基于DRNN的多回路参数在线自整定控制解耦算法  

Multi-Loop Parameter Online Auto-Tuning Control Algorithm Based on DRNN Decoupling

作　　者：肖宏启 XIAO Hongqi(Department of Computer Science,Guizhou Institute of Aerospace Technology,Zunyi 563000,China)

机构地区：[1]贵州航天职业技术学院计算机科学系

出　　处：《兵器装备工程学报》2019年第12期115-119,共5页Journal of Ordnance Equipment Engineering

摘　　要：提出了一种基于DRNN解耦的控制参数在线自整定控制算法。分析了关联耦合回路控制中存在的问题,推导了基于DRNN解耦的控制参数在线自整定算法。以某军用弹药仓库库房的温、湿度高精度控制为例,借助仿真验证了采用前馈补偿的DRNN解耦可消除关联控制通道间影响,通过解耦后的各个控制通道间几乎互不影响,在解耦的同时完成控制参数的自整定。研究结果表明,对于关联耦合回路,采用基于DRNN的解耦控制算法是合理、可行和有效的。The paper explored an online auto-tuning control algorithm based on DRNN decoupling.In the paper,it made the anatomy of the puzzles existing in the control of correlation coupling loops,discussed the concept of control decoupling,and deduced the control algorithm based on diagonal regression neural network decoupling of coupling multiple loops.Taking the high-precision control of temperature and humidity in a storeroom of military ammunition warehouse as an example,the simulation response demonstrated that the decoupling of feed forward compensation based on DRNN could eliminate the influence among the correlation coupling control channels,and after through decoupling,each control channel has little influence on the other,and the auto-tuning of control parameters could also be completed at the same time.The research results show that the decoupling control algorithm based on DRNN is reasonable,feasible and effective for the correlation coupled multi-loop control.

关 键 词：耦合变量 关联耦合回路控制 参数自整定 对角回归神经网络 解耦控制算法 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]