Data-driven fault diagnosis of power converter systems
This project aims to develop a hardware platform which can operate a three-phase synchronous motor and simulate the open-circuit fault state of the IGBTs in the three-phase power inverter, acquire the time-varying current data of each three phase, and diagnosis the fault type by data-driven methods...
| Main Author: | |
|---|---|
| Other Authors: | |
| Format: | Thesis-Master by Coursework |
| Language: | English |
| Published: |
Nanyang Technological University
2023
|
| Subjects: | |
| Online Access: | https://hdl.handle.net/10356/167346 |
| _version_ | 1826113474110947328 |
|---|---|
| author | Su, Yuancheng |
| author2 | Xu Yan |
| author_facet | Xu Yan Su, Yuancheng |
| author_sort | Su, Yuancheng |
| collection | NTU |
| description | This project aims to develop a hardware platform which can operate a three-phase synchronous motor and simulate the open-circuit fault state of the IGBTs in the three-phase power inverter, acquire the time-varying current data of each three phase, and diagnosis the fault type by data-driven methods according to the acquired current data in real time. To realize those functions mentioned above, the TI controlCARD evaluation board with the hardware-in-the-loop is applied to control the operation of the motor, the equipment based on STM32 is designed to acquire data, and the raspberry pi is applied to process the acquired data and realize fault diagnosis. In the first part of this thesis, a detailed instruction to the operation of the built hardware platform is developed in order to help others who intend to do further study using this platform get a quick start. In the second part, the data-driven methods in open-circuit fault diagnosis of the three-phase pulse-width modulation (PWM) inverter are discussed, and the performances of random forest (RF), support vector machine (SVM), k-nearest neighborhood (KNN) and extreme learning machine (ELM) are tested in the condition of both simulation environment and the built hardware platform environment. After getting and analyzing the performances, further thoughts about how to improve the performance of the algorithm in real operation and how to improve the data quality of the built hardware platform are proposed in the last part. Among those, missing data problem is mainly focused. |
| first_indexed | 2024-10-01T03:23:48Z |
| format | Thesis-Master by Coursework |
| id | ntu-10356/167346 |
| institution | Nanyang Technological University |
| language | English |
| last_indexed | 2024-10-01T03:23:48Z |
| publishDate | 2023 |
| publisher | Nanyang Technological University |
| record_format | dspace |
| spelling | ntu-10356/1673462023-07-04T16:53:48Z Data-driven fault diagnosis of power converter systems Su, Yuancheng Xu Yan School of Electrical and Electronic Engineering xuyan@ntu.edu.sg Engineering::Electrical and electronic engineering This project aims to develop a hardware platform which can operate a three-phase synchronous motor and simulate the open-circuit fault state of the IGBTs in the three-phase power inverter, acquire the time-varying current data of each three phase, and diagnosis the fault type by data-driven methods according to the acquired current data in real time. To realize those functions mentioned above, the TI controlCARD evaluation board with the hardware-in-the-loop is applied to control the operation of the motor, the equipment based on STM32 is designed to acquire data, and the raspberry pi is applied to process the acquired data and realize fault diagnosis. In the first part of this thesis, a detailed instruction to the operation of the built hardware platform is developed in order to help others who intend to do further study using this platform get a quick start. In the second part, the data-driven methods in open-circuit fault diagnosis of the three-phase pulse-width modulation (PWM) inverter are discussed, and the performances of random forest (RF), support vector machine (SVM), k-nearest neighborhood (KNN) and extreme learning machine (ELM) are tested in the condition of both simulation environment and the built hardware platform environment. After getting and analyzing the performances, further thoughts about how to improve the performance of the algorithm in real operation and how to improve the data quality of the built hardware platform are proposed in the last part. Among those, missing data problem is mainly focused. Master of Science (Power Engineering) 2023-05-15T08:05:33Z 2023-05-15T08:05:33Z 2023 Thesis-Master by Coursework Su, Y. (2023). Data-driven fault diagnosis of power converter systems. Master's thesis, Nanyang Technological University, Singapore. https://hdl.handle.net/10356/167346 https://hdl.handle.net/10356/167346 en application/pdf Nanyang Technological University |
| spellingShingle | Engineering::Electrical and electronic engineering Su, Yuancheng Data-driven fault diagnosis of power converter systems |
| title | Data-driven fault diagnosis of power converter systems |
| title_full | Data-driven fault diagnosis of power converter systems |
| title_fullStr | Data-driven fault diagnosis of power converter systems |
| title_full_unstemmed | Data-driven fault diagnosis of power converter systems |
| title_short | Data-driven fault diagnosis of power converter systems |
| title_sort | data driven fault diagnosis of power converter systems |
| topic | Engineering::Electrical and electronic engineering |
| url | https://hdl.handle.net/10356/167346 |
| work_keys_str_mv | AT suyuancheng datadrivenfaultdiagnosisofpowerconvertersystems |