Anticipation Augment-Assisted Active Disturbance Rejection Control and Its Application in SCR Denitrification Systems

Selective catalytic reduction (SCR) denitrification systems in coal-fired power plants exhibit large inertia, pronounced dynamic lag, time-varying parameters, catalyst ammonia-storage effects, and multiple disturbances, which limit the control performance of outlet nitrogen oxides (NOx). To address these issues, an anticipation augment-assisted active disturbance rejection control (AA-ADRC) framework is proposed and implemented in a linear form. The dynamic characteristics of classical mechanistic, improved mechanistic, and existing industrially identified SCR models are first analyzed, and a control-oriented linear active disturbance rejection control (LADRC) model is established based on the identified second-order ammonia-injection control channel. A finite-bandwidth filtered-derivative component is then extracted from the ammonia-injection command and converted into a finite-bandwidth anticipatory process increment. The augmented signal is introduced into the extended state observer (ESO), allowing input-variation information to participate earlier in state estimation and disturbance compensation while retaining the estimation and compensation of the generalized disturbance. Comparative simulations with proportional–integral (PI) and LADRC under inlet NOx step disturbances, composite inlet NOx disturbances, outlet NOx disturbances, and ammonia-injection disturbances show that AA-ADRC provides faster transient recovery in the step-disturbance case and reduces ammonia-command variation.