On-line estimation of parameters in a respiration model using a continuous-discrete type of recursive prediction error algorithm

Estimates of biokinetic parameters for an activated sludge respiration model are obtained using an `on-line´ combined state/parameter estimation algorithm. The analysis suggests that the continuous-discrete type of recursive prediction error algorithm can be used to (i) obtain improved estimates of certain identifiable parameter combinations based on oxygen data only, and (ii) to evaluate the model `on-line´ through recursive estimates of the steady state Kalman gain matrix K_∞.