Scholarlink

<p>The nonlinear and complex dynamics of nuclear reactors pose significant challenges in designing simple yet effective controllers for power level regulation, particularly during load-following operations, where the reactor must adjust its output to match varying power demands while maintaining safety and stability. This study presents a novel discrete adaptive learning-based active disturbance rejection control (ADRC) law, specifically developed for such operations and suitable for digital implementation on embedded controllers. The proposed framework integrates an adaptive learning extended state observer (ESO) that dynamically tunes its estimation gains to improve the accuracy of system state and lumped and input disturbance estimations in real time. Based on these enhanced ESO estimates, a backstepping-based discrete control law is constructed to ensure precise tracking and disturbance compensation. Stability of the closed-loop system is rigorously verified using Lyapunov theory. Simulation results show that the proposed controller achieves up to 74% improvement in mean absolute tracking error (MATE), 88% reduction in average absolute tracking error (AATE), and over 90% reduction in steady-state tracking error (SDTE) compared to conventional discrete PID (DPID) and discrete sliding mode control (DSMC) schemes, demonstrating faster response, zero steady-state error, and superior disturbance rejection.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Discrete Adaptive Learning Active Disturbance Rejection Control Law Designed for Nuclear Reactor Under Load Following Operation

  • 期刊论文

  • 发表日期：

    2026年2月25日

查看全文
  • Boubacar Kirgni Hamza,
  • Wang Junling

摘要

The nonlinear and complex dynamics of nuclear reactors pose significant challenges in designing simple yet effective controllers for power level regulation, particularly during load-following operations, where the reactor must adjust its output to match varying power demands while maintaining safety and stability. This study presents a novel discrete adaptive learning-based active disturbance rejection control (ADRC) law, specifically developed for such operations and suitable for digital implementation on embedded controllers. The proposed framework integrates an adaptive learning extended state observer (ESO) that dynamically tunes its estimation gains to improve the accuracy of system state and lumped and input disturbance estimations in real time. Based on these enhanced ESO estimates, a backstepping-based discrete control law is constructed to ensure precise tracking and disturbance compensation. Stability of the closed-loop system is rigorously verified using Lyapunov theory. Simulation results show that the proposed controller achieves up to 74% improvement in mean absolute tracking error (MATE), 88% reduction in average absolute tracking error (AATE), and over 90% reduction in steady-state tracking error (SDTE) compared to conventional discrete PID (DPID) and discrete sliding mode control (DSMC) schemes, demonstrating faster response, zero steady-state error, and superior disturbance rejection.