摘要
提出了一种基于回归分析弱化系统耦合度的方法,将以精确数学模型为基础进行耦合度分析的问题转化为建立输入输出之间最优回归方程的问题,通过回归方程系数的显著性检验来确定对输出影响显著的输入量。与已有的方法比,回归分析方法不要求精确的数学模型,而且适用于输入与输出变量个数不同的非方形系统的耦合度分析。理论与实验分析验证了所提方法的有效性。
A regression analysis-based method was proposed to weaken linear system＇s coupling degree,and transform the accurate mathematical model-based coupled analysis into establishing an optimal regression equation between the input and output,and to determine the input quantity that significantly influences the output through a significance test of regression equation coefficient.Compared with existing methods,the theory analysis and simulation results show that regression analysis-based method requires no precise mathematical model,and can be applied to the coupling degree analysis of non-square systems boasting of different input and output variables.
出处
《化工自动化及仪表》
CAS
北大核心
2011年第1期34-36,共3页
Control and Instruments in Chemical Industry
关键词
多变量系统
耦合度
回归分析
显著性
multiple variation system
coupling degree
regression analysis
estimation significance
