International Conference
| Multi-Channel Neural Networks-Based Thermal Monitoring of Electric Motor | |
|---|---|
| Year | 2023 |
| Month | August |
| Journal | 2023 IEEE International Symposium on Sensorless Control for Electrical Drives (SLED) |
| Author | Jaehoon Shim; Jonghun Choi; Sangwon Lee; Jung-Ik Ha |
| Link |
http://ieeexplore.ieee.org/abstract/document/10261386
28회 연결
|
This paper presents a thermal monitoring model for electric drive motors using Multi-Channel Neural Networks (MCNNs). The proposed structure consists of three key components. Firstly, a data-driven approach is adopted, eliminating the need for additional sensors, except during the training phase. Secondly, it employs a multi-channel architecture that processes two sets of inputs. One channel handles input power-related thermal modeling, while the other focuses on modeling thermal losses. Consequently, these channels are combined to create a data-driven thermal system. Thirdly, a suitable structure is proposed for the input data, tailored differently for each channel. Furthermore, the MCNNs are trained end-to-end using the provided loss function. The model's performance is assessed using data collected from an 89 kW induction motor. Additionally, comparative models based on previously studied machine learning-based thermal models are utilized for comparison. |
|
Seoul National University, 1, Gwanak-ro, Gwanak-gu,
Seoul, Republic of Korea, 08826 Bldg. 301, Rm. 606