Wind Turbine Gearbox Condition Monitoring Using Hybrid Attentions and Spatio-Temporal BiConvLSTM Network
Gearbox fault deterioration can significantly impact the safety, reliability, and efficiency of wind turbines, resulting in substantial economic losses for wind farms. However, current condition monitoring methods face challenges in effectively mining the hidden spatio-temporal features within SCADA...
Main Authors: | , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
MDPI AG
2023-09-01
|
Series: | Energies |
Subjects: | |
Online Access: | https://www.mdpi.com/1996-1073/16/19/6786 |
_version_ | 1797575936310247424 |
---|---|
author | Junshuai Yan Yongqian Liu Xiaoying Ren Li Li |
author_facet | Junshuai Yan Yongqian Liu Xiaoying Ren Li Li |
author_sort | Junshuai Yan |
collection | DOAJ |
description | Gearbox fault deterioration can significantly impact the safety, reliability, and efficiency of wind turbines, resulting in substantial economic losses for wind farms. However, current condition monitoring methods face challenges in effectively mining the hidden spatio-temporal features within SCADA data and establishing reasonable weight allocations for model input variables. To tackle these issues, we proposed a novel condition monitoring method for wind turbine gearboxes called HBCE, which integrated a feature-time hybrid attention mechanism (HA), the bidirectional convolutional long short-term memory networks (BiConvLSTM), and an improved exponentially weighted moving-average (iEWMA). Specifically, utilizing historical health SCADA data acquired through the modified Thompson tau data-cleaning algorithm, a normal behavior model (HA-BiConvLSTM) of gearbox was constructed to effectively extract the spatio-temporal features and learn normal behavior patterns. An iEWMA-based outlier detection approach was employed to set dynamic adaptive thresholds, and real-time monitor the prediction residuals of HA-BiConvLSTM to identify the early faults of gearbox. The proposed HBCE method was validated through actual gearbox faults and compared with conventional spatio-temporal models (i.e., CNN-LSTM and CNN&LSTM). The results illustrated that the constructed HA-BiConvLSTM model achieved superior prediction precision in terms of RMSE, MAE, MAPE, and R<sup>2</sup>, and the proposed method HBCE can effectively and reliably identify early anomalies of a wind turbine gearbox in advance. |
first_indexed | 2024-03-10T21:45:22Z |
format | Article |
id | doaj.art-11678a54a27a4523b71e87f1e1b14a11 |
institution | Directory Open Access Journal |
issn | 1996-1073 |
language | English |
last_indexed | 2024-03-10T21:45:22Z |
publishDate | 2023-09-01 |
publisher | MDPI AG |
record_format | Article |
series | Energies |
spelling | doaj.art-11678a54a27a4523b71e87f1e1b14a112023-11-19T14:18:52ZengMDPI AGEnergies1996-10732023-09-011619678610.3390/en16196786Wind Turbine Gearbox Condition Monitoring Using Hybrid Attentions and Spatio-Temporal BiConvLSTM NetworkJunshuai Yan0Yongqian Liu1Xiaoying Ren2Li Li3School of New Energy, North China Electric Power University, Beijing 102206, ChinaSchool of New Energy, North China Electric Power University, Beijing 102206, ChinaSchool of New Energy, North China Electric Power University, Beijing 102206, ChinaSchool of New Energy, North China Electric Power University, Beijing 102206, ChinaGearbox fault deterioration can significantly impact the safety, reliability, and efficiency of wind turbines, resulting in substantial economic losses for wind farms. However, current condition monitoring methods face challenges in effectively mining the hidden spatio-temporal features within SCADA data and establishing reasonable weight allocations for model input variables. To tackle these issues, we proposed a novel condition monitoring method for wind turbine gearboxes called HBCE, which integrated a feature-time hybrid attention mechanism (HA), the bidirectional convolutional long short-term memory networks (BiConvLSTM), and an improved exponentially weighted moving-average (iEWMA). Specifically, utilizing historical health SCADA data acquired through the modified Thompson tau data-cleaning algorithm, a normal behavior model (HA-BiConvLSTM) of gearbox was constructed to effectively extract the spatio-temporal features and learn normal behavior patterns. An iEWMA-based outlier detection approach was employed to set dynamic adaptive thresholds, and real-time monitor the prediction residuals of HA-BiConvLSTM to identify the early faults of gearbox. The proposed HBCE method was validated through actual gearbox faults and compared with conventional spatio-temporal models (i.e., CNN-LSTM and CNN&LSTM). The results illustrated that the constructed HA-BiConvLSTM model achieved superior prediction precision in terms of RMSE, MAE, MAPE, and R<sup>2</sup>, and the proposed method HBCE can effectively and reliably identify early anomalies of a wind turbine gearbox in advance.https://www.mdpi.com/1996-1073/16/19/6786wind turbine gearboxcondition monitoringattention mechanismconvolutional long short-term memory networkadaptive threshold |
spellingShingle | Junshuai Yan Yongqian Liu Xiaoying Ren Li Li Wind Turbine Gearbox Condition Monitoring Using Hybrid Attentions and Spatio-Temporal BiConvLSTM Network Energies wind turbine gearbox condition monitoring attention mechanism convolutional long short-term memory network adaptive threshold |
title | Wind Turbine Gearbox Condition Monitoring Using Hybrid Attentions and Spatio-Temporal BiConvLSTM Network |
title_full | Wind Turbine Gearbox Condition Monitoring Using Hybrid Attentions and Spatio-Temporal BiConvLSTM Network |
title_fullStr | Wind Turbine Gearbox Condition Monitoring Using Hybrid Attentions and Spatio-Temporal BiConvLSTM Network |
title_full_unstemmed | Wind Turbine Gearbox Condition Monitoring Using Hybrid Attentions and Spatio-Temporal BiConvLSTM Network |
title_short | Wind Turbine Gearbox Condition Monitoring Using Hybrid Attentions and Spatio-Temporal BiConvLSTM Network |
title_sort | wind turbine gearbox condition monitoring using hybrid attentions and spatio temporal biconvlstm network |
topic | wind turbine gearbox condition monitoring attention mechanism convolutional long short-term memory network adaptive threshold |
url | https://www.mdpi.com/1996-1073/16/19/6786 |
work_keys_str_mv | AT junshuaiyan windturbinegearboxconditionmonitoringusinghybridattentionsandspatiotemporalbiconvlstmnetwork AT yongqianliu windturbinegearboxconditionmonitoringusinghybridattentionsandspatiotemporalbiconvlstmnetwork AT xiaoyingren windturbinegearboxconditionmonitoringusinghybridattentionsandspatiotemporalbiconvlstmnetwork AT lili windturbinegearboxconditionmonitoringusinghybridattentionsandspatiotemporalbiconvlstmnetwork |