A filtered Smith predictor based subspace predictive controller

Abstract This paper presents a SPC based in the filtered Smith predictor structure in order to improve the performance of SPC when applied to a stable or integrative dead-time processes. This technique combines the robustness of subspace identification algorithms, the ability of predictive controllers to deal with multi variable processes and operational constraints and robustness of filtered Smith predictor in presence of model uncertainties. The proposed controller is applied in two simulation cases. The first is a boiler temperature control and the second is the control of effluent concentration and temperature in a stirred tank reactor. The obtained results show that the proposed strategy gives good performance when controlling dead-time processes considering estimation errors in dynamics and dead time.