Artificial Intelligence Aided Automated Design for Reliability of Power Electronic Systems

Research output: Contribution to journalJournal articleResearchpeer-review

2 Citations (Scopus)
354 Downloads (Pure)

Abstract

This paper proposes a new methodology for automated design of power electronic systems realized through the use of artificial intelligence. Existing approaches do not consider the system's reliability as a performance metric or are limited to reliability evaluation for a certain fixed set of design parameters. The method proposed in this paper establishes a functional relationship between design parameters and reliability metrics, and uses them as the basis for optimal design. The first step in this new framework is to create a nonparametric surrogate model of the power converter that can quickly map the variables characterizing the operating conditions (e.g., ambient temperature and irradiation) and design parameters (e.g., switching frequency and dc link voltage) into variables characterizing the thermal stress of a converter (e.g., mean temperature and temperature variation of its devices). This step can be carried out by training a dedicated artificial neural network (ANN) either on experimental or simulation data. The resulting network is named as text{ANN}-{1} and can be deployed as an accurate surrogate converter model. This model can then be used to quickly map the yearly mission profile into a thermal stress profile of any selected device for a large set of design parameter values. The resulting data is then used to train text{ANN}-{2}, which becomes an overall system representation that explicitly maps the design parameters into a yearly lifetime consumption. To verify the proposed methodology, text{ANN}-{2} is deployed in conjunction with the standard converter design tools on an exemplary grid-connected PV converter case study. This study showed how to find the optimal balance between the reliability and output filter size in the system with respect to several design constraints. This paper is also accompanied by a comprehensive dataset that was used for training the ANNs.

Original languageEnglish
Article number8584133
JournalI E E E Transactions on Power Electronics
Volume34
Issue number8
Pages (from-to)7161 - 7171
Number of pages11
ISSN0885-8993
DOIs
Publication statusPublished - Aug 2019

Fingerprint

Power electronics
Artificial intelligence
Neural networks
Thermal stress
Power converters
Switching frequency
Temperature
Irradiation
Electric potential

Keywords

  • Automated design for reliability (ADfR)
  • Artificial intelligence
  • Power electronic systems

Cite this

@article{683c612dbc7e4848a86eca333e9891ea,
title = "Artificial Intelligence Aided Automated Design for Reliability of Power Electronic Systems",
abstract = "This paper proposes a new methodology for automated design of power electronic systems realized through the use of artificial intelligence. Existing approaches do not consider the system's reliability as a performance metric or are limited to reliability evaluation for a certain fixed set of design parameters. The method proposed in this paper establishes a functional relationship between design parameters and reliability metrics, and uses them as the basis for optimal design. The first step in this new framework is to create a nonparametric surrogate model of the power converter that can quickly map the variables characterizing the operating conditions (e.g., ambient temperature and irradiation) and design parameters (e.g., switching frequency and dc link voltage) into variables characterizing the thermal stress of a converter (e.g., mean temperature and temperature variation of its devices). This step can be carried out by training a dedicated artificial neural network (ANN) either on experimental or simulation data. The resulting network is named as text{ANN}-{1} and can be deployed as an accurate surrogate converter model. This model can then be used to quickly map the yearly mission profile into a thermal stress profile of any selected device for a large set of design parameter values. The resulting data is then used to train text{ANN}-{2}, which becomes an overall system representation that explicitly maps the design parameters into a yearly lifetime consumption. To verify the proposed methodology, text{ANN}-{2} is deployed in conjunction with the standard converter design tools on an exemplary grid-connected PV converter case study. This study showed how to find the optimal balance between the reliability and output filter size in the system with respect to several design constraints. This paper is also accompanied by a comprehensive dataset that was used for training the ANNs.",
keywords = "Automated design for reliability (ADfR), Artificial intelligence, Power electronic systems",
author = "Tomislav Dragicevic and Patrick Wheeler and Frede Blaabjerg",
year = "2019",
month = "8",
doi = "10.1109/TPEL.2018.2883947",
language = "English",
volume = "34",
pages = "7161 -- 7171",
journal = "I E E E Transactions on Power Electronics",
issn = "0885-8993",
publisher = "IEEE",
number = "8",

}

Artificial Intelligence Aided Automated Design for Reliability of Power Electronic Systems. / Dragicevic, Tomislav; Wheeler, Patrick; Blaabjerg, Frede.

In: I E E E Transactions on Power Electronics, Vol. 34, No. 8, 8584133, 08.2019, p. 7161 - 7171.

Research output: Contribution to journalJournal articleResearchpeer-review

TY - JOUR

T1 - Artificial Intelligence Aided Automated Design for Reliability of Power Electronic Systems

AU - Dragicevic, Tomislav

AU - Wheeler, Patrick

AU - Blaabjerg, Frede

PY - 2019/8

Y1 - 2019/8

N2 - This paper proposes a new methodology for automated design of power electronic systems realized through the use of artificial intelligence. Existing approaches do not consider the system's reliability as a performance metric or are limited to reliability evaluation for a certain fixed set of design parameters. The method proposed in this paper establishes a functional relationship between design parameters and reliability metrics, and uses them as the basis for optimal design. The first step in this new framework is to create a nonparametric surrogate model of the power converter that can quickly map the variables characterizing the operating conditions (e.g., ambient temperature and irradiation) and design parameters (e.g., switching frequency and dc link voltage) into variables characterizing the thermal stress of a converter (e.g., mean temperature and temperature variation of its devices). This step can be carried out by training a dedicated artificial neural network (ANN) either on experimental or simulation data. The resulting network is named as text{ANN}-{1} and can be deployed as an accurate surrogate converter model. This model can then be used to quickly map the yearly mission profile into a thermal stress profile of any selected device for a large set of design parameter values. The resulting data is then used to train text{ANN}-{2}, which becomes an overall system representation that explicitly maps the design parameters into a yearly lifetime consumption. To verify the proposed methodology, text{ANN}-{2} is deployed in conjunction with the standard converter design tools on an exemplary grid-connected PV converter case study. This study showed how to find the optimal balance between the reliability and output filter size in the system with respect to several design constraints. This paper is also accompanied by a comprehensive dataset that was used for training the ANNs.

AB - This paper proposes a new methodology for automated design of power electronic systems realized through the use of artificial intelligence. Existing approaches do not consider the system's reliability as a performance metric or are limited to reliability evaluation for a certain fixed set of design parameters. The method proposed in this paper establishes a functional relationship between design parameters and reliability metrics, and uses them as the basis for optimal design. The first step in this new framework is to create a nonparametric surrogate model of the power converter that can quickly map the variables characterizing the operating conditions (e.g., ambient temperature and irradiation) and design parameters (e.g., switching frequency and dc link voltage) into variables characterizing the thermal stress of a converter (e.g., mean temperature and temperature variation of its devices). This step can be carried out by training a dedicated artificial neural network (ANN) either on experimental or simulation data. The resulting network is named as text{ANN}-{1} and can be deployed as an accurate surrogate converter model. This model can then be used to quickly map the yearly mission profile into a thermal stress profile of any selected device for a large set of design parameter values. The resulting data is then used to train text{ANN}-{2}, which becomes an overall system representation that explicitly maps the design parameters into a yearly lifetime consumption. To verify the proposed methodology, text{ANN}-{2} is deployed in conjunction with the standard converter design tools on an exemplary grid-connected PV converter case study. This study showed how to find the optimal balance between the reliability and output filter size in the system with respect to several design constraints. This paper is also accompanied by a comprehensive dataset that was used for training the ANNs.

KW - Automated design for reliability (ADfR)

KW - Artificial intelligence

KW - Power electronic systems

UR - http://www.scopus.com/inward/record.url?scp=85058877050&partnerID=8YFLogxK

U2 - 10.1109/TPEL.2018.2883947

DO - 10.1109/TPEL.2018.2883947

M3 - Journal article

VL - 34

SP - 7161

EP - 7171

JO - I E E E Transactions on Power Electronics

JF - I E E E Transactions on Power Electronics

SN - 0885-8993

IS - 8

M1 - 8584133

ER -