Objectives

Fault detection and diagnosis play an important industrial role. Indeed, they contribute, by early detection and early to earn points and availability of production capital invested in production facilities. The renewed interest shown by the different industrial sectors and the world of Research shows that the area of fault detection and diagnosis is a very lucrative niche. The work developed within this theme concerned the initial detection and diagnosis of faults in electric motors (mainly asynchronous cage), then gradually extended to the converter associated (mainly a UPS). The choice of induction motor cage is justified by the fact that it is heavily used in industry due to its robustness and flexibility.

The main topics covered are:

• Diagnosis of bearing defects by Concordia transforms special.
• Diagnosis of defects in the arms of inverters by hybrid methods (transform Concordia neural networks and fuzzy logic).
• Fault diagnosis: Eccentricity and bearings.

The work was gradually extended to more complex systems especially in renewable energy and context Brest (offshore wind farm projects under the Pôle Mer Bretagne). Indeed, we have initiated work theses whose main objective is the monitoring and diagnosis of failures in wind energy systems (particularly offshore). Indeed, from the fact that large wind turbines are very difficult to access (at sea or extremely high towers exceeding the height of 20 m), there is a strong need to reduce maintenance costs and operating wind. This will remove certain difficulties due to problems of accessibility and will detect an early, mechanical defects and / or providing electrical failures of different components (electrical and mechanical) of the wind, facilitating a proactive response , minimizing sudden stops caused by breakdowns and consequently maximizing productivity. The most effective method for reducing these costs is to continuously monitor a state of the wind and especially its electric generator for use in diagnosis.

Our first results in this framework to begin recovery in international journals. The main results obtained are mainly related to the use of advanced signal processing power from the double-fed induction generator wind turbine. In addition, collaboration is being institutionalized with AREVA T & A Cadarache on Diagnosis of Defects in Nuclear Power Plant Asynchronous Pumps

Fig 1. Spectrogram of the stator current in the presence of a defect in turnover	Fig 2. Scalogram of the stator current in the presence of a defect in turnover	Fig 3. Wigner Ville stator current in the presence of a defect in turnover