Multimodel generalized predictive control of a heat-pipe reactor coupled with an open-air Brayton cycle

Abstract

The heat-pipe reactor is based on the design concept of a solid-state reactor, and has gradually become a preferred reactor type for future space nuclear power and new nuclear power technology applications. However, due to the all-solid-state assembly mode, dynamic and control characteristics of the heat-pipe reactor are different from those of traditional reactors. In this regard, the paper proposed a multimodel generalized predictive control (MMGPC) method for a heat-pipe reactor coupled with an open-air Brayton cycle. First, based on the dynamic characteristics, the control strategy of the heat-pipe reactor system was formulated. Then, decoupling controllers and generalized predictive controllers of the reactor power and rotating shaft speed were designed under the control strategy, and all the dependent models in the controllers were established with the multimodel modeling principle to construct the MMGPC system. Finally, to verify the performance of the control system, step load-change transients of the heat-pipe reactor system were simulated. The results demonstrates that the MMGPC method is applicable to and has excellent control performances for the heat-pipe reactor coupled with an open-air Brayton cycle.