Commande numérique des machines

The articleDigital control of converter-machine assemblies is the subject of three articles:

• [D 3 640] Evolution of orders

• DC motor

• Three-phase system

and the subjects covered are not independent of each other. Readers will often need to refer to other articles.

The control of electrical machines has undergone two major revolutions: the rise in power electronics components and the development of programmed digital devices, microprocessors and microcontrollers. The digital control of machines promises considerable progress in the quality of the performance obtained, while at the same time posing particular problems requiring specific solutions. This article attempts to give an overview of these problems, based on a few typical examples.

The problems posed by analog technology (drift, long-term stability, difficulties in realizing functions such as trigonometric lines) prompted the switch to digital technology in the 1970s. The advent of powerful microprocessors in the 1980s led to the emergence of a new family of controls, vector control, giving AC machines dynamic performance comparable to that of DC machines. These techniques have also enabled the implementation of optimized controls (in terms of harmonic elimination) for pulse-width modulation inverters. Generally speaking, digital technology allows the practical implementation of algorithms and functions that are far more complex than those tolerated by analog technologies.

While digital technology has enabled the development of intrinsic performance, it has also brought decisive industrial benefits. The same digital device can handle relations with the human operator, take charge of stop-start sequences and self-diagnosis; all these complex functions practically require the use of digital devices. The same microprocessor, with the same hardware architecture, can be used for all kinds of applications, simply by changing the memory content. Last but not least, thanks to advances in integrated circuitry, digital controls enable highly complex designs in remarkably compact volumes.

However, digital control has a number of requirements: the establishment of control algorithms usually requires a good knowledge of the process to be controlled and its mathematical model (transfer function, equation of state), a good knowledge of automatic control methods (continuous control, sampled control, state feedback, optimal control, robustness) and signal processing (to make good use of signals supplied by sensors). The signals have high frequencies, and there is a certain sensitivity to interference transmitted or radiated...