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

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。这种情况促使在先进控制中越来越多地直接采用非线性分布参数机理

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

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.