Modified Singular Spectrum Decomposition and Its Application to Composite Fault Diagnosis of Gearboxes
Under the strong noise environment, the composite fault signal of gearbox is weak, which makes it difficult to extract fault features. For this problem, based on noise-assisted method, we propose a novel method called Modified Singular Spectrum Decomposition (MSSD). Singular Spectrum Decomposition (...
| Main Authors: | , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2018-12-01
|
| Series: | Sensors |
| Subjects: | |
| Online Access: | http://www.mdpi.com/1424-8220/19/1/62 |
| _version_ | 1811184732606562304 |
|---|---|
| author | Junyuan Wang Xiaofeng Han Zhijian Wang Wenhua Du Jie Zhou Jiping Zhang Huihui He Xiaoming Guo |
| author_facet | Junyuan Wang Xiaofeng Han Zhijian Wang Wenhua Du Jie Zhou Jiping Zhang Huihui He Xiaoming Guo |
| author_sort | Junyuan Wang |
| collection | DOAJ |
| description | Under the strong noise environment, the composite fault signal of gearbox is weak, which makes it difficult to extract fault features. For this problem, based on noise-assisted method, we propose a novel method called Modified Singular Spectrum Decomposition (MSSD). Singular Spectrum Decomposition (SSD) has many advantages such as high decomposition precision and strong ability to restrain mode mixing, etc. However, the ability of SSD to extract a weak signal is not ideal, the decomposition results usually contain a lot of redundant noise and mode mixing caused by intermittency, which is also a troubling problem. In order to improve the decomposition efficiency and make up for the defects of SSD, the new method MSSD adds an adaptive and particular noise in every SSD decomposition stage for each trial, and in addition, whenever the input signal is decomposed to obtain an intrinsic module function (IMF), a unique residual is obtained. After multiple decomposition, the average value of the residual is used as input to the next stage, until the residual cannot continue to decompose, which means that the residual component has, at most, one extreme value. Finally, analyzing simulated signals to explain the advantages of MSSD compared to ensemble empirical mode decomposition (EEMD) and complete ensemble local mean decomposition with adaptive noise (CEEMDAN). In order to further prove the effectiveness of MSSD, this new method, MSSD, is applied to the fault diagnosis of an engineering gearbox test stand in an actual engineer project case. The final results show that MSSD can extract more fault feature information, and mode mixing has been improved and suffers less interference compared to SSD. |
| first_indexed | 2024-04-11T13:18:29Z |
| format | Article |
| id | doaj.art-c4a9974b3e0849f5915685cc8bb1d5a0 |
| institution | Directory Open Access Journal |
| issn | 1424-8220 |
| language | English |
| last_indexed | 2024-04-11T13:18:29Z |
| publishDate | 2018-12-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Sensors |
| spelling | doaj.art-c4a9974b3e0849f5915685cc8bb1d5a02022-12-22T04:22:20ZengMDPI AGSensors1424-82202018-12-011916210.3390/s19010062s19010062Modified Singular Spectrum Decomposition and Its Application to Composite Fault Diagnosis of GearboxesJunyuan Wang0Xiaofeng Han1Zhijian Wang2Wenhua Du3Jie Zhou4Jiping Zhang5Huihui He6Xiaoming Guo7College of Mechanical Engineering, North University of China, Taiyuan 030051, ChinaCollege of Mechanical Engineering, North University of China, Taiyuan 030051, ChinaCollege of Mechanical Engineering, North University of China, Taiyuan 030051, ChinaCollege of Mechanical Engineering, North University of China, Taiyuan 030051, ChinaCollege of Mechanical Engineering, North University of China, Taiyuan 030051, ChinaCollege of Mechanical Engineering, North University of China, Taiyuan 030051, ChinaCollege of Mechanical Engineering, North University of China, Taiyuan 030051, ChinaCollege of Mechanical Engineering, North University of China, Taiyuan 030051, ChinaUnder the strong noise environment, the composite fault signal of gearbox is weak, which makes it difficult to extract fault features. For this problem, based on noise-assisted method, we propose a novel method called Modified Singular Spectrum Decomposition (MSSD). Singular Spectrum Decomposition (SSD) has many advantages such as high decomposition precision and strong ability to restrain mode mixing, etc. However, the ability of SSD to extract a weak signal is not ideal, the decomposition results usually contain a lot of redundant noise and mode mixing caused by intermittency, which is also a troubling problem. In order to improve the decomposition efficiency and make up for the defects of SSD, the new method MSSD adds an adaptive and particular noise in every SSD decomposition stage for each trial, and in addition, whenever the input signal is decomposed to obtain an intrinsic module function (IMF), a unique residual is obtained. After multiple decomposition, the average value of the residual is used as input to the next stage, until the residual cannot continue to decompose, which means that the residual component has, at most, one extreme value. Finally, analyzing simulated signals to explain the advantages of MSSD compared to ensemble empirical mode decomposition (EEMD) and complete ensemble local mean decomposition with adaptive noise (CEEMDAN). In order to further prove the effectiveness of MSSD, this new method, MSSD, is applied to the fault diagnosis of an engineering gearbox test stand in an actual engineer project case. The final results show that MSSD can extract more fault feature information, and mode mixing has been improved and suffers less interference compared to SSD.http://www.mdpi.com/1424-8220/19/1/62Gearbox composite faultalgorithmfault diagnosismodified singular spectrum decomposition |
| spellingShingle | Junyuan Wang Xiaofeng Han Zhijian Wang Wenhua Du Jie Zhou Jiping Zhang Huihui He Xiaoming Guo Modified Singular Spectrum Decomposition and Its Application to Composite Fault Diagnosis of Gearboxes Sensors Gearbox composite fault algorithm fault diagnosis modified singular spectrum decomposition |
| title | Modified Singular Spectrum Decomposition and Its Application to Composite Fault Diagnosis of Gearboxes |
| title_full | Modified Singular Spectrum Decomposition and Its Application to Composite Fault Diagnosis of Gearboxes |
| title_fullStr | Modified Singular Spectrum Decomposition and Its Application to Composite Fault Diagnosis of Gearboxes |
| title_full_unstemmed | Modified Singular Spectrum Decomposition and Its Application to Composite Fault Diagnosis of Gearboxes |
| title_short | Modified Singular Spectrum Decomposition and Its Application to Composite Fault Diagnosis of Gearboxes |
| title_sort | modified singular spectrum decomposition and its application to composite fault diagnosis of gearboxes |
| topic | Gearbox composite fault algorithm fault diagnosis modified singular spectrum decomposition |
| url | http://www.mdpi.com/1424-8220/19/1/62 |
| work_keys_str_mv | AT junyuanwang modifiedsingularspectrumdecompositionanditsapplicationtocompositefaultdiagnosisofgearboxes AT xiaofenghan modifiedsingularspectrumdecompositionanditsapplicationtocompositefaultdiagnosisofgearboxes AT zhijianwang modifiedsingularspectrumdecompositionanditsapplicationtocompositefaultdiagnosisofgearboxes AT wenhuadu modifiedsingularspectrumdecompositionanditsapplicationtocompositefaultdiagnosisofgearboxes AT jiezhou modifiedsingularspectrumdecompositionanditsapplicationtocompositefaultdiagnosisofgearboxes AT jipingzhang modifiedsingularspectrumdecompositionanditsapplicationtocompositefaultdiagnosisofgearboxes AT huihuihe modifiedsingularspectrumdecompositionanditsapplicationtocompositefaultdiagnosisofgearboxes AT xiaomingguo modifiedsingularspectrumdecompositionanditsapplicationtocompositefaultdiagnosisofgearboxes |