Remaining useful lifetime estimation for discrete power electronic devices using physics-informed neural network
Abstract Estimation of Remaining Useful Lifetime (RUL) of discrete power electronics is important to enable predictive maintenance and ensure system safety. Conventional data-driven approaches using neural networks have been applied to address this challenge. However, due to ignoring the physical pr...
Main Authors: | , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
Nature Portfolio
2023-06-01
|
Series: | Scientific Reports |
Online Access: | https://doi.org/10.1038/s41598-023-37154-5 |
_version_ | 1797795717421465600 |
---|---|
author | Zhonghai Lu Chao Guo Mingrui Liu Rui Shi |
author_facet | Zhonghai Lu Chao Guo Mingrui Liu Rui Shi |
author_sort | Zhonghai Lu |
collection | DOAJ |
description | Abstract Estimation of Remaining Useful Lifetime (RUL) of discrete power electronics is important to enable predictive maintenance and ensure system safety. Conventional data-driven approaches using neural networks have been applied to address this challenge. However, due to ignoring the physical properties of the target RUL function, neural networks can result in unreasonable RUL estimates such as going upwards and wrong endings. In the paper, we apply the fundamental principle of Physics-Informed Neural Network (PINN) to enhance Recurrent Neural Network (RNN) based RUL estimation methods. Through formulating proper constraints into the loss function of neural networks, we demonstrate in our experiments with the NASA IGBT dataset that PINN can make the neural networks trained more realistically and thus achieve performance improvements in estimation error and coefficient of determination. Compared to the baseline vanilla RNN, our physics-informed RNN can improve Mean Squared Error (MSE) of out-of-sample estimation on average by 24.7% in training and by 51.3% in testing; Compared to the baseline Long Short Term Memory (LSTM, a variant of RNN), our physics-informed LSTM can improve MSE of out-of-sample estimation on average by 15.3% in training and 13.9% in testing. |
first_indexed | 2024-03-13T03:22:15Z |
format | Article |
id | doaj.art-7909ace54c25414b96100fe69a50c1f4 |
institution | Directory Open Access Journal |
issn | 2045-2322 |
language | English |
last_indexed | 2024-03-13T03:22:15Z |
publishDate | 2023-06-01 |
publisher | Nature Portfolio |
record_format | Article |
series | Scientific Reports |
spelling | doaj.art-7909ace54c25414b96100fe69a50c1f42023-06-25T11:17:15ZengNature PortfolioScientific Reports2045-23222023-06-0113111510.1038/s41598-023-37154-5Remaining useful lifetime estimation for discrete power electronic devices using physics-informed neural networkZhonghai Lu0Chao Guo1Mingrui Liu2Rui Shi3Department of Electrical Engineering, KTH Royal Institute of TechnologyDepartment of Electrical Engineering, KTH Royal Institute of TechnologyDepartment of Electrical Engineering, KTH Royal Institute of TechnologyDepartment of Electrical Engineering, KTH Royal Institute of TechnologyAbstract Estimation of Remaining Useful Lifetime (RUL) of discrete power electronics is important to enable predictive maintenance and ensure system safety. Conventional data-driven approaches using neural networks have been applied to address this challenge. However, due to ignoring the physical properties of the target RUL function, neural networks can result in unreasonable RUL estimates such as going upwards and wrong endings. In the paper, we apply the fundamental principle of Physics-Informed Neural Network (PINN) to enhance Recurrent Neural Network (RNN) based RUL estimation methods. Through formulating proper constraints into the loss function of neural networks, we demonstrate in our experiments with the NASA IGBT dataset that PINN can make the neural networks trained more realistically and thus achieve performance improvements in estimation error and coefficient of determination. Compared to the baseline vanilla RNN, our physics-informed RNN can improve Mean Squared Error (MSE) of out-of-sample estimation on average by 24.7% in training and by 51.3% in testing; Compared to the baseline Long Short Term Memory (LSTM, a variant of RNN), our physics-informed LSTM can improve MSE of out-of-sample estimation on average by 15.3% in training and 13.9% in testing.https://doi.org/10.1038/s41598-023-37154-5 |
spellingShingle | Zhonghai Lu Chao Guo Mingrui Liu Rui Shi Remaining useful lifetime estimation for discrete power electronic devices using physics-informed neural network Scientific Reports |
title | Remaining useful lifetime estimation for discrete power electronic devices using physics-informed neural network |
title_full | Remaining useful lifetime estimation for discrete power electronic devices using physics-informed neural network |
title_fullStr | Remaining useful lifetime estimation for discrete power electronic devices using physics-informed neural network |
title_full_unstemmed | Remaining useful lifetime estimation for discrete power electronic devices using physics-informed neural network |
title_short | Remaining useful lifetime estimation for discrete power electronic devices using physics-informed neural network |
title_sort | remaining useful lifetime estimation for discrete power electronic devices using physics informed neural network |
url | https://doi.org/10.1038/s41598-023-37154-5 |
work_keys_str_mv | AT zhonghailu remainingusefullifetimeestimationfordiscretepowerelectronicdevicesusingphysicsinformedneuralnetwork AT chaoguo remainingusefullifetimeestimationfordiscretepowerelectronicdevicesusingphysicsinformedneuralnetwork AT mingruiliu remainingusefullifetimeestimationfordiscretepowerelectronicdevicesusingphysicsinformedneuralnetwork AT ruishi remainingusefullifetimeestimationfordiscretepowerelectronicdevicesusingphysicsinformedneuralnetwork |