A comparison between model reduction and controller reduction: application to a PWR nuclear plant

摘要

The most classical way of obtaining a low-order model-based controller for a high-order system is to apply reduction techniques to an accurate high-order model or controller of the plant. An alternative is to use identification for control with alow-order model set. In [1], we compared model reduction, both in open loop and in closed loop, with closed-loop identification of a low-order model. Both methods were applied to the design of a controller for the secondary circuit of a nuclearPressurized Water Reactor, leading to the conclusions that system identification is a viable alternative to model reduction, and that the key feature for a successful control design is not so much the choice between order reduction or identification, butbetween open-loop and closed-loop techniques. In the present paper, we go further in this study and we compare model reduction with controller reduction for the same PWR system. We show that closed-loop techniques are more powerful than open-loop ones,and that controller reduction gives better results than model reduction when appropriate frequency weightings are chosen.