De-noising of spectral contents for diagnosis purpose using morphological filter

Hubert Razik (1) , Jose Antonino Daviu (2)

1 MIS - Ampère, Département Méthodes pour l'Ingénierie des Systèmes
2 UPV - Universitat Politècnica de València = Universitad Politecnica de Valencia = Polytechnic University of Valencia

This paper presents the use of a novel method dealing with mathematical morphology to de-noise spectral contents. This approach is based on two fundamental operators which are dilation and erosion. This type of filter does not need a huge computing time because the operations involved are simple. The goal of this filter in the application considered in this paper is to de-noise the spectrum of the stator current demanded by an induction motor in order to facilitate the self-extraction of occasional fault components. Furthermore, the induction motors are widely used under speed variation conditions, usually fed by an inverter. Experiments showed in this paper are convincing and suggest the possible extension of the method to a wide range of applications.

Mots clés: Broken bar fault; induction motor; morphological filter

Domaines: Energie électrique

Hubert Razik, Jose Antonino Daviu. De-noising of spectral contents for diagnosis purpose using morphological filter. 43rd IEEE IECON, Oct 2017, Pékin, China. ⟨10.1109/IECON.2017.8216650⟩. ⟨hal-01899867⟩