Artificial neural networks based fault detection in 3-Phase PMSM traction motor - Archive ouverte HAL
Communication Dans Un Congrès Année : 2012

Artificial neural networks based fault detection in 3-Phase PMSM traction motor

Résumé

Traction Motors Condition Monitoring is one of the important factors in increasing motor life time and prevention of any vehicle sudden stop in its track and thereupon avoiding of risking the safety of drivers or passengers. In this paper, a neural network based method for detecting unbalanced voltage fault which is one of the various faults in 3-phase traction motors was surveyed. Proposed method is independent from load state and fault percentage, which means neural network, is able to detect fault and load condition without any assumption about the state of the load and fault. In proposed method, two MLP (Multi Layer Perceptron) separate neural networks are used for solving of each problem. Experimental acquired data is used to train neural networks. Based on first test results, for detecting of unbalanced voltage fault percentage and also based on second test results for detecting of load condition accurately, the neural network could detect close to 100% of the tested cases.
Fichier non déposé

Dates et versions

hal-03223590 , version 1 (11-05-2021)

Identifiants

  • HAL Id : hal-03223590 , version 1

Citer

Seyed Saeid Moosavi, Abdesslem Djerdir, Youcef Ait Amirat, Davod Arab Khaburi. Artificial neural networks based fault detection in 3-Phase PMSM traction motor. International Conference on Electrical Machines, Sep 2012, Marseille, France. ⟨hal-03223590⟩
20 Consultations
0 Téléchargements

Partager

More