| 系统降阶的马尔柯夫参数拟合法 |
| |
| 引用本文： | 王士杰.系统降阶的马尔柯夫参数拟合法[J].东华大学学报(自然科学版),1992(1). |
| |
| 作者姓名： | 王士杰 |
| |
| 作者单位： | 中国纺织大学自动化及计算机系过程控制及仪表教研室 |
| |
| 摘 要： | 降阶是应用大规模动态系统的重要问题。本文提出的方法是:求出原系统大量的马尔柯夫参数,并根据应用要求确定降阶模型的阶数。然后根据由本文作者提出的准则——模型的马尔柯夫参数尽可能与原系统相一致同时模型的静态增益却应与原系统必须相等,采用最小二乘拟合法来求出降阶模型的参数。在降阶模型中引入了纯滞后,这样就能显著地提高拟合的精度。本文通过几个数例证明作者提出的方法,计算方便,拟合精度较高,可以用来解决多数高阶单输入单输出稳定系统的降阶问题。
|
| 关 键 词： | 离散系统 连续系统 最小二乘拟合 系统降阶 马尔柯夫参数 双线性变换 |
METHOD FOR DEVELOPING REDUCED-ORDER MODELS BY LEAST SQUARE FITTING OF MARKOV PARAMETERS |
| |
| Wang Shijie.METHOD FOR DEVELOPING REDUCED-ORDER MODELS BY LEAST SQUARE FITTING OF MARKOV PARAMETERS[J].Journal of Donghua University,1992(1). |
| |
| Authors: | Wang Shijie |
| |
| Institution: | Department of Automation and Computer Science |
| |
| Abstract: | Order reduction is an important problem for applications of large scale dynamic systems.
This paper presents a method for developing reduced-order models by least square fitting of
Markov parameters.First,Markov parameters of original high order system are calculated
and reduced-order of the model is determined by requirements of application.Then,accord-
ing to the criterion raised by author of this paper that Markov parameters of the model and
of the original must be as almost identical as possible and their static gains must be equal,the
parameters of reduced-order model are obtained by least square fitting.The time delay led in
the model can remarkably increase the fitting accuracy.Some examples in this paper proved
that such a method is effective and simple,capable of solving order reduction problems of
most high order SISO stable systems. |
| |
| Keywords: | discrete systems continuous systems least square fitting system order reduction markov parameters bilinear transformation |
| 本文献已被 CNKI 等数据库收录！ |
