Four actions will be undertaken in this topic:

1. Optimal design machine

2. Development of models adapted to the control

Work on the identification of the parameters of the asynchronous machine (Kalman filter) was initiated with the LEC, by Luc LORON from the very start of the 1990's.

The originality of this work relies upon a will to establish the tool for real time identification in embedded industrial applications.Thus the efforts were not only related to the conditioning of the signals, but also to the adjustment of the Kalman filter and on an adequate modeling of the actuator. A model of the second order was developed in particular to allow a use of the Kalman filter over long periods of sampling and thus reducing the necessary computing power. This competence, taken over again by Christophe FORGEZ is maintained within the LEC, as well through the identification of the parameters of the machines, as of the estimate of certain operation parameters (speed or temperature). Studies were in particular undertaken for VALEO within the framework of the thermal monitoring of an asynchronous integrated alterno-starter. Within the framework of this specific application, the work initiated by Luc LORON was reinforced by a coupling of knowledge ' a priori' with the Kalman filter in order to stage the inaccuracy of embedded sensors. Such real time identification techniques are planned for the determination of some parameters of electrochemical storage devices in order to determine their state of load or health.

4.Analysis of acoustic noises generated during electric machine - electronic power converter association