A Time-Distributed Spatiotemporal Feature Learning Method for Machine Health Monitoring with Multi-Sensor Time Series
Data-driven methods with multi-sensor time series data are the most promising approaches for monitoring machine health. Extracting fault-sensitive features from multi-sensor time series is a daunting task for both traditional data-driven methods and current deep learning models. A novel hybrid end-t...
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2018-09-01
|
| Series: | Sensors |
| Subjects: | |
| Online Access: | http://www.mdpi.com/1424-8220/18/9/2932 |
| _version_ | 1798039045375262720 |
|---|---|
| author | Huihui Qiao Taiyong Wang Peng Wang Shibin Qiao Lan Zhang |
| author_facet | Huihui Qiao Taiyong Wang Peng Wang Shibin Qiao Lan Zhang |
| author_sort | Huihui Qiao |
| collection | DOAJ |
| description | Data-driven methods with multi-sensor time series data are the most promising approaches for monitoring machine health. Extracting fault-sensitive features from multi-sensor time series is a daunting task for both traditional data-driven methods and current deep learning models. A novel hybrid end-to-end deep learning framework named Time-distributed ConvLSTM model (TDConvLSTM) is proposed in the paper for machine health monitoring, which works directly on raw multi-sensor time series. In TDConvLSTM, the normalized multi-sensor data is first segmented into a collection of subsequences by a sliding window along the temporal dimension. Time-distributed local feature extractors are simultaneously applied to each subsequence to extract local spatiotemporal features. Then a holistic ConvLSTM layer is designed to extract holistic spatiotemporal features between subsequences. At last, a fully-connected layer and a supervised learning layer are stacked on the top of the model to obtain the target. TDConvLSTM can extract spatiotemporal features on different time scales without any handcrafted feature engineering. The proposed model can achieve better performance in both time series classification tasks and regression prediction tasks than some state-of-the-art models, which has been verified in the gearbox fault diagnosis experiment and the tool wear prediction experiment. |
| first_indexed | 2024-04-11T21:48:30Z |
| format | Article |
| id | doaj.art-5f73ada7aa3e40cfa7b49edc9e31d4b8 |
| institution | Directory Open Access Journal |
| issn | 1424-8220 |
| language | English |
| last_indexed | 2024-04-11T21:48:30Z |
| publishDate | 2018-09-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Sensors |
| spelling | doaj.art-5f73ada7aa3e40cfa7b49edc9e31d4b82022-12-22T04:01:19ZengMDPI AGSensors1424-82202018-09-01189293210.3390/s18092932s18092932A Time-Distributed Spatiotemporal Feature Learning Method for Machine Health Monitoring with Multi-Sensor Time SeriesHuihui Qiao0Taiyong Wang1Peng Wang2Shibin Qiao3Lan Zhang4Key Laboratory of Mechanism Theory and Equipment Design of Ministry of Education, Tianjin University, Tianjin 300350, ChinaKey Laboratory of Mechanism Theory and Equipment Design of Ministry of Education, Tianjin University, Tianjin 300350, ChinaKey Laboratory of Mechanism Theory and Equipment Design of Ministry of Education, Tianjin University, Tianjin 300350, ChinaInstitute for Special Steels, Central Iron and Steel Research Institute, Beijing 100081, ChinaKey Laboratory of Mechanism Theory and Equipment Design of Ministry of Education, Tianjin University, Tianjin 300350, ChinaData-driven methods with multi-sensor time series data are the most promising approaches for monitoring machine health. Extracting fault-sensitive features from multi-sensor time series is a daunting task for both traditional data-driven methods and current deep learning models. A novel hybrid end-to-end deep learning framework named Time-distributed ConvLSTM model (TDConvLSTM) is proposed in the paper for machine health monitoring, which works directly on raw multi-sensor time series. In TDConvLSTM, the normalized multi-sensor data is first segmented into a collection of subsequences by a sliding window along the temporal dimension. Time-distributed local feature extractors are simultaneously applied to each subsequence to extract local spatiotemporal features. Then a holistic ConvLSTM layer is designed to extract holistic spatiotemporal features between subsequences. At last, a fully-connected layer and a supervised learning layer are stacked on the top of the model to obtain the target. TDConvLSTM can extract spatiotemporal features on different time scales without any handcrafted feature engineering. The proposed model can achieve better performance in both time series classification tasks and regression prediction tasks than some state-of-the-art models, which has been verified in the gearbox fault diagnosis experiment and the tool wear prediction experiment.http://www.mdpi.com/1424-8220/18/9/2932multi-sensor time seriesdeep learningmachine health monitoringtime-distributed ConvLSTM modelspatiotemporal feature learning |
| spellingShingle | Huihui Qiao Taiyong Wang Peng Wang Shibin Qiao Lan Zhang A Time-Distributed Spatiotemporal Feature Learning Method for Machine Health Monitoring with Multi-Sensor Time Series Sensors multi-sensor time series deep learning machine health monitoring time-distributed ConvLSTM model spatiotemporal feature learning |
| title | A Time-Distributed Spatiotemporal Feature Learning Method for Machine Health Monitoring with Multi-Sensor Time Series |
| title_full | A Time-Distributed Spatiotemporal Feature Learning Method for Machine Health Monitoring with Multi-Sensor Time Series |
| title_fullStr | A Time-Distributed Spatiotemporal Feature Learning Method for Machine Health Monitoring with Multi-Sensor Time Series |
| title_full_unstemmed | A Time-Distributed Spatiotemporal Feature Learning Method for Machine Health Monitoring with Multi-Sensor Time Series |
| title_short | A Time-Distributed Spatiotemporal Feature Learning Method for Machine Health Monitoring with Multi-Sensor Time Series |
| title_sort | time distributed spatiotemporal feature learning method for machine health monitoring with multi sensor time series |
| topic | multi-sensor time series deep learning machine health monitoring time-distributed ConvLSTM model spatiotemporal feature learning |
| url | http://www.mdpi.com/1424-8220/18/9/2932 |
| work_keys_str_mv | AT huihuiqiao atimedistributedspatiotemporalfeaturelearningmethodformachinehealthmonitoringwithmultisensortimeseries AT taiyongwang atimedistributedspatiotemporalfeaturelearningmethodformachinehealthmonitoringwithmultisensortimeseries AT pengwang atimedistributedspatiotemporalfeaturelearningmethodformachinehealthmonitoringwithmultisensortimeseries AT shibinqiao atimedistributedspatiotemporalfeaturelearningmethodformachinehealthmonitoringwithmultisensortimeseries AT lanzhang atimedistributedspatiotemporalfeaturelearningmethodformachinehealthmonitoringwithmultisensortimeseries AT huihuiqiao timedistributedspatiotemporalfeaturelearningmethodformachinehealthmonitoringwithmultisensortimeseries AT taiyongwang timedistributedspatiotemporalfeaturelearningmethodformachinehealthmonitoringwithmultisensortimeseries AT pengwang timedistributedspatiotemporalfeaturelearningmethodformachinehealthmonitoringwithmultisensortimeseries AT shibinqiao timedistributedspatiotemporalfeaturelearningmethodformachinehealthmonitoringwithmultisensortimeseries AT lanzhang timedistributedspatiotemporalfeaturelearningmethodformachinehealthmonitoringwithmultisensortimeseries |