Search

The main problems that appear when using a long cable connection between a Variable Frequency Drive (VFD) and the AC motor are over-voltages and bearing currents. The last one concerns especially EMC, while the over-voltages transients can be fatal to the motor in long motor duties. Using LF/HF models the paper investigates the behaviour of this motor and drive system concerning over-voltages. As a result of the theoretical study of over-voltages problem, a PWM algorithm that decreases the amplitude of the over-voltages transients is proposed. It is based on an online identification of the cable. Experimental results are shown.

Nowadays, controlling the position of alternating current machines in a high performant way is necessary in applications that need a high precision mobility. In this aim, many variable frequency AC drives are constituted by a three-phase synchronous motor fed by a static converter. As a matter of fact, because of its dynamic characteristics, the synchronous motor is usually used in view of position tracking performance.

The first China International Conference on Electrical Machines was held in Wuhan on September 18-20, 1991. The conference was supported by China Association for Science and Technology and ICEM (Europe), and was sponsored by CSEE, CES and IEE (London). A total of 165 participants attended the conference, coming from relevant academic, research and educational institutions as well as R D and engineering faculties in the power and manufacturing industries in China and abroad. The conference proceedings contained 186 papers contributed by authors of 18 different countries.; International audience; Controlling the acceleration of a synchronous motor drive is necessary in applications such as passengers transport systems or high precision cutting machines. Within this context a digital cascade control structure including an acceleration loop is studied in view of acceleration, speed, and position tracking performance. In order to compensate the load torque, a deterministic observer is developed, and the load torque estimate is then used to adapt the control law. The efficiency of the overall control law is investigated by simulation yielding system characteristics that do no more depend on the load torque.

Le contrôle d'un actionneur électromécanique constitué d'une machine synchrone associée à un onduleur de tension à modulation de largeur d'impulsion (MLI) par des régulateurs en cascade, possède des caractéristiques dépendantes de la charge entraînée. On peut améliorer la réponse dynamique du système, en présence d'un couple de charge constant, par l'utilisation des techniques de réglage d'état digital. Cependant en présence d'un couple variable (variations du frottement et de l'inertie) cette technique s'avère également insuffisante. Pour rendre le fonctionnement indépendant de ces variations, il est alors nécessaire de les compenser dans le système de commande. Dans ce but nous avons étudié une structure de contrôle d'état numérique, où la loi de commande présentée intègre un observateur du couple résistant associé à un retour d'état. Ainsi les performances dynamiques de l'ensemble restent indépendantes de la charge entraînée.

The speed control of a self controlled Pulse Width Modulated converter fed synchronous motor drive, is usually done by an Integral-Proportional regulator. A digital simulation process shows that this of control guarantees, only for a particular speed range, a good dynamic behaviour. To overcome this disadvantage, the authors propose the implementation of a system which permits a real time adaptation of the regular parameters to the operational point. For this purpose, the identification of the transfer function is introduced. After a dynamic analysis of the system, this transfer function is choosen in such a way that the dynamic behaviour of the system is preserved. The resulted second order transfer function describes in a satisfactory way the behaviour of the system. The recursive least squares algorithm permits to calculate the different coefficients of the transfer function. This function which represents the system is not stationary, a "forgot factor" is added to the algorithm which privileged the new measurements with respect to the old one. in the second part of this paper, this identification is used to define an optimal regulator. Some simulations results show an adaptive character of the regulator with respect to operational points are presented.

The introduction of a micro-processor into a command loop permits us to envisage new control structures. In line with this, we propose to set up modern adjusters to control the electro-magnetic torque in the synchronous machine fed by a PMW inverter. The first technique, called STATE SPACE FEEDBACK, consists in decomposing the dynamics of the procedure into two time scales, and in giving a reduced-order action model. This reduced-state model serves as a base for the construction of a dynamic flux observer. We next develop a state space feedback allowing control of the transitory evolution of state variables in a speed variator. The second technique, called ADAPTIVE CONTROL, is concerned with setting-up a self-tuning regulator ensuring adaptation of the regulator's parameters as the function point. This technique requires a line identification procedure which permits knowledge of the different conefficients of the equivalent-transfer function.

A) Synthèse de la commande algorithmique Les variateurs de vitesse ou de position constitués de machines électriques à courant alternatif associées à des convertisseurs statiques nécessitent la mise au point de commandes et régulations ayant des performances de plus en plus élevées pour satisfaire les cahiers de charge. La mise au point de ces commandes demande la résolution d'un certain nombre de problèmes d'automatique, d'informatique industrielle et de traitement des signaux qui vont être précisés ci-dessous. On en verra des réalisations concrètes sur la machine asynchrone et la machine synchrone dans la suite de ce papier. On peut, d'une manière non exhaustive, énumérer ci-dessous les principales étapes de la mise au point de la commande : I - détermination d'un modèle de connaissance du processus,II - identification paramétrique,III - structures de régulation,IV - mesures ou observation et traitement des signaux nécessaires comme entrées de la commande, V - implémentation des algorithmes sur microprocesseur pour une commande en temps réel.

The introduction of microprocessors in a control loop allows the implementation of new control structures. Based on this idea, we propose modern regulation techniques for the control of the electromagnetic torque of a synchronous motor. From the state-space representation of systems, we propose the decomposition of the process dynamic into two time scales, in order to give a reduced-order action model of the device. This reduced state-model will be the basis for the construction of a flux observer, with a structure being simple enough to be implemented, in real time, in the control loop of the motor. At last, we develop a modal control system, which allows the control of the transient-state evolution of state variables in the speed drive.