基于改进差分进化算法-径向基神经网络的电热水浴串级控制系统研究

doi:10.11949/j.issn.0438-1157.20190885

摘要/Abstract

摘要：

针对电热水浴装置温度控制中被控对象存在的大惯性、非线性、大延迟等特点，设计了一种基于改进差分进化(improved differential evolution, IDE)算法的径向基（radial basis function, RBF）神经网络串级控制系统。采用IDE算法对RBF神经网络的初始参数进行优化，采用优化后的RBF神经网络辨识主控制回路被控对象的Jacobian信息，进而实现主控制回路PID(proportional integration differentiation)控制器参数的在线调整。针对主控制回路控制器包含输出噪声，导致控制性能下降的问题，引入Kalman 滤波器对串级控制的主回路进行重新设计，控制对象的输出值经过Kalman 滤波算法处理后再返回闭环控制系统。以微化工领域常用电热水浴装置为对象，对IDE-RBF-PID-PI串级控制系统进行仿真实验，结果表明，IDE-RBF-PID-PI串级控制系统相较于常规串级控制，大大提高了控制性能，主控制回路引入的Kalman滤波算法有效消减控制系统的输出噪声，控制效果接近于无噪声的理想状态。

关键词: 过程控制, 神经网络, 差分进化, 串级控制, 电热水浴

Abstract:

Aiming at the large inertia, nonlinearity and large delay of the controlled object in the temperature control of the electric heated water bath device, RBF (radial basis function) neural network cascade control system based on the IDE(improved differential evolution) algorithm is designed. The IDE algorithm is used to optimize the initial parameters of the RBF neural network. The optimized RBF neural network is used to identify the Jacobian information of the controlled object of the main control loop, and then the online adjustment of the parameters of the main control loop PID (proportional integration differentiation) controller is realized. Aiming at the problem that the main control loop controller contains output noise which leads to the decline of control performance, the Kalman filter is introduced to redesign the main loop of the cascade control. The output value of the control object is processed by the Kalman filter algorithm and then returned to the closed loop control system. The simulation test of the IDE-RBF-PID-PI cascade control system is carried out for the common electric heated water bath device in the micro chemical industry. The results show that the IDE-RBF-PID-PI cascade control system greatly improves control performance compared to conventional cascade control. The Kalman filtering algorithm introduced in the main control loop effectively reduces the output noise of the control system, and the control effect is close to the ideal state without noise.

Key words: process control, neural network, differential evolution, cascade control, electric heated water bath

中图分类号:

TP 273

于蒙, 邹志云. 基于改进差分进化算法-径向基神经网络的电热水浴串级控制系统研究[J]. 化工学报, 2019, 70(12): 4680-4688.

Meng YU, Zhiyun ZOU. Electric heated water bath cascade control system research based on improved differential evolution algorithm-radial basis function neural network[J]. CIESC Journal, 2019, 70(12): 4680-4688.

图/表 11

图1 RBF神经网络结构图

Fig.1 Structure of RBF neural network

图2 差分进化算法优化RBF神经网络流程图

Fig.2 Flow chart of differential evolution algorithm optimizing RBF neural network

图3 Kalman滤波算法优化控制系统架构

Fig.3 Structure of Kalman filter algorithm optimizing control system

图4 电热水浴装置结构图

Fig.4 Structure of electric heated water bath

图5 改进差分进化算法代价函数J的优化过程

Fig.5 Optimization process of cost function J of improved differential evolution algorithm

图6 三种串级控制策略控制对比图

Fig.6 Comparison chart of three cascade control strategies

图7 RBF神经网络辨识Jacobian信息

Fig.7 Identification of Jacobian information by RBF neural network

图7 主控制回路PID控制器权值变化

Fig.7 Weight change of PID controller in main control loop

图9 含噪声无滤波器阶跃响应图

Fig.9 Step response graph with noise without filter

图10 含噪声有滤波器阶跃响应图

Fig.10 Step response graph with noise and filter

表1 不同状态下跟踪绝对误差

Table 1 Tracking absolute error under different states

数据状态	绝对误差积分
理想情况	0.01149
有噪声	0.08365
滤波后	0.01298

参考文献 35

1	刘凤萍 . 电热恒温水浴锅温度场(温度均匀性)测量值的不确定度评定[J]. 计量与测试技术, 2016, 43(1): 84-85.
	Liu F P . Evaluation of uncertainty about measurement results of electric heating constant temperature water-bath temperature uniformity[J]. Metrology & Measurement Technique, 2016, 43(1): 84-85.
2	王志甄, 邹志云 .电热水浴装置的MPC-PID串级控制研究[J]. 石油化工自动化, 2018, 54(3): 36-40.
	Wang Z Z , Zou Z Y . Research on MPC-PID cascade control of electric-heated water bath device[J]. Automation in Petro-Chemical Industry, 2018, 54(3): 36-40.
3	Karthikeyan R , Manickavasagam K , Tripathi S , et al . Neuro-fuzzy-based control for parallel cascade control[J]. Chemical Product and Process Modeling, 2013, 8(1): 15-25.
4	Sakthivel R , Sathishkumar M , Ren Y , et al . Fault-tolerant sampled-data control of singular networked cascade control systems[J]. International Journal of Systems Science, 2017, 48(10): 2079-2090.
5	Hu Y F , Chen C H , Wu H C , et al . Study on structural optimization design and cascade PID control of maglev actuator for active vibration isolation system[J]. Journal of Vibration and Control, 2018, 24(10): 1829-1847.
6	Khalil H K . Cascade high-gain observers in output feedback control[J]. Automatica, 2017, 80(4): 110-118.
7	林巍, 王亚刚 . 串级控制系统闭环辨识及PID参数整定[J]. 控制工程, 2018, 25(1): 11-18.
	Lin W , Wang Y G . Modeling for cascade control systems based on frequency domain and PID parameter tuning[J]. Control Engineering of China, 2018, 25(1): 11-18.
8	柴杰, 江青茵, 曹志凯 . RBF神经网络的函数逼近能力及其算法[J]. 模式识别与人工智能, 2002, 15(3): 310-316.
	Chai J , Jiang Q Y , Cao Z K . Function approximation capability and algorithms of RBF neural networks[J]. Pattern Recognition and Artificial Intelligence, 2002, 15(3): 310-316.
9	Behnam T , Fariba Z , Houman Z , et al . Utilization of RBF-ANN as a novel approach for estimation of asphaltene inhibition efficiency[J]. Petroleum Science and Technology, 2018, 36(16): 1216-1221.
10	丰会萍, 胡亚南, 李明辉, 等 . 基于RBF-PID的多功能包装机热封温度控制[J]. 制造业自动化, 2018, 40(1): 16-20.
	Feng H P , Hu Y N , Li M H , et al . Temperature control of heat-sealing of multi-function packaging machine based on RBF-PID[J]. Manufacturing Automation, 2018, 40(1): 16-20.
11	郑太雄, 潘松, 李永福，等 . 基于黑箱模型和RBF-PID的HCCI发动机燃烧正时控制[J]. 仪器仪表学报, 2015, 36(11): 2510-2518.
	Zheng T X , Pan S , Li Y F , et al . Combustion timing control of homogeneous charge compression ignition engine based on RBF-PID and black-box model[J]. Chinese Journal of Scientific Instrument, 2015, 36(11): 2510-2518.
12	Lionel L S , Godpromesse K , Andrew M F . A new static synchronous series compensator control strategy based on RBF neuro-sliding mode technique for power flow control and DC voltage regulation[J]. Electric Power Components and Systems, 2018, 46(4): 456-471.
13	Tu W P , Yang Y H , Du B , et al . Towards a real-time production of immersive spatial audio of high individuality with an RBF neural network[J]. Journal of Parallel and Distributed Computing, 2019, 13(1): 124-131.
14	Yang H J , Liu J K . An adaptive RBF neural network control method for a class of nonlinear systems[J]. Automatica, 2018, 5(2): 457-462.
15	Alireza Z , Afshin T . Utilization of the RBF network to model the nucleate pool boiling heat transfer properties of refrigerant-oil mixtures with nanoparticles[J]. Journal of Molecular Liquids, 2017, 24(7): 252-261.
16	Moscoso M , Bayona V , Kindelan M . Gaussian RBF-FD weights and its corresponding local truncation errors[J]. Engineering Analysis with Boundary Elements, 2012, 36(9): 1361-1369.
17	邓凯文, 陈海昕 . 基于差分进化和RBF响应面的混合优化算法[J]. 力学学报, 2017, 49(2): 441-455.
	Deng K W , Chen H X . Hybrid optimization algorithm based on differential evolutio and RBF response surface[J]. Chinese Journal of Theoretical and Applied Mechanics, 2017, 49(2): 441-455.
18	薛晓岑, 向文国, 吕剑虹 . 基于差分进化与 RBF 神经网络的热工过程辨识[J]. 东南大学学报(自然科学版), 2014, (4): 769-774.
	Xue X C , Xiang W G , Lyu J H . Thermal process identification based on differential evolution and RBF neural network[J]. Journal of Southeast University (Natural Science Edition), 2014, (4): 769-774.
19	Ji N , Xu D Z , Liu F . A novel adaptive neural network constrained control for solid oxide fuel cells via dynamic anti-windup[J]. Neurocomputing, 2016, 21(9): 134-142.
20	Liu K , Wu Y , Zhu T M , et al . Improved RBF network torque control in flexible manipulator actuated by PMAs[J]. Robotica, 2019, 37(2): 264-280.
21	Storn R . On the usage of differential evolution for function optimization[C]//Proceedings of North American Fuzzy Information Processing. Berkeley, CA, USA, 1996.
22	Li Y C , Li Z P , Yang L Q , et al . An improved quantum differential evolution algorithm for optimization and control in power systems including DGs[J]. Automatica, 2017, 43(7): 1280-1288.
23	Dong Y , Zhao R . Solve train stowage planning problem of steel coil using a pointer-based discrete differential evolution[J]. International Journal of Production Research, 2018, 56(22): 6937-6955.
24	Subrat K N , Pravat K R , Alok J , et al . Elitism-based multi-objective differential evolution with extreme learning machine for feature selection: a novel searching technique[J]. Connection Science, 2018, 30(4): 362-387.
25	Addawe R C , Addawe J M , Sueño M R K , et al . Differential evolution-simulated annealing for multiple sequence alignment[J]. Journal of Physics: Conference Series, 2017, 8(1): 120-131.
26	Ponnuthurai N , Mostafa Z . Population topologies for particle swarm optimization and differential evolution[J]. Swarm and Evolutionary Computation, 2018, 39(4): 24-35.
27	刘金琨 . 先进PID 控制MATLAB 仿真[M]. 北京: 电子工业出版社, 2016: 301-319.
	Liu J K . Advanced PID Control MATLAB Simulation[M]. Beijing: Electronic Industry Press, 2016: 301-319.
28	Bourmaud G , Megret R , Arnaudon M , et al . Continuous-discrete extended Kalman filter on matrix lie groups using concentrated gaussian distributions [J]. Journal of Mathematical Imaging and Vision, 2015, 51(1): 209-228.
29	Torabi H , Pariz N , Karimpour A . A novel cubature statistically linearized Kalman filter for fractional-order nonlinear discrete-time stochastic systems[J]. Journal of Vibration and Control, 2018, 24(24): 5880-5897.
30	Daniel T , Gary V . Attribution of hedge fund returns using a Kalman filter[J]. Applied Economics, 2018, 50(9): 1043-1058.
31	Zhao Y B , Li X F , Wang Y , et al . Biased constrained hybrid Kalman filter for range-based indoor localization[J]. IEEE Sensors Journal, 2018, 18(4): 1647-1655.
32	吴晓明, 马立廷, 郑协, 等 . 改进的RBF神经网络PID 算法在电液伺服系统中应用[J]. 机床与液压, 2015, 43(11): 63-66.
	Wu X M , Ma L T , Zheng X , et al . Improved RBF neural network PID control strategy used in electro-hydraulic servo system[J]. Machine Tool & Hydraulics, 2015, 43(11): 63-66.
33	Xie Y F , Yu J J , Xie S W , et al . On-line prediction of ferrous ion concentration in goethite process based on self-adjusting structure RBF neural network[J]. Neural Networks, 2019, 5(2): 132-141.
34	Kwedlo W , Bandurski K . A parallel differential evolution algorithm [C]//Parallel Computing in Electrical Engineering.IEEE, 2006, 12(9): 454-467.
35	Sento A , Kitjaidure Y . Neural network controller based on PID using an extended Kalman filter algorithm for multi⁃variable non⁃linear control system [C]//The Eighth International Conference on Advanced Computational Intelligence. Chiang Mai: IEEE, 2016: 302⁃309.

编辑推荐 0

Metrics

阅读次数

全文

181

HTML			PDF

最新录用	在线预览	正式出版	最新录用	在线预览	正式出版
0	0	2	11	0	168

来源	本网站	其他网站

次数	113	68
比例	62%	38%

摘要

441

最新录用	在线预览	正式出版

15	0	426

来源	本网站	其他网站

次数	311	130
比例	71%	29%

[1]	曹跃, 余冲, 李智, 杨明磊. 工业数据驱动的加氢裂化装置多工况切换过渡状态检测[J]. 化工学报, 2023, 74(9): 3841-3854.
[2]	康飞, 吕伟光, 巨锋, 孙峙. 废锂离子电池放电路径与评价研究[J]. 化工学报, 2023, 74(9): 3903-3911.
[3]	温凯杰, 郭力, 夏诏杰, 陈建华. 一种耦合CFD与深度学习的气固快速模拟方法[J]. 化工学报, 2023, 74(9): 3775-3785.
[4]	尹刚, 李伊惠, 何飞, 曹文琦, 王民, 颜非亚, 向禹, 卢剑, 罗斌, 卢润廷. 基于KPCA和SVM的铝电解槽漏槽事故预警方法[J]. 化工学报, 2023, 74(8): 3419-3428.
[5]	诸程瑛, 王振雷. 基于改进深度强化学习的乙烯裂解炉操作优化[J]. 化工学报, 2023, 74(8): 3429-3437.
[6]	闫琳琦, 王振雷. 基于STA-BiLSTM-LightGBM组合模型的多步预测软测量建模[J]. 化工学报, 2023, 74(8): 3407-3418.
[7]	徐野, 黄文君, 米俊芃, 申川川, 金建祥. 多源信息融合的离心式压缩机喘振诊断方法[J]. 化工学报, 2023, 74(7): 2979-2987.
[8]	高学金, 姚玉卓, 韩华云, 齐咏生. 基于注意力动态卷积自编码器的发酵过程故障监测[J]. 化工学报, 2023, 74(6): 2503-2521.
[9]	黄磊, 孔令学, 白进, 李怀柱, 郭振兴, 白宗庆, 李平, 李文. 油页岩添加对准东高钠煤灰熔融行为影响的研究[J]. 化工学报, 2023, 74(5): 2123-2135.
[10]	苏晓丹, 朱干宇, 李会泉, 郑光明, 孟子衡, 李防, 杨云瑞, 习本军, 崔玉. 湿法磷酸半水工艺考察与石膏结晶过程研究[J]. 化工学报, 2023, 74(4): 1805-1817.
[11]	贠程, 王倩琳, 陈锋, 张鑫, 窦站, 颜廷俊. 基于社团结构的化工过程风险演化路径深度挖掘[J]. 化工学报, 2023, 74(4): 1639-1650.
[12]	吴选军, 王超, 曹子健, 蔡卫权. 数据与物理信息混合驱动的固定床吸附穿透深度学习模型[J]. 化工学报, 2023, 74(3): 1145-1160.
[13]	张中秋, 李宏光, 石逸林. 基于人工预测调控策略的复杂化工过程多任务学习方法[J]. 化工学报, 2023, 74(3): 1195-1204.
[14]	张江淮, 赵众. 碳三加氢装置鲁棒最小协方差约束控制及应用[J]. 化工学报, 2023, 74(3): 1216-1227.
[15]	吴心远, 刘奇磊, 曹博渊, 张磊, 都健. Group2vec：基于无监督机器学习的基团向量表示及其物性预测应用[J]. 化工学报, 2023, 74(3): 1187-1194.