水箱液位多变量系统辨识方法的比较

Comparision of Gradient Based Identification for Three Water Tank Level Systems

针对流程工业中广泛使用的多反应塔液位控制系统,以三水箱液位系统为例,利用伯努利流体力学原理,推导了液位系统的多变量非线性数学模型.采用线性化和离散化方法,获得系统的状态空间模型和传递矩阵模型,分析辨识该多输入多输出模型的遗忘梯度算法、子系统遗忘梯度算法和递阶遗忘梯度算法,并对这3种算法进行仿真比较.结果表明,递介遗忘梯度算法计算量最小,计算效率最高,但参数估计性能介于遗忘梯度算法和子系统遗忘梯度算法之间.

For many reaction tower level systems multivariable nonlinear models are derived for a three Bernouilli＇s fluid principle. The state space models existing widely in process industries, the water tank level system as an example using and transfer matrix models of the nonlinear systems are obtained by using linearization and discretization, and the identification methods of the obtained models are studied by using the stochastic gradient algorithm, subsystem stochastic gradient algorithm and hierarchical stochastic gradient algorithm, The simulation and comparison are given for these identification methods. The hierarchical stochastic gradient algorithm has the lowest computation load and highest efficiency among the three algorithms, but the parameter estimation performance is situated between the stochastic gradient algorithm and the subsystem stochastic gradient algorithm.