Simultaneous state and parameter estimation method for a conventional ozonation system

This article presents a simultaneous state (via a nonlinear form of Luenberger observer) and parameter (using a proportional–integral least mean square form) estimator design method for a conventional ozonation system. The suggested state observer assumes that the only available output signal is the concentration of the ozone gas at the output of the reactor. The estimation of the reaction rate constants of ozonation in the presence of contaminants uses the suggested proportional–integral estimation method. The convergence proof of the developed state-parameter identification method was confirmed using a Lyapunov based stability analysis. This analysis characterizes the quality of estimation considering the presence of modeled uncertainties and external perturbations. The implementation of the super-twisting algorithm as a robust and exact differentiator allowed to perform the estimation of the reaction rate constants of the ozonation, the temporal evolution of the dissolved ozone and the evolution of contaminants concentrations. The simultaneous state and parameter estimator design method was implemented in real-time using phenol as a model contaminant. The numerically simulated and real-time implementations showed that the method provides accurate estimates of the contaminant concentration and the reaction rate coefficient in all the evaluated cases.