Nonlinear model predictive control applied to multivariable thermal and chemical control of selective catalytic reduction aftertreatment

Manufacturers of diesel engines are under increasing pressure to meet progressively stricter NOx emissions limits. A key NOx abatement technology is selective catalytic reduction (SCR) in which ammonia, aided by a catalyst, reacts with NOx in the exhaust stream to produce nitrogen and water. The conversion efficiency is temperature dependent: at low temperature, reaction rates are temperature limited, resulting in suboptimal NOx removal, whereas at high temperatures, they are mass transfer limited. Maintaining sufficiently high temperature to allow maximal conversion is a challenge, particularly after cold start, as well as during conditions in which exhaust heat is insufficient, such as periods of low load or idling. In this work, a nonlinear model predictive controller simultaneously manages urea injection and power to an electric catalyst heater, in the presence of constraints.