Multivariate Close-loop Identification

In industry control, nowsdays, the control algorithm which includes PID control and many advanced control algorithm such as predictive control and inner model control almost based on model, therefore,system identification become one of the control theory.In PID control, because there are still many single loop, single identification play an important role presently This paper has studied the continuous-model identification based on NLJ algorithm which has already apply to industry process successfully. We have already developed a control software which includes data-processing, identification based on NLJ algorithm and controller design based on NLJ algorithm and IMC-PID.Advanced control always use the multivariate model, so multivariate close-loop identification is one hot point of research. Aims at the most of the systems is close-loop system, this paper particularly studies the multivariate close-loop identification methods. Most of subspace identification algorithms at home have errors in the errors-in-variable situation and close-loop situation. This paper introduces a new identification algorithm that uses principle component analysis (PCA) to identify the system matrices. That avoids the projection in other algorithms, so it can be applied to errors-in-variable and close-loop situation.At last, some key technologies has been discussed which used in the control software, such as how to get the wavelet coefficient in VC++, matrix-connection and polynomial-connection methods. This paper also demonstrates all the function of the control software.