基于子空间方法的超超临界机组过热蒸汽系统模型辨识  

Model Identification Based on Subspace Model Identification of Superheated Steam System in Ultra-Supercritical Coal-Fired Power Unit

作　　者：孟庆阳[1] 阎威武[1] 胡勇[1] 程建林[1] 陈世和 张曦 

机构地区：[1]上海交通大学自动化系,上海200240 [2]广东电网电力科学研究院,广州510080

出　　处：《上海交通大学学报》2017年第6期672-678,共7页Journal of Shanghai Jiaotong University

基　　金：国家自然科学基金项目(60974119;61533013)资助

摘　　要：针对超超临界机组过热蒸汽系统模型辨识,采用基于数据驱动的闭环子空间辨识方法直接得到系统的阶跃响应系数,结合传统的SISO(Single Input Single Output)辨识方法,并利用最小二乘算法回归出系统的低阶传递函数模型.通过对某电厂超超临界机组过热蒸汽系统进行闭环辨识,结果表明,该策略很好地融合了子空间方法的简便性以及传统SISO辨识方法的最优性,并成功用于超超临界机组过热蒸汽系统模型辨识.The superheated steam system ultra-supercritical unit is a multi-input and multi-output system. In this paper, a data-driven subspace approach is introduced for the model identification o{ the superheated steam system. In the approach, the unit step response coefficients are drawn directly, then the low-order transfer function of the system is obtained by least-square regression. Finally, closed loop identification based on data-driven subspace approach is applied to the superheated steam system of ultra-supercritical coal-fired power unit. Real application shows that proposed identification strategy has a small computation- al cost and can meet the accuracy requirements in industrial sites.

关 键 词：超超临界机组 过热蒸汽系统 子空间辨识 闭环辨识方法 

分 类 号：TP271[自动化与计算机技术—检测技术与自动化装置]