High thermal efficiency and low engine-out emissions including nitrogen oxides
(NOx) and particulate matter (PM) make low-temperature combustion (LTC)
favorable for use in engine technologies. Homogeneous charge compression
ignition (HCCI), partially premixed charge compression ignition (PPCI), and
reactivity controlled compression ignition (RCCI) are among the common LTC
modes. These three LTC modes can be achieved on the same dual-fuel engine
platform; thus, an engine controller can choose the best LTC mode for each
target engine load and speed. To this end, a multi-mode engine controller is
needed to adjust the engine control variables for each LTC mode.
This article presents a model-based control development of a 2.0-liter multi-mode
LTC engine for cycle-to-cycle combustion control. The engine is equipped with
port fuel injectors (PFI) and direct injectors (DI). All combustion modes are
achieved with dual fuels (iso-octane and n-heptane) under
naturally aspirated conditions. Using experimental data, control-oriented models
(COMs) are developed for HCCI, PPCI, and RCCI combustion modes on a
cycle-to-cycle basis. The COMs for HCCI, PPCI, and RCCI modes can predict the
combustion phasing (CA50, the crank angle by which 50% of the fuel mass is
burned) with average errors of 1.3 crank angle degrees (CAD), 1.5 CAD, and 1
CAD, respectively. The average errors in predicting the indicated mean effective
pressure (IMEP) for HCCI, PPCI, and RCCI modes are 18 kPa, 34 kPa, and 43 kPa,
respectively. Multi-input and multi-output (MIMO) adaptive model predictive
controllers (MPCs) with linear parameter varying (LPV) models are designed for
the LTC modes. CA50 and IMEP are controlled by adjusting the premixed ratio (PR)
of the fuels, start of injection (SOI) timing, and fuel quantity (FQ). The
results show that the designed MPCs are able to track both CA50 and IMEP in all
combustion modes, with average tracking errors of less than 1 CAD and 5.2 kPa,
respectively.
