System identification for a system subjected to persistent disturbances

This paper is aimed to analyse possible approaches for addressing multivariable system identification in presence of persistent disturbances. The main objective is to find a solution that is applicable to systems like wastewater treatment plants. Indeed, this type of processes can effectively improve their efficiency by means of advanced controllers, but their identification is quite difficult for the presence of continuous time varying characteristics of the feed flow rate. Optimal design of experiments is applied to the Benchmark Simulation Model no. 1 (Gernaey et al., 2014), and an adequate estimation of the process gains is obtained with multivariable identification if nonlinear models are used to describe the manipulated inputs – outputs relationships.