Abstract:

As the main steam temperature system is a multivariable and strong coupling system, a large number of field operational data is collected. Through principal component analysis(PCA)，five main variables are extracted from 11 procedure variables as the input of the main steam temperature system. Simultaneously, the mechanism modeling, experimental modeling and intelligent optimization algorithms modeling are combined. Finally, the multipleinput singleoutput multivariable model is established by using the field data. This solution solves the problem of frequent and largescale step disturbance tests in which it is not allowed or cannot to get to the productive site. It overcomes the uncertainty of the range of initial values and model structure for multivariable intelligent optimizing, and established a multivariable model that can be used for studying main steam temperature system control.

Key words:

main steam temperature, multivariable modeling, principal component analysis, field data