Robust estimation for an uncertain linear model of an activated sludge process

A two-step strategy is developed to estimate nitrogen constituents in a single reactor activated sludge process with alternated aeration. The process is represented by a reduced-order linear model. Unmeasured components in the influent (as the ammonia concentration) and state variables are estimated. Parameter uncertainty is considered in the dynamics and input matrices of the system. Results are presented for simulated data with variations of the operating conditions for which the robustness of the approach is evaluated.