Enhancing Remaining Useful Life Predictions in Predictive Maintenance of MOSFETs: The Efficacy of Integrated Particle Filter-Gaussian Process Regression Models
In the field of predictive maintenance, accurately predicting the remaining useful life (RUL) of equipment is critical to optimizing operations and avoiding unexpected failures. Traditional models such as Gaussian Process Regression (GPR), Particle Filter (PF), and Kalman Filter (KF) have been widel...
Saved in:
| Main Authors: | , , , , , |
|---|---|
| Other Authors: | |
| Format: | Article |
| Published: |
International Information and Engineering Technology Association
2025
|
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1833350976856129536 |
|---|---|
| author | Hadi E.F. bin Baharuddin M.Z. Zuhdi A.W.M. Ghadir G.K. Al-Tmimi H.M. Mustafa M.A. |
| author2 | 59508922600 |
| author_facet | 59508922600 Hadi E.F. bin Baharuddin M.Z. Zuhdi A.W.M. Ghadir G.K. Al-Tmimi H.M. Mustafa M.A. |
| author_sort | Hadi E.F. |
| building | UNITEN Library |
| collection | Institutional Repository |
| content_provider | Universiti Tenaga Nasional |
| content_source | UNITEN Institutional Repository |
| continent | Asia |
| country | Malaysia |
| description | In the field of predictive maintenance, accurately predicting the remaining useful life (RUL) of equipment is critical to optimizing operations and avoiding unexpected failures. Traditional models such as Gaussian Process Regression (GPR), Particle Filter (PF), and Kalman Filter (KF) have been widely used, each with their own strengths and limitations. The motivation behind this study stems from the need to improve the accuracy and reliability of RUL predictions. Given the critical importance of predictive maintenance across a variety of industries, improved predictive models can provide significant operational and economic benefits. The main issues addressed are the limitations inherent in individual predictive models. GPR prediction gap or high initial error of his KF. This can lead to suboptimal RUL estimates. To address this problem, an integrated approach combining particle filtering and Gaussian process regression (PF-GPR) was proposed and developed. This integration aims to leverage the strengths of PF and GPR and potentially alleviate the limitations of the individual models. The performance of the PF-GPR model is evaluated and compared with the standalone His GPR, PF, and KF models using the prediction error and root mean square error (RMSE) at different points in the aging process of device #36. The results show that the PF-GPR model consistently outperforms the individual models in terms of both prediction error and RMSE, and significantly improves the accuracy and accuracy of RUL prediction. The PF-GPR model showed excellent performance in his RUL prediction for device #36, achieving the lowest prediction error and RMSE values at all time points as follows: B. The prediction error is only 2.30E-1 and the RMSE after 100 minutes is 7. 12E-3, which significantly outperforms his standalone GPR, PF, and KF models. The PF-GPR model demonstrated its excellent ability to provide robust and reliable predictions, highlighting the benefits of integrating different prediction methods into predictive maintenance applications. Copyright: ?2024 The authors. This article is published by IIETA. |
| format | Article |
| id | my.uniten.dspace-36657 |
| institution | Universiti Tenaga Nasional |
| publishDate | 2025 |
| publisher | International Information and Engineering Technology Association |
| record_format | dspace |
| spelling | my.uniten.dspace-366572025-03-03T15:43:42Z Enhancing Remaining Useful Life Predictions in Predictive Maintenance of MOSFETs: The Efficacy of Integrated Particle Filter-Gaussian Process Regression Models Hadi E.F. bin Baharuddin M.Z. Zuhdi A.W.M. Ghadir G.K. Al-Tmimi H.M. Mustafa M.A. 59508922600 59120482600 56589966300 58288988500 58946141500 57212482984 In the field of predictive maintenance, accurately predicting the remaining useful life (RUL) of equipment is critical to optimizing operations and avoiding unexpected failures. Traditional models such as Gaussian Process Regression (GPR), Particle Filter (PF), and Kalman Filter (KF) have been widely used, each with their own strengths and limitations. The motivation behind this study stems from the need to improve the accuracy and reliability of RUL predictions. Given the critical importance of predictive maintenance across a variety of industries, improved predictive models can provide significant operational and economic benefits. The main issues addressed are the limitations inherent in individual predictive models. GPR prediction gap or high initial error of his KF. This can lead to suboptimal RUL estimates. To address this problem, an integrated approach combining particle filtering and Gaussian process regression (PF-GPR) was proposed and developed. This integration aims to leverage the strengths of PF and GPR and potentially alleviate the limitations of the individual models. The performance of the PF-GPR model is evaluated and compared with the standalone His GPR, PF, and KF models using the prediction error and root mean square error (RMSE) at different points in the aging process of device #36. The results show that the PF-GPR model consistently outperforms the individual models in terms of both prediction error and RMSE, and significantly improves the accuracy and accuracy of RUL prediction. The PF-GPR model showed excellent performance in his RUL prediction for device #36, achieving the lowest prediction error and RMSE values at all time points as follows: B. The prediction error is only 2.30E-1 and the RMSE after 100 minutes is 7. 12E-3, which significantly outperforms his standalone GPR, PF, and KF models. The PF-GPR model demonstrated its excellent ability to provide robust and reliable predictions, highlighting the benefits of integrating different prediction methods into predictive maintenance applications. Copyright: ?2024 The authors. This article is published by IIETA. Final 2025-03-03T07:43:42Z 2025-03-03T07:43:42Z 2024 Article 10.18280/ijsse.140230 2-s2.0-85192779096 https://www.scopus.com/inward/record.uri?eid=2-s2.0-85192779096&doi=10.18280%2fijsse.140230&partnerID=40&md5=33ea42271efdbdd194b4c491203db553 https://irepository.uniten.edu.my/handle/123456789/36657 14 2 647 669 International Information and Engineering Technology Association Scopus |
| spellingShingle | Hadi E.F. bin Baharuddin M.Z. Zuhdi A.W.M. Ghadir G.K. Al-Tmimi H.M. Mustafa M.A. Enhancing Remaining Useful Life Predictions in Predictive Maintenance of MOSFETs: The Efficacy of Integrated Particle Filter-Gaussian Process Regression Models |
| title | Enhancing Remaining Useful Life Predictions in Predictive Maintenance of MOSFETs: The Efficacy of Integrated Particle Filter-Gaussian Process Regression Models |
| title_full | Enhancing Remaining Useful Life Predictions in Predictive Maintenance of MOSFETs: The Efficacy of Integrated Particle Filter-Gaussian Process Regression Models |
| title_fullStr | Enhancing Remaining Useful Life Predictions in Predictive Maintenance of MOSFETs: The Efficacy of Integrated Particle Filter-Gaussian Process Regression Models |
| title_full_unstemmed | Enhancing Remaining Useful Life Predictions in Predictive Maintenance of MOSFETs: The Efficacy of Integrated Particle Filter-Gaussian Process Regression Models |
| title_short | Enhancing Remaining Useful Life Predictions in Predictive Maintenance of MOSFETs: The Efficacy of Integrated Particle Filter-Gaussian Process Regression Models |
| title_sort | enhancing remaining useful life predictions in predictive maintenance of mosfets: the efficacy of integrated particle filter-gaussian process regression models |
| url_provider | http://dspace.uniten.edu.my/ |