Adaptive backstepping controller with extended state observer for load following of nuclear power plant

Abstract

Load following is one of the most important subjects for nuclear power plants. In this paper, an adaptive backstepping control strategy with extended state observer is proposed for load following of nuclear power plant. The model of nuclear reactor based on point-reactor kinetics equations with one delayed neutron groups is described firstly. Subsequently, considering the unmeasurable limitation of densities of delayed neutron precursor and average fuel temperature, an extended state observer (ESO) is designed to estimate unmeasured states of system and output disturbances online, and the stability of the observer error dynamic is analyzed. Then, an adaptive backstepping load following control approach is proposed, based on radial basis function neural network (RBFNN) for nuclear reactor in the presence of reactivity disturbances, where the system uncertainties including unknown reactivity disturbances caused by xenon and samarium poisoning, power defect etc. are approximated by using RBFNN. In addition, it is demonstrated though Lyapunov stability theory that the overall control system is ultimately uniformly bounded (UUB). Finally, simulation results are given to verify the effectiveness of the proposed controller compared with conventional sliding mode controller in terms of load following performance, namely, better following accuracy and stronger ability against disturbances.