Time optimal digital computer control of nuclear reactor, (3). Experiment

Time optimal control has been applied to a digital computer control system governing a research reactor (Toshiba Training Reactor). A program was developed for the control computer, with which actual experiments were performed. The results demonstrated the applicability of the various modern control theories for designing such systems. Discussion is first presented on some considerations required in realizing the results envisaged on the basis of a simplified model of the system. In view of the use of reactivity as variable representing the state of reactor system, a method has been developed that permits simple and accurate on-line calculation of the reactivity from changes in reactor power. The control is determined in the whole phase plane (neutron density vs. reactivity), and is programmed with the use of appropriate simplifications and approximations. Experiments were first performed on a simulated system of the reactor and then on the actual reactor system. The results described cover experiments on steady state performance as well as transient performance with the reactor power varied in ranges extending to 2 decades. (author)