| 基于神经网络的间接矢量控制中转矩辨识算法的研究 |
| |
| 引用本文： | 刘成昊,马吉恩,方攸同,刘星,邱麟.基于神经网络的间接矢量控制中转矩辨识算法的研究[J].微电机,2021,0(3):9-13. |
| |
| 作者姓名： | 刘成昊 马吉恩 方攸同 刘星 邱麟 |
| |
| 作者单位： | (浙江大学电气工程学院,杭州310027) |
| |
| 摘 要： | 本文提出了一种基于神经网络的间接矢量控制系统中的转矩辨识算法，该算法将电机输入输出序列作为神经网络的输入，且神经网络的结果由分步计算实现；同时，该算法中训练数据集的标注采用加速度为核心的标定方法，并设计了数据集中多工况的转矩给定轨迹。相较于传统的神经网络方案，此方案在引入输入输出序列后，使网络对暂态过程的拟合能力得到提升；此外，采用所设计的多工况下的转矩轨迹和基于加速度的标定方法，不仅能够避免模型对特定工况的过拟合，而且能够在兼顾辨识精度的同时在标定过程中无需采用转矩感仪器。仿真和实验结果对比表明该方案精度高，鲁棒性强且适用于暂态过程，整个算法流程在工程上易于实现。
|
| 关 键 词： | 间接矢量控制系统 转矩辨识 神经网络 鲁棒性能 数据集制作 |
Research on Torque Identification Algorithm Based on Neural Network in Indirect Vector Control |
| |
| LIU Chenghao,MA Jien,FANG Youtong,LIU Xing,QIU Lin.Research on Torque Identification Algorithm Based on Neural Network in Indirect Vector Control[J].Micromotors,2021,0(3):9-13. |
| |
| Authors: | LIU Chenghao MA Jien FANG Youtong LIU Xing QIU Lin |
| |
| Affiliation: | (College of Electrical EngineeringZhejiang University, Hangzhou 310027China) |
| |
| Abstract: | For indirect vector control system,a torque identification algorithm based on neural network was proposed.The feature of the method was that the relevant variables of current time and previous time were taken as the input of neural network,and the result of neural network was realized by a two-step calculation.At the same time,the calibration method based on acceleration was used to label the training data set in the method,and the given trajectory of torque under multiple working conditions in the data set was designed.Compared with the traditional identification method,this algorithm improved the ability of neural network to fit the transient process after introducing the previous time variables.On the other hand,the designed torque trajectory avoided the over fitting of the model to the specific working condition.Also,the acceleration scheme made the calibration process without torque sensor.Simulation and experiment results show that the algorithm has high accuracy,robustness and is suitable for transient process.The whole process is easy to realize in engineering. |
| |
| Keywords: | indirect vector control system torque identification neural network robust performance data set generation |
| 本文献已被 CNKI 维普 等数据库收录！ |
| 点击此处可从《微电机》浏览原始摘要信息 |
|
点击此处可从《微电机》下载全文 |
|
