Predictive control oriented identification of finite step response model

Bao-cang DING,Qiao CHEN,Ya-jun XIE
DOI: https://doi.org/10.13195/j.kzyjc.2015.1307
2016-01-01
Abstract:An algorithm for identifying the finite step response(FSR) model is given, which is utilized in the predictive control software development. This algorithm includes the stamp and interpolation of bad data, data filtering, identification case grouping and handling of the integral output, which is applied to the software development and engineering implementations. In calculating the FSR model, the classical least squares method and increasing penalty factors are applied. The contributions include an unified procedure for handling both stable and integral outputs, and the incremental method for both the data and the identification algorithm, which are consistent with the actually applied dynamic matrix control(DMC) techniques. The simulation example verifies the effectiveness of the proposed algorithm.
What problem does this paper attempt to address?