Adaptive CI-RCCI modeling and hierarchical LPV-MPC control for energy-efficiency mode transition in marine methanol/diesel dual-fuel engine

As global decarbonization targets tighten, accurate prediction and control of CI-RCCI combustion mode transitions are essential for achieving high efficiency and ultra-low emission in methanol/diesel dual-fuel marine engines. This study developed a cycle-resolved adaptive CI-RCCI model with a shared thermodynamic structure, mode-specific ignition kernels, fuel-transport dynamics, and a first-cycle residual-temperature correction to capture RCCI’s strong sensitivity to residual energy. An LPV state-space model is then identified via LS-SVM, enabling real-time prediction of combustion phasing and load as functions of methanol substitution and residual states, specifically the residual gas temperature and residual gas fraction. Building on this model, a hierarchical mode-transition control (HMTC) framework is established, consisting of an upper-layer coordinator and a lower-layer LPV-MPC. The coordinator determines feasible modes, triggers CI↔RCCI transitions, and generates smooth CA50 trajectories, while the LPV-MPC regulates CA50 and IMEP within a single control structure under combustion and actuator constraints. Comprehensive validations under intake-pressure steps, elevated residual fractions, and ±15 °C intake-temperature disturbances demonstrate that the proposed controller achieves fast, smooth, and stable transitions compared with uncontrolled operation. It reduces transition duration by up to 65%, lowers COVIMEP by more than 30%, and eliminates knock and misfire, while consistently maintaining phasing limits and preserving steady-state efficiency-emission characteristics. The proposed framework offers a robust, repeatable, and emission-consistent solution for intelligent low-carbon marine industry.