Skip to main content
placeholder image

Three-Phase Inverter Faults Diagnosis Using Unsupervised Sparse Auto-Encoder

Conference Paper


Abstract


  • Fault detection and classification is vital for ensuring the safety of power electronic converters employed in the power electronic device. In order to prevent the propagation of faults to the other electronic components, the accurate detection and classification of faults is a must considerable thing. Although modern fault detection methods can perform this task accurately, their accuracy limits to a number of a specific type of fault. Considering the unpredictable nature of the faults, this paper applies a sparse auto-encoder (SAE) to detect and classify the faults. In contrary to the conventional methods, this proposed method can extract the features automatically from the image representation of the signals which increases the generalizability of the proposed method. The results show in this paper confirms the reliability of the methods in performance.

Publication Date


  • 2020

Citation


  • Fahim, S. R., Sarker, S. K., Das, S. K., Islam, M. R., Kouzani, A. Z., & Parvez Mahmud, M. A. (2020). Three-Phase Inverter Faults Diagnosis Using Unsupervised Sparse Auto-Encoder. In 2020 IEEE International Conference on Applied Superconductivity and Electromagnetic Devices, ASEMD 2020. doi:10.1109/ASEMD49065.2020.9276267

Scopus Eid


  • 2-s2.0-85099302831

Web Of Science Accession Number


Abstract


  • Fault detection and classification is vital for ensuring the safety of power electronic converters employed in the power electronic device. In order to prevent the propagation of faults to the other electronic components, the accurate detection and classification of faults is a must considerable thing. Although modern fault detection methods can perform this task accurately, their accuracy limits to a number of a specific type of fault. Considering the unpredictable nature of the faults, this paper applies a sparse auto-encoder (SAE) to detect and classify the faults. In contrary to the conventional methods, this proposed method can extract the features automatically from the image representation of the signals which increases the generalizability of the proposed method. The results show in this paper confirms the reliability of the methods in performance.

Publication Date


  • 2020

Citation


  • Fahim, S. R., Sarker, S. K., Das, S. K., Islam, M. R., Kouzani, A. Z., & Parvez Mahmud, M. A. (2020). Three-Phase Inverter Faults Diagnosis Using Unsupervised Sparse Auto-Encoder. In 2020 IEEE International Conference on Applied Superconductivity and Electromagnetic Devices, ASEMD 2020. doi:10.1109/ASEMD49065.2020.9276267

Scopus Eid


  • 2-s2.0-85099302831

Web Of Science Accession Number