工业过程稳态优化ISOPE方法的鲁棒性分析

Robustness Analysis of ISOPE Methods for Steady-State Optimizing Control of Industrial Processes

对于大多数的实际工业过程，系统的稳态输出常常含有量测噪声．因而不可能得到系统的真实稳态输出值及过程导数真值，因而也不可能得到实际最优解。本文针对ＩＳＯＰＥ方法（系统优化和多数估计的综合方法）给出了两种评价其算法鲁棒性的指标，并考察了其算法的收敛性与次优性。最后对几种典型的ＩＳＯＰＥ方法进行了鲁棒性分析。

In general,the real process steady-state outputs are often contaminated with measumentnoise and hence,auurate measuments of real steady-state outputs and estimations of processderivatives cannot be assumed so,the real optimal solution cannot be obtained at all by using theISOPE methods,In the paper,two indexes which are used to assess Robustness of ISOPE meth-ods against noise are first presented,and the suboptimality and convergence of the ISOPE math-ods are examined.Finally,the Robustness of some typical ISOPE methods are discussed in detail.