Model predictive control of HEVs with exhaust aftertreatment system at low ambient temperatures

Online control of hybrid electric vehicles (HEV) is mostly centered on fuel consumption and battery management while emissions are seldom considered. This is frequently correct, as warmed up exhaust aftertreatment systems show an extremely high conversion efficiency. However, HEVs typically shut off the engine during low load phases, which extends coldstart periods and prevents the aftertreatment to work at the correct temperature. Latest regulations put additional focus on real driving coldstart performance though. Against this background, this paper analyzes the potential relevance and presents a possible solution by considering it explicitly in the control approach. It is shown that an implementable model predictive control (MPC) strategy can recover a large part of the theoretical performance as computed by dynamic programming (DP). Optimal MPC parameter tuning is efficiently performed utilizing available DP solutions from comparable scenarios. A potential saving of up to 40% NOx with equal consumption is shown in this example.