基于现场数据热工对象建模的可辨识性

为克服现有机理建模和试验建模方法的不足,从基本建模理论出发,研究了利用现场数据对热工对象进行建模的问题。基于现场数据建模需要对热工过程输入信号的可激励性、控制系统闭环可辨识性、多变量系统输入信号的无关性等进行判断。结果表明:在较大的机组负荷波动过程中采集的运行数据能够满足输入信号的可激励性;热工过程的大滞后现象对保证热工控制系统的闭环可辨识性有利;含有噪声的实测信号能够保证多变量系统的可辨识性。在满足以上条件的情况下,可以基于现场数据对热工对象进行建模研究。

Identifiability of building thermal system models using on-line data

Models of thermal systems were developed using basic modeling principles and on-line data to improve the conventional mechanistic method and testing method. The stimulation of the input signal, identifying closed loops of control system, and modeling of multi-variable systems were analyzed to meet the requirements of building thermal system model using on-line data. The stimulation of the input signal can be guaranteed by data obtained during power load changing in a large range; identifying closed loops of control systems benefit from the lag effects of thermal systems, and identifying of multi-variable systems can be guaranteed by the measured signals with noise. Dynamic models of thermal systems using on-line data can be obtained with the above conditions being met.