Deep learning-based prediction model for combustion emissions and limits

Deep learning model for RCCI (diesel-natural gas) combustion

• Development of a reinforcement learning–based model to predict combustion mode (conventional diffusion flame vs. RCCI-based homogeneous combustion), emissions, and operational stability in diesel–natural gas dual-fuel engines.

• Model training using diverse RCCI combustion cases capable of achieving ~0.2 g/kWh NOx and ~1.468 mg/kWh PM at gIMEP 15.35 bar. [1]

• Sechul is doing this work in collaboration with KIMM and Prof. Seung Hyup Shin (Sejong Univ.).

References

[1] Sechul Oh, et al. "Effects of piston shape and nozzle specifications on part-load operation of natural gas–diesel dual-fuel RCCI engine and its application to high load extension." Fuel 328 (2022): 125361.