Context:

This paper presents the design of a Linear Quadratic Gaussian controller to regulate the voltage of a DC generator, and its performance is compared with that of a self-tuning regulator based on pole placement.

Method:

To implement these control strategies, a prototype DC generator is used, and the mathematical model of the system is obtained by means of non-parametric identification. The theoretical bases and the design of the two control strategies are presented, and the performance comparison of the two controllers is performed by using metrics from their integral error and time response.

Results:

Integral error and time response metrics show that, although the performance of the Linear Quadratic Gaussian controller is slightly higher than that presented by the self-tuning controller, system responses are stable with both control strategies, as well as with a good speed response and a steady-state error equal to zero.

Conclusions:

A Linear Quadratic Gaussian controller and self-tuned control strategies are a good alternative to conventional control and generate an adequate process control that is in line with system requirements.

        · abstract in Spanish     · text in Spanish     · Spanish ( pdf )