Reactor power control performance improvement based on fuzzy neural network of a small pressurized water reactor
Juan Liu,Ge Gao,Ziqi Fan,Changhao Wu,Peiwei Sun,Xinyu Wei
DOI: https://doi.org/10.1016/j.nucengdes.2024.113730
IF: 1.7
2024-12-06
Nuclear Engineering and Design
Abstract:The modular design of small reactors has potential application prospects in nuclear power plants such as surface ships and land-based applications, and requires higher safety and flexibility. The ship nuclear power plant has the requirement of maneuvering and fast operation. As the power source, the small pressurized water reactor adopts the traditional control technology represented by PID (proportional-integral–differential), which cannot effectively meet the requirements of different operating conditions for control performance, especially the overshoot of reactor power is too large. Therefore, advanced control algorithms are needed to improve the control performance. In order to improve the learning and adaptive ability of fuzzy control, this paper takes a small PWR with 150 MW as the research object and proposes a reactor power control system based on fuzzy neural network. The simulation test results show that compared with fuzzy control and traditional control, the fuzzy neural network control has better learning and adaptive ability in the transient operation process, can adapt to various working conditions and effectively reduce the overshoot of reactor power. The overshoot of reactor power can be reduced by more than 15 % compared with the traditional control, and can be reduced by more than 8 % compared with the fuzzy control, and the control rod travel is also significantly reduced.
nuclear science & technology