Research of ALSTM-GPC in Coordinated Control System of Nuclear Power Plant
-
摘要: 针对常规比例积分微分(PID)控制器面对复杂系统时控制效果欠佳的问题,充分结合深度学习在特征提取、回归预测以及预测控制在处理多变量、强耦合等问题的优势,先通过ALSTM深度网络构建预测模型控制器,该预测模型以对象一维时序信号作为输入,以长短期记忆网络(LSTM)完成特征提取,然后通过引入注意力机制,对LSTM提取特征的权重参数进行优化,筛选保留目标特征,滤除冗杂特征,实现精准提取有效时序特征。再将广义预测控制(GPC)作为滚动优化控制器,搭建注意力长短期记忆网络-广义预测控制器(ALSTM-GPC)。随后在核电站双入双出多变量协调控制系统进行仿真实验,通过设定值扰动、内扰以及外扰等一系列仿真实验的验证,与常规PID控制相比,ALSTM-GPC控制器具有更优的控制效果。
-
关键词:
- 长短期记忆网络(LSTM) /
- 注意力机制 /
- 阶梯式广义预测控制 /
- 协调系统 /
- 核电厂
Abstract: Aiming at solving the problem of poor control effect of conventional proportion integration differentiation (PID) controller when dealing with complex system, combining the advantages of deep learning in feature extraction, regression prediction and predictive control in dealing with multivariable and strong coupling, a predictive model controller is built by ALSTM deep networks, the model takes one-dimensional time series signal of object as its input, completes feature extraction through LSTM. Then, by introducing the attention mechanism, the weight parameters of LSTM extraction features are optimized, the target features are screened and retained, and the redundant features are filtered to achieve accurate extraction of effective time series features. And then a generalized predictive control (GPC) is used as a rolling optimization controller, thus building attention long short-term memory-generalized predictive controller (ALSTM-GPC). Then, the simulation experiment was carried out in the nuclear power plant's double-input and double-output multivariable coordinated control system. Through a series of simulations such as set-point disturbance, internal disturbance and external disturbance, ALSTM-GPC controller is proven having better control effect than conventional PID control. -
表 1 弗朗明歇距离对比
Table 1. Frechet Distance Comparison
传递函数 弗朗明歇距离 BP ALSTM-GPC G11 0.0957 0.0517 G21 0.00759 0.00643 G12 0.00630 0.00171 G22 0.00763 0.00146 -
[1] 王蕾,宋文忠. PID控制[J]. 自动化仪表,2004, 25(4): 1-6. doi: 10.3969/j.issn.1000-0380.2004.04.001 [2] 沈永福,吴少军,邓方林. 智能PID控制综述[J]. 工业仪表与自动化装置,2002(6): 11-13,24. doi: 10.3969/j.issn.1000-0682.2002.06.003 [3] 施小成,辛成东,边信黔,等. 船用动力装置智能协调控制策略研究[J]. 船舶工程,1997(2): 18-20. [4] 徐立伟. 核电站控制系统模糊控制应用研究[D]. 南京: 东南大学, 2014. [5] 李佳文,王景升. 基于深度学习的短时交通流预测[J]. 山东交通科技,2018(5): 87-90. doi: 10.3969/j.issn.1673-8942.2018.05.025 [6] SCHUSTER M, PALIWAL K K. Bidirectional recurrent neural networks[J]. IEEE Transactions on Signal Processing, 1997, 45(11): 2673-2681. doi: 10.1109/78.650093 [7] 朱俊丞,杨之乐,郭媛君,等. 深度学习在电力负荷预测中的应用综述[J]. 郑州大学学报(工学版),2019, 40(5): 12-21. [8] 易灵芝,常峰铭,龙谷宗,等. 基于进化深度学习短期负荷预测的应用研究[J]. 电力系统及其自动化学报,2020, 32(3): 1-6,13. [9] 李少远. 工业过程系统的预测控制[J]. 控制工程,2010, 17(4): 407-415. doi: 10.3969/j.issn.1671-7848.2010.04.001 [10] HOCHREITER S, SCHMIDHUBER J. Long short-term memory[J]. Neural Computation, 1997, 9(8): 1735-1780. doi: 10.1162/neco.1997.9.8.1735 [11] 刘建帮,孙威,张宪霞,等. 多变量预测控制工程应用的控制模型前馈解耦策略[J]. 控制与决策,2019, 34(5): 1094-1102. [12] 栾秀春,李士勇. 基于局部神经网络模型的过热汽温多模型预测控制的研究[J]. 中国电机工程学报,2004, 24(8): 190-195. doi: 10.3321/j.issn:0258-8013.2004.08.038 [13] 张建华,董菲,侯国莲,等. 基于神经网络预测控制的单元机组协调控制策略[J]. 动力工程,2006, 26(3): 392-395. [14] 赵洁,刘涤尘,吴耀文. 压水堆核电厂接入电力系统建模[J]. 中国电机工程学报,2009, 29(31): 8-13. doi: 10.3321/j.issn:0258-8013.2009.31.002 [15] 赵洁,刘涤尘,熊莉,等. 基于PSASP自定义模型的核电机组动态响应仿真[J]. 核动力工程,2010, 31(3): 113-117+142. [16] 王宝生,王冬青,张建民,等. 压水堆核电厂蒸汽排放控制系统实时仿真研究[J]. 核动力工程,2011, 32(5): 38-44.