Learning-Based Decision Making for Combustion Phasing Control in Multi-Fuel CI Engines with Latent Fuel Reactivity Estimation
Researchers have developed a new reinforcement learning framework to improve combustion phasing control in multi-fuel compression-ignition engines. This system addresses the challenge of uncertain and time-varying fuel reactivity, quantified by cetane number (CN), which complicates precise control. The proposed framework utilizes a gated recurrent unit (GRU) to learn a compact representation of fuel reactivity from combustion history, enabling the control policy to make decisions based on this estimated signal rather than relying on an oracle CN value. This approach aims to prevent train-deploy inconsistencies and achieve stable CA50 regulation with low tracking errors, even with rapidly evolving CN. AI
IMPACT This research could lead to more efficient and adaptable multi-fuel engines by improving combustion control through advanced AI techniques.