Adaptive power transformer lifetime predictions through machine learning and uncertainty modelling in nuclear power plants

Jose Ignacio Aizpurua, Stephen D. J. McArthur, Brian G. Stewart, Brandon Lambert, James G. Cross, Victoria M. Catterson

Research output: Contribution to journalArticle

5 Citations (Scopus)

Abstract

The remaining useful life (RUL) of transformer insulation paper is largely determined by the winding hot-spot temperature (HST). Frequently the HST is not directly monitored and it is inferred from other measurements. However, measurement errors affect prediction models and if uncertain variables are not taken into account this can lead to incorrect maintenance decisions. Additionally, existing analytic models for HST calculation are not always accurate because they cannot generalize the properties of transformers operating in different contexts. In this context, this paper presents a novel transformer condition assessment approach integrating uncertainty modeling, data-driven forecasting models and model-based experimental models to increase the prediction accuracy and handle uncertainty. The proposed approach quantifies the effect of measurement errors on transformer RUL predictions and confirms that temperature and load measurement errors affect the RUL estimation. Forecasting results show that the extreme gradient boosting (XGB) algorithm best captures the non-linearities of the thermal model and improves the prediction accuracy amongst a number of forecasting approaches. Accordingly, the XGB model is integrated with experimental models in a Particle Filtering framework to improve thermal modelling and RUL prediction tasks. Models are tested and validated using a real dataset from a power transformer operating in a nuclear power plant.
LanguageEnglish
Pages4726-4737
Number of pages12
JournalIEEE Transactions on Industrial Electronics
Volume66
Issue number6
Early online date1 Aug 2018
DOIs
Publication statusPublished - 1 Jun 2019

Fingerprint

Power transformers
Nuclear power plants
Learning systems
Measurement errors
Uncertainty
Temperature
Data structures
Insulation

Keywords

  • condition assessment
  • forecasting
  • prognostics and health management
  • sensitivity
  • transformers

Cite this

@article{5a0e1bf3a32e4cbf98acf92679bd8c5a,
title = "Adaptive power transformer lifetime predictions through machine learning and uncertainty modelling in nuclear power plants",
abstract = "The remaining useful life (RUL) of transformer insulation paper is largely determined by the winding hot-spot temperature (HST). Frequently the HST is not directly monitored and it is inferred from other measurements. However, measurement errors affect prediction models and if uncertain variables are not taken into account this can lead to incorrect maintenance decisions. Additionally, existing analytic models for HST calculation are not always accurate because they cannot generalize the properties of transformers operating in different contexts. In this context, this paper presents a novel transformer condition assessment approach integrating uncertainty modeling, data-driven forecasting models and model-based experimental models to increase the prediction accuracy and handle uncertainty. The proposed approach quantifies the effect of measurement errors on transformer RUL predictions and confirms that temperature and load measurement errors affect the RUL estimation. Forecasting results show that the extreme gradient boosting (XGB) algorithm best captures the non-linearities of the thermal model and improves the prediction accuracy amongst a number of forecasting approaches. Accordingly, the XGB model is integrated with experimental models in a Particle Filtering framework to improve thermal modelling and RUL prediction tasks. Models are tested and validated using a real dataset from a power transformer operating in a nuclear power plant.",
keywords = "condition assessment, forecasting, prognostics and health management, sensitivity, transformers",
author = "Aizpurua, {Jose Ignacio} and McArthur, {Stephen D. J.} and Stewart, {Brian G.} and Brandon Lambert and Cross, {James G.} and Catterson, {Victoria M.}",
note = "{\circledC} 2018 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works.",
year = "2019",
month = "6",
day = "1",
doi = "10.1109/TIE.2018.2860532",
language = "English",
volume = "66",
pages = "4726--4737",
journal = "IEEE Transactions on Industrial Electronics",
issn = "0278-0046",
number = "6",

}

Adaptive power transformer lifetime predictions through machine learning and uncertainty modelling in nuclear power plants. / Aizpurua, Jose Ignacio; McArthur, Stephen D. J.; Stewart, Brian G.; Lambert, Brandon; Cross, James G.; Catterson, Victoria M.

In: IEEE Transactions on Industrial Electronics, Vol. 66, No. 6, 01.06.2019, p. 4726-4737.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Adaptive power transformer lifetime predictions through machine learning and uncertainty modelling in nuclear power plants

AU - Aizpurua, Jose Ignacio

AU - McArthur, Stephen D. J.

AU - Stewart, Brian G.

AU - Lambert, Brandon

AU - Cross, James G.

AU - Catterson, Victoria M.

N1 - © 2018 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works.

PY - 2019/6/1

Y1 - 2019/6/1

N2 - The remaining useful life (RUL) of transformer insulation paper is largely determined by the winding hot-spot temperature (HST). Frequently the HST is not directly monitored and it is inferred from other measurements. However, measurement errors affect prediction models and if uncertain variables are not taken into account this can lead to incorrect maintenance decisions. Additionally, existing analytic models for HST calculation are not always accurate because they cannot generalize the properties of transformers operating in different contexts. In this context, this paper presents a novel transformer condition assessment approach integrating uncertainty modeling, data-driven forecasting models and model-based experimental models to increase the prediction accuracy and handle uncertainty. The proposed approach quantifies the effect of measurement errors on transformer RUL predictions and confirms that temperature and load measurement errors affect the RUL estimation. Forecasting results show that the extreme gradient boosting (XGB) algorithm best captures the non-linearities of the thermal model and improves the prediction accuracy amongst a number of forecasting approaches. Accordingly, the XGB model is integrated with experimental models in a Particle Filtering framework to improve thermal modelling and RUL prediction tasks. Models are tested and validated using a real dataset from a power transformer operating in a nuclear power plant.

AB - The remaining useful life (RUL) of transformer insulation paper is largely determined by the winding hot-spot temperature (HST). Frequently the HST is not directly monitored and it is inferred from other measurements. However, measurement errors affect prediction models and if uncertain variables are not taken into account this can lead to incorrect maintenance decisions. Additionally, existing analytic models for HST calculation are not always accurate because they cannot generalize the properties of transformers operating in different contexts. In this context, this paper presents a novel transformer condition assessment approach integrating uncertainty modeling, data-driven forecasting models and model-based experimental models to increase the prediction accuracy and handle uncertainty. The proposed approach quantifies the effect of measurement errors on transformer RUL predictions and confirms that temperature and load measurement errors affect the RUL estimation. Forecasting results show that the extreme gradient boosting (XGB) algorithm best captures the non-linearities of the thermal model and improves the prediction accuracy amongst a number of forecasting approaches. Accordingly, the XGB model is integrated with experimental models in a Particle Filtering framework to improve thermal modelling and RUL prediction tasks. Models are tested and validated using a real dataset from a power transformer operating in a nuclear power plant.

KW - condition assessment

KW - forecasting

KW - prognostics and health management

KW - sensitivity

KW - transformers

UR - https://ieeexplore.ieee.org/xpl/RecentIssue.jsp?punumber=41

U2 - 10.1109/TIE.2018.2860532

DO - 10.1109/TIE.2018.2860532

M3 - Article

VL - 66

SP - 4726

EP - 4737

JO - IEEE Transactions on Industrial Electronics

T2 - IEEE Transactions on Industrial Electronics

JF - IEEE Transactions on Industrial Electronics

SN - 0278-0046

IS - 6

ER -