The control of nuclear reactors is one of the most challenging requirements in performing the load following operation mode; however, most of the previous studies either do not consider the adverse effects of the lumped uncertainty or rely on the assumption of possessing healthy actuators without any faults. Besides, the recently published backstepping load following controllers may result in the problem of “explosion of complexity” due to repeatedly differentiating the virtual control laws. To this end, this paper provides an adaptive active fault-tolerant dynamic surface control strategy for the load following problem of a modular high-temperature gas-cooled reactor suffering from lumped uncertainty and actuator faults. The main contributions of this paper are threefold. (i) a radial basis function neural network is introduced to real-timely learn the lumped uncertainty that contains model uncertainties, parameter variations, external disturbances, and unmeasured states; (ii) a dynamic surface control technique is employed to cope with the problem of “explosion of complexity” induced by traditional backstepping approach; and (iii) an active fault-tolerant controller, incorporated with a fault detection observer and a fault estimator, is developed such that the reliability and safety of the modular high-temperature gas-cooled reactor system could be enhanced. With Lyapunov stability theory, both the stability of the closed-loop system and the error convergence are proved. At the end, comparative simulation studies and analysis using quantitative performance indices indicate the proposed control strategy’s superiority in dealing with lumped uncertainty and actuator faults.
Read full abstract