A dynamic model for the advanced neutron source reactor including the xenon oscillation and the utilization of optimal control as a diagnostic tool

The research performed in this Ph.D. dissertation has been in support of the Advanced Neutron Source Reactor (ANS), a national experimental facility to be built at the Oak Ridge National Laboratory (ORNL). The goal of this work was to calculate kinetic parameters and develop algorithms for the setting-up of an appropriate control architecture for the ANS reactor.

One-and two-dimensional static calculations were performed using the BOLD VENTURE computation system to obtain the global spatial dependence of the neutron flux throughout the reactor core and reflector. These calculations were used to determine the amount of built-in reactivity and control required to compensate for the excess reactivity contained in the initial fuel loading, to allow for flexible and safe reactor operation and to calculate reactor kinetic parameters (neutron lifetime and the effective delayed neutron fraction).

Regarding the development of support algorithms to be integrated in the control architecture of the ANS reactor, three issues appeared to be of crucial interest: (i) start-up and approach to criticality, (ii) xenon and samarium poisoning, and (iii) diagnostic tools for the unmeasurable quantities of the ANS reactor.

A two-Point-two-group kinetics model was developed to study the startup and criticality issues, which can be implemented as an on - line computer algorithm. This development was needed since IV the disparity of neutron spectrum in the reflector and core makes the usual one point reactor kinetics approach a doubtful one, especially for severe reactor transients.

The xenon poisoning is a process of crucial relevance in view of the high neutron flux present in the ANS reactor. This process affects both the control algorithm, and more importantly the ability to restart the reactor after unanticipated shutdowns. It was found that the presence of samarium poisoning will stabilize the xenon oscillations, a certainly welcome result. Additionally, extensive calculations were performed of the after-shutdown reactivity due to xenon and samarium build-ups.

Diagnostics tools were developed, using optimal control techniques, to estimate the average thickness of aluminum oxide on the fuel plates, the average temperature of the fuel plates, which is not a measured state variable in the ANS reactor, and the coolant flow.