Multi-objective optimization of thermal power and outlet steam temperature for a nuclear steam supply system with deep reinforcement learning

Tianhao Zhang, Zhe Dong and Xiaojin Huang

Energy, 2024, vol. 286, issue C

Abstract: The nuclear steam supply system (NSSS) is a critical component of a nuclear power plant that produces steam for electricity or cogeneration. However, the performance of current control strategies in the NSSS is limited since the coupling in the system is too complex and non-linear to be modeled precisely. Thus, there is an urgent need to study the model-free optimization method of NSSS. Motivated by this, this article proposes a novel multi-objective optimization approach based on deep reinforcement learning (DRL). With a hierarchical structure, the proposed method improves the response performances of both thermal power and outlet steam temperature by dynamically adjusting the reference values of existing proportional–integral–differential (PID) controllers within NSSS. This structure combines the stability of PID’s closed-loop control and the optimization capabilities of DRL for a safe and efficient operation. Moreover, a safe DRL method with an event-triggered mechanism is proposed to further ensure safety throughout the optimization process. The numerical simulations demonstrate the effectiveness and superiority of the proposed method, which results in significant improvements in the transient response compared to traditional PID controllers.

Keywords: Nuclear power plant; Coordinated control; Multi-objective optimization; Safe deep reinforcement learning (search for similar items in EconPapers)
Date: 2024
References: View references in EconPapers View complete reference list from CitEc
Citations: View citations in EconPapers (2)

Downloads: (external link)
http://www.sciencedirect.com/science/article/pii/S0360544223029201
Full text for ScienceDirect subscribers only

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:eee:energy:v:286:y:2024:i:c:s0360544223029201

DOI: 10.1016/j.energy.2023.129526

Access Statistics for this article

Energy is currently edited by Henrik Lund and Mark J. Kaiser

More articles in Energy from Elsevier
Bibliographic data for series maintained by Catherine Liu ().