摘要
以新型干法水泥分解炉为研究对象,提出自适应多维泰勒网控制方案,实现了分解炉出口温度的控制。首先基于互相关及递归最小二乘法建立多输入单输出含滞后环节分解炉出口温度数学模型;然后分析并设计自适应多维泰勒网控制器;最后使用Matlab对水泥分解炉出口温度的基于改进单纯形法的PID优化控制、自适应BP神经网络PID控制和自适应多维泰勒网控制进行仿真对比研究。结果表明,在相同条件下,自适应多维泰勒网控制具有最好的快速响应性和稳定性,最小的超调量和稳态误差。
In this paper,an adaptive multi-dimensional Taylor network(Adaptive MTN) control strategy is proposed to realize calciner outlet temperature control by taking the new dry-process cement calciner as the research object.Firstly,the multi-input-single-output outlet temperature mathematic model with time delay is established based on cross-correlation and recursive least squares method.Secondly,the adaptive MTN controller is analyzed and designed.Finally,Matlab is used to simulate and compare the improved simplex method based PID optimal control,adaptive BP neural network PID control and adaptive MTN control of calciner outlet temperature.
引文
[1]万春红,柬洪春,张东宁,等.水泥分解炉温度控制过程的阶跃响应建模与仿真[J].化工自动化及仪表,2011,38(12):1420-1424
[2]郝晓弘.模糊预测控制在新型干法水泥生产线的应用[J].计算机仿真,2010,27(5):281-283
[3]张志刚.神经动态规划在水泥分解炉温度控制中的应用研究[D].桂林:广西大学,2007
[4]王子峰.水泥分解炉温度控制系统研究[D].长春:长春工业大学,2016
[5]冯丽辉,吕智愚,董乃飞,等.水泥分解炉出口温度的多变量模糊控制[J].控制工程,2014,21(3):374-377
[6]Zhang J J, Yan H S. Nonlinear time-varying systems identification based on multi-dimensional Taylor network and variable forgetting factor recursive least squares algorithm[C]//1st IFAC Conference on Modeling, Identification and Control of Nonlinear Systems. Holland:Elsevier Ltd, 2015:1103-1107
[7]严洪森.多维泰勒网优化控制[R].南京:东南大学自动化学院制造系统控制与优化研究所,2010
[8]严洪森.多维泰勒网优化控制[EB/OL].http://automation.seu.edu.cn/dongnan-web/teacher/publish/info/30, 2018-09-11
[9]张德攀.基于神经网络PID的温度控制系统实时仿真研究[D].沈阳:东北大学,2007