A Dynamic Adam Based Deep Neural Network for Fault Diagnosis of Oil-Immersed Power Transformers
This paper presents a Dynamic Adam and dropout based deep neural network (DADDNN) for fault diagnosis of oil-immersed power transformers. To solve the problem of incomplete extraction of hidden information with data driven, the gradient first-order moment estimate and second-order moment estimate ar...
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2019-03-01
|
| Series: | Energies |
| Subjects: | |
| Online Access: | http://www.mdpi.com/1996-1073/12/6/995 |
| _version_ | 1811213360818028544 |
|---|---|
| author | Minghui Ou Hua Wei Yiyi Zhang Jiancheng Tan |
| author_facet | Minghui Ou Hua Wei Yiyi Zhang Jiancheng Tan |
| author_sort | Minghui Ou |
| collection | DOAJ |
| description | This paper presents a Dynamic Adam and dropout based deep neural network (DADDNN) for fault diagnosis of oil-immersed power transformers. To solve the problem of incomplete extraction of hidden information with data driven, the gradient first-order moment estimate and second-order moment estimate are used to calculate the different learning rates for all parameters with stable gradient scaling. Meanwhile, the learning rate is dynamically attenuated according to the optimal interval. To prevent over-fitted, we exploit dropout technique to randomly reset some neurons and strengthen the information exchange between indirectly-linked neurons. Our proposed approach was utilized on four datasets to learn the faults diagnosis of oil-immersed power transformers. Besides, four benchmark cases in other fields were also utilized to illustrate its scalability. The simulation results show that the average diagnosis accuracies on the four datasets of our proposed method were 37.9%, 25.5%, 14.6%, 18.9%, and 11.2%, higher than international electro technical commission (IEC), Duval Triangle, stacked autoencoders (SAE), deep belief networks (DBN), and grid search support vector machines (GSSVM), respectively. |
| first_indexed | 2024-04-12T05:45:32Z |
| format | Article |
| id | doaj.art-9080435884db40beb2c4ee3fe32e5bb0 |
| institution | Directory Open Access Journal |
| issn | 1996-1073 |
| language | English |
| last_indexed | 2024-04-12T05:45:32Z |
| publishDate | 2019-03-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Energies |
| spelling | doaj.art-9080435884db40beb2c4ee3fe32e5bb02022-12-22T03:45:29ZengMDPI AGEnergies1996-10732019-03-0112699510.3390/en12060995en12060995A Dynamic Adam Based Deep Neural Network for Fault Diagnosis of Oil-Immersed Power TransformersMinghui Ou0Hua Wei1Yiyi Zhang2Jiancheng Tan3Guangxi Key Laboratory of Power System Optimization and Energy Technology, Guangxi University, Nanning 530004, ChinaGuangxi Key Laboratory of Power System Optimization and Energy Technology, Guangxi University, Nanning 530004, ChinaGuangxi Key Laboratory of Power System Optimization and Energy Technology, Guangxi University, Nanning 530004, ChinaCollege of Electrical Engineering, Guangxi University, Nanning 530004, ChinaThis paper presents a Dynamic Adam and dropout based deep neural network (DADDNN) for fault diagnosis of oil-immersed power transformers. To solve the problem of incomplete extraction of hidden information with data driven, the gradient first-order moment estimate and second-order moment estimate are used to calculate the different learning rates for all parameters with stable gradient scaling. Meanwhile, the learning rate is dynamically attenuated according to the optimal interval. To prevent over-fitted, we exploit dropout technique to randomly reset some neurons and strengthen the information exchange between indirectly-linked neurons. Our proposed approach was utilized on four datasets to learn the faults diagnosis of oil-immersed power transformers. Besides, four benchmark cases in other fields were also utilized to illustrate its scalability. The simulation results show that the average diagnosis accuracies on the four datasets of our proposed method were 37.9%, 25.5%, 14.6%, 18.9%, and 11.2%, higher than international electro technical commission (IEC), Duval Triangle, stacked autoencoders (SAE), deep belief networks (DBN), and grid search support vector machines (GSSVM), respectively.http://www.mdpi.com/1996-1073/12/6/995power transformerfault diagnosisdissolved gas analysisdeep neural networkDynamic Adamdropout |
| spellingShingle | Minghui Ou Hua Wei Yiyi Zhang Jiancheng Tan A Dynamic Adam Based Deep Neural Network for Fault Diagnosis of Oil-Immersed Power Transformers Energies power transformer fault diagnosis dissolved gas analysis deep neural network Dynamic Adam dropout |
| title | A Dynamic Adam Based Deep Neural Network for Fault Diagnosis of Oil-Immersed Power Transformers |
| title_full | A Dynamic Adam Based Deep Neural Network for Fault Diagnosis of Oil-Immersed Power Transformers |
| title_fullStr | A Dynamic Adam Based Deep Neural Network for Fault Diagnosis of Oil-Immersed Power Transformers |
| title_full_unstemmed | A Dynamic Adam Based Deep Neural Network for Fault Diagnosis of Oil-Immersed Power Transformers |
| title_short | A Dynamic Adam Based Deep Neural Network for Fault Diagnosis of Oil-Immersed Power Transformers |
| title_sort | dynamic adam based deep neural network for fault diagnosis of oil immersed power transformers |
| topic | power transformer fault diagnosis dissolved gas analysis deep neural network Dynamic Adam dropout |
| url | http://www.mdpi.com/1996-1073/12/6/995 |
| work_keys_str_mv | AT minghuiou adynamicadambaseddeepneuralnetworkforfaultdiagnosisofoilimmersedpowertransformers AT huawei adynamicadambaseddeepneuralnetworkforfaultdiagnosisofoilimmersedpowertransformers AT yiyizhang adynamicadambaseddeepneuralnetworkforfaultdiagnosisofoilimmersedpowertransformers AT jianchengtan adynamicadambaseddeepneuralnetworkforfaultdiagnosisofoilimmersedpowertransformers AT minghuiou dynamicadambaseddeepneuralnetworkforfaultdiagnosisofoilimmersedpowertransformers AT huawei dynamicadambaseddeepneuralnetworkforfaultdiagnosisofoilimmersedpowertransformers AT yiyizhang dynamicadambaseddeepneuralnetworkforfaultdiagnosisofoilimmersedpowertransformers AT jianchengtan dynamicadambaseddeepneuralnetworkforfaultdiagnosisofoilimmersedpowertransformers |