Global physical and numerical stability of a nuclear reactor core

Low order models are used to investigate the influence of integration methods on observed power oscillations of some nuclear reactor simulators. The zero-power point reactor kinetics with six-delayed neutron precursor groups are time discretized using explicit, implicit and Crank–Nicholson methods, and the stability limit of the time mesh spacing is exactly obtained by locating their characteristic poles in the z-transform plane. These poles are the s to z mappings of the inhour equation roots and, except for one of them, they show little or no dependence on the integration method. Conditions for stable power oscillations can be also obtained by tracking when steady state output signals resulting from reactivity oscillations in the s-Laplace plane cross the imaginary axis. The dynamics of a BWR core operating at power conditions is represented by a reduced order model obtained by adding three ordinary differential equations, which can model void and Doppler reactivity feedback effects on power, and collapsing all delayed neutron precursors in one group. Void dynamics are modeled as a second order system and fuel heat transfer as a first order system. This model shows rich characteristics in terms of indicating the relative importance of different core parameters and conditions on both numerical and physical oscillations observed by large computer code simulations. A brief discussion of the influence of actual core and coolant conditions on the reduced order model is presented.