A fault monitoring approach using model-based and neural network techniques applied to input–output feedback linearization control induction motor

被引:0
|
作者
Imadeddine Harzelli
Arezki Menacer
Tarek Ameid
机构
[1] Biskra University,LGEB Laboratory, Electrical Engineering Department
来源
Journal of Ambient Intelligence and Humanized Computing | 2020年 / 11卷
关键词
Induction motor (IM); Input–output feedback linearization (IOFL) control; Fault monitoring; Residual speed; Neural network (NN); Hilbert transform (HT);
D O I
暂无
中图分类号
学科分类号
摘要
This paper presents a contribution to the fault monitoring approach and input–output feedback linearization control of the induction motor (IM) in the closed-loop drive. Two kinds of faults are considered in the machine, particularly the broken rotor bars and stator inter-turn short circuit faults. This approach has been employed to detect and identify simple and mixed defects during motor operation by utilizing advanced techniques. To achieve it, two procedures are applied for the fault monitoring: The model-based strategy, which used to generate a residual speed signal to indicate the presence of possible failures, by means the high gain observer in the closed-loop drive. However, this strategy is not able to recognise the type of faults but it can be affected by the disturbances. Therefore, the neural network (NN) technique is applied in order to identify the faults and distinguish them. However, the NN required a relevant database to achieve satisfactory results. Hence, the stator current analysis based on the HFFT combination of the Hilbert transform and fast Fourier transform is applied to extract the amplitude of the harmonics and used them as an input data set for NN. The obtained results show the efficiency of the fault monitoring system and its ability to detect and diagnosis any minor faults in a closed loop of the IM.
引用
收藏
页码:2519 / 2538
页数:19
相关论文
empty
未找到相关数据