Skip to Main content Skip to Navigation
Theses

Contribution à l'analyse de défauts de roulement induits par les défauts électriques de machines tournantes

Abstract : Premature bearing aging is one of the main reasons of forced outages of electrical machines, especially under converter operation. This work contributes to the analysis and the detection of the generalized roughness bearing fault in small machines (up to 20 kW) under converter operation. In this configuration, this fault is caused by capacitive discharges creating an electroerosion of the raceways. To detect this fault, we proposed a realistic aging process that reproduces the emerging state of the generalized roughness bearing fault. This experimental process consists of an inverter located between the raceways and designed to increase the discharge activity while limiting the energy of a single discharge to a realistic threshold. Operating conditions, which accelerate the bearing aging by increasing the discharge activity, are experimentally determined. In this thesis, we also tackle the problem of detecting the generalized roughness bearing fault. In accordance with the maintenance constraints and the effects produced by the fault on vibration signals, we selected appropriate signal processing tools. The creation of synthetic vibration signals, enriched with some properties of the fault, allowed a comparative study of the proposed detection methods, which are based on the Wiener filter, the spectral subtraction and the spectral kurtosis.
Complete list of metadatas

https://hal.univ-lorraine.fr/tel-02528936
Contributor : Thèses Ul <>
Submitted on : Thursday, April 2, 2020 - 9:06:45 AM
Last modification on : Monday, May 25, 2020 - 8:43:14 PM

File

DDOC_T_2019_0187_PLAZENET.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : tel-02528936, version 1

Collections

Citation

Thibaud Plazenet. Contribution à l'analyse de défauts de roulement induits par les défauts électriques de machines tournantes. Energie électrique. Université de Lorraine, 2019. Français. ⟨NNT : 2019LORR0187⟩. ⟨tel-02528936⟩

Share

Metrics

Record views

166

Files downloads

530