An inherently fault tolerant control of sodium cooled fast reactors and its stability analysis using particle swarm optimization

Abstract

In this study we propose a simple, yet inherently fault-tolerant and robust controller utilizing a combination of both feedforward and feedback control schemes. Feedback controllers are known to be robust, but feedforward controllers can be reliable in the context of safety as they can be less dependent on measurements. Though, many combined feedforward and feedback controllers have been proposed in the literature, to the best of our knowledge a controller tolerant to the failure of power feedback signal is nowhere demonstrated. In this scheme, major corrections are effected by a model predictive forward control unit, while bounded uncertain part is corrected by a feedback control unit. The feedforward control is implemented using the inverse equations of the nuclear reactor point kinetics model. The bounded error control is effected by a simpler limited Proportional–Integral–Derivative (PID) feedback controller. Stability analysis of the controller is carried out numerically by Lyapunov’s direct method with the help of Particle Swarm Optimization (PSO) technique. The proposed controller (referred to as Inverse Dynamics Corrected Control (IDCC)), is studied for reactor power control with model error plus uncertainty and a principal feedback signal failure case ( i.e. power sensor failure). It is shown that the IDCC has excellent load tracking performance for a challenging demand profile with model error, in comparison to Sliding Mode Controller (SMC) and a manually tuned PID controller. For the loss of principal feedback signal case, there is a trade off between the tracking performance of IDCC and safety aspects. While both PID and SMC drive the reactor power to unsafe levels, IDCC maintains the reactor power within a safer bound.