基于特征线法的一阶双曲型分布参数模型辨识及其在烟气脱硫过程中的应用  

Characteristics based parameter identification for hyperbolic distributed systems and its application in flue gas desulfurization process

作　　者：范丽婷[1] 王福利[1] 李鸿儒[1] 

机构地区：[1]东北大学信息科学与工程学院,辽宁沈阳110004

出　　处：《化工学报》2013年第7期2543-2549,共7页CIESC Journal

基　　金：国家自然科学基金项目(61074074);国家重点基础研究发展计划项目(2009CB320601)~~

摘　　要：引言在现代控制工程领域中,许多工业对象实际上是非线性分布参数系统。由于这类对象的复杂性,原始模型常常进行集中线性化处理后分析和设计控制系统,然而系统本质的分布特性以及非线性引起的模型失配将造成控制的失败。More and more nonlinear distributed parameter model described by partial differential equations have been directly adopted in advanced control strategies to improve control performance.However,some model parameter is highly sensitive to the operating condition that regularly changes.A novel characteristics based identification algorithm was proposed for estimating the time-varying parameter for a class of hyperbolic distributed model.Based on the method of characteristics,the partial differential equation（PDE）was transformed into a system of ordinary differential equations（ODEs）.According to the analytical expression of the state variable,the model parameter was optimized and updated iteratively based on the latest measurable data sequence.Simulation results showed that the proposed identification algorithm achieved good performance and the time-varying parameters could rapidly converge to its true values.The experimental application on a circulating fluidized bed scrubber achieved an average prediction error within ±6.5%,which provided the basis to implement advanced control strategies with mechanism model described by partial differential equations.

关 键 词：分布参数系统 特征线法 参数辨识 循环流化床 烟气脱硫 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]