Towards cycle-to-cycle control: AI-based heat release rate prediction of a diesel engine operated in low temperature combustion mode

Recent market trends suggest that internal combustion engines will still have a significant role in transportation. Commercial diesel engines have to be further developed to meet strict environmental legislation. Low temperature combustion (LTC) is a new technology that can enhance the efficiency and reduce the pollutant emissions of compression ignition engines. However, some LTC methods deal with problematic combustion control that could be handled by model predictive cycle-to-cycle control. The applied physicochemical models require high computational capacity, increasing engine control costs. Artificial intelligence (AI) can provide a better alternative approach since well-trained AI models can perform calculations much faster. Therefore, this work aims to create AI models to predict combustion characteristics under varying conditions, including conventional and LTC operations. 113 different intake oxygen concentrations were set by exhaust gas recirculation, and the heat release rates and combustion pressures were predicted with neural networks. The results demonstrated that the AI can accurately predict both characteristics, providing an excellent possibility for further research where an AI-based model predictive controller could be built.