. IEEE, 2022. (2022 IEEE 13th International Symposium on Power Electronics for Distributed Generation Systems, PEDG 2022).

Research output: Contribution to book/anthology/report/conference proceeding › Article in proceeding › Research › peer-review

TY - GEN

T1 - Neural Network Models and Transfer Learning for Impedance Modeling of Grid-Tied Inverters

AU - Li, Yufei

AU - Liao, Yicheng

AU - Wang, Xiongfei

AU - Nordstrom, Lars

AU - Mittal, Prateek

AU - Chen, Minjie

AU - Poor, H. Vincent

N1 - Funding Information: This work was supported in part by the C3.ai Digital Transformation Institute as a collaboration between Princeton University and the KTH Royal Institute of Technology Fig. 1. Modeling inverter impedances as neural networks for stability analysis of an active distribution network with many different inverters. Publisher Copyright: © 2022 IEEE.

PY - 2022

Y1 - 2022

N2 - The future power grid will be supported by a large number of grid-tied inverters whose dynamics are critical for grid stability and power flow control. The operating conditions of these inverters vary across a wide range, leading to different small-signal impedances and different grid-interface behaviors. Analytical impedance models derived at specific operating points can hardly capture nonlinearities and nonidealities of the physical systems. The applicability of electromagnetic transient (EMT) simulations is often limited by the system complexity and the available computational resources. This paper applies neural network and transfer learning to impedance modeling of grid-tied inverters. It is shown that a neural network (NN) trained by data automatically acquired from EMT simulations outperforms the one trained by traditional analytical models when unknown information exist in simulations. Pre-training the NN with analytically calculated data can greatly reduce the amount of simulation data needed to achieve good modeling results.

AB - The future power grid will be supported by a large number of grid-tied inverters whose dynamics are critical for grid stability and power flow control. The operating conditions of these inverters vary across a wide range, leading to different small-signal impedances and different grid-interface behaviors. Analytical impedance models derived at specific operating points can hardly capture nonlinearities and nonidealities of the physical systems. The applicability of electromagnetic transient (EMT) simulations is often limited by the system complexity and the available computational resources. This paper applies neural network and transfer learning to impedance modeling of grid-tied inverters. It is shown that a neural network (NN) trained by data automatically acquired from EMT simulations outperforms the one trained by traditional analytical models when unknown information exist in simulations. Pre-training the NN with analytically calculated data can greatly reduce the amount of simulation data needed to achieve good modeling results.

KW - grid-tied inverter

KW - impedance

KW - machine learning

KW - Neural network

KW - small-signal model

KW - transfer learning

UR - http://www.scopus.com/inward/record.url?scp=85142099958&partnerID=8YFLogxK

U2 - 10.1109/PEDG54999.2022.9923064

DO - 10.1109/PEDG54999.2022.9923064

M3 - Article in proceeding

AN - SCOPUS:85142099958

T3 - 2022 IEEE 13th International Symposium on Power Electronics for Distributed Generation Systems, PEDG 2022

BT - 2022 IEEE 13th International Symposium on Power Electronics for Distributed Generation Systems, PEDG 2022

PB - IEEE

T2 - 13th IEEE International Symposium on Power Electronics for Distributed Generation Systems, PEDG 2022

Y2 - 26 June 2022 through 29 June 2022

ER -

Li Y, Liao Y, Wang X, Nordstrom L, Mittal P, Chen M et al. Neural Network Models and Transfer Learning for Impedance Modeling of Grid-Tied Inverters. In 2022 IEEE 13th International Symposium on Power Electronics for Distributed Generation Systems, PEDG 2022. IEEE. 2022. (2022 IEEE 13th International Symposium on Power Electronics for Distributed Generation Systems, PEDG 2022). doi: 10.1109/PEDG54999.2022.9923064

Neural Network Models and Transfer Learning for Impedance Modeling of Grid-Tied Inverters

Abstract

Conference

Bibliographical note

Keywords

Access to Document

AUB Link

Other files and links

Fingerprint

Cite this