Research on Optimization of Core Power Regulation System of Swimming Pool Reactor Based on PSO-BP Neural Network
-
摘要: 基于MATLAB/Simulink平台构建49-2游泳池式反应堆堆芯功率调节系统和一回路传热系统的仿真模型，开展外界反应性扰动仿真试验验证模型的准确性。采用粒子群算法（PSO）与反向传播（BP）神经网络相结合的比例-积分-微分（PID）控制器作为主控制器，模拟堆芯反应性和堆芯进口温度扰动下调节系统的响应情况，与游泳池式反应堆原控制器和传统BP神经网络控制器的响应情况相比较。结果表明，外界存在扰动时，基于PSO-BP神经网络的PID控制器可以使堆芯迅速达到稳定状态，调节时间更短、超调量更小，具有更好的鲁棒性和稳定性。
-
关键词:
- 粒子群算法（PSO） /
- 反向传播（BP）神经网络 /
- 比例-积分-微分（PID）控制器 /
- 反应堆功率调节
Abstract: Based on the MATLAB/Simulink, the simulation model of the power regulation system and the primary heat transfer system of the 49-2 swimming pool reactor was constructed, and the external reactive disturbance simulation test was carried out to verify the accuracy of the model. The proportion integration differentiation (PID) controller combined with particle swarm optimization (PSO) and BP neural network was used as the main controller, and the response of the regulating system under core reactivity and core inlet temperature disturbance was simulated, which was compared with that of the original controller of swimming pool reactor and the traditional BP neural network controller. The results show that the PID controller based on PSO-BP neural network can make the core reach a stable state quickly, with shorter regulating time and smaller overshoot, and has better robustness and stability. -
图 1 游泳池式堆功率调节系统工作原理示意图
Figure 1. Working Principle of Swimming Pool Reactor Power Regulation System
图 2 基于PSO -BP神经网络的PID控制器
r—目标值；y—系统输出值
Figure 2. PID Controller Based on PSO-BP Neural Network
图 6 30pcm反应性阶跃扰动下堆芯归一化功率变化
Figure 6. Change of Core Normalized Power under 30pcm Step Disturbance of Reactivity
图 7 30pcm反应性阶跃扰动下PSO-BP算法与传统BP神经网络算法性能指标函数值对比
Figure 7. Contrast of E(k) between PSO-BP Algorithm and Traditional BP Neural Network under 30pcm Step Disturbance of Reactivity
图 8 30pcm反应性阶跃扰动下PSO-BP神经网络控制器的$ {K}_{\mathrm{P}} $、$ {K}_{\mathrm{I}} $、$ {K}_{\mathrm{D}} $的变化曲线
Figure 8. $ {K}_{\mathrm{P}} $, $ {K}_{\mathrm{I}} $ and $ {K}_{\mathrm{D}} $ Curves of PSO-BP Neural Network Controller under 30pcm Step Disturbance of Reactivity
图 9 50pcm反应性阶跃扰动下堆芯归一化功率的变化
Figure 9. Change of Core Normalized Power under 50pcm Step Disturbance of Reactivity
图 10 50pcm反应性线性扰动下堆芯归一化功率的变化
Figure 10. Change of Core Normalized Power under 50pcm Linear Disturbance of Reactivity
图 11 2℃阶跃温度扰动下堆芯归一化功率的变化
Figure 11. Change of Core Normalized Power under 2℃Step Temperature Disturbance
表 1 冷却剂温度对比验证
Table 1. Comparison Verification of Coolant Temperature
参数 反应堆运行值 程序计算值 冷却剂平均温度/℃ 37.4 37.41 冷却剂出口温度/℃ 38.8 38.81 
下载: 导出CSV
-
[1] 张亚东. 49-2游泳池反应堆安全分析报告:ZYY.HSY.DG.224[R]. 北京: 中国原子能科学研究院,2016. [2] 何丽华,于涛,郑平卫. 反应堆功率控制系统PID控制器参数整定及仿真[J]. 核电子学与探测技术,2013, 33(6): 775-777,786. doi: 10.3969/j.issn.0258-0934.2013.06.029 [3] 李献,骆志伟,于晋臣. MATLAB/Simulink系统仿真[M]. 北京: 清华大学出版社,2017:396-398. [4] 黄轲,王琳,肖凯,等. 基于粒子群算法的反应堆功率调节系统优化研究[J]. 科技视界,2019(10): 64-66,70. [5] 曾雄飞. 基于粒子群算法优化BP神经网络的PID控制算法[J]. 电子设计工程,2022, 30(11): 69-73,78. [6] 黄忠霖. 控制系统MATLAB计算及仿真[M]. 第二版. 北京: 国防工业出版社,2004:284-285. [7] LORENZI S, CAMMI A, BORTOT S, et al. Analytical models for a small LFR core dynamics studies[J]. Nuclear Engineering and Design, 2013, 254: 67-88. doi: 10.1016/j.nucengdes.2012.09.001 [8] 罗璋琳,段天英. 反应堆控制与仪表[M]. 北京: 中国原子能出版社,1993:311-314. [9] 熊华胜,李铎. 基于Unscented卡尔曼滤波器的反应堆周期计算算法研究[J]. 原子能科学技术,2014, 48(S1): 568-571. [10] KIM Y B, VISTA IV F P, CHO S B, et al. Digitalization of the Ex-core neutron flux monitoring system for APR1400 nuclear power plant[J]. Applied Sciences, 2020, 10(23): 8331. doi: 10.3390/app10238331 [11] 朱珈辰. 49-2游泳池反应堆两套功率调节系统切换仿真分析[J]. 仪器仪表用户,2016, 23(5): 71-73. doi: 10.3969/j.issn.1671-1041.2016.05.022 -
计量
- 文章访问数: 48
- HTML全文浏览量: 13
- PDF下载量: 17
- 被引次数: 0
