Nonlinear Predictive Control of Combustion and Emissions in Direct Injection Engines with Nozzle Aging

Current developments in emissions legislation mean that real driving emissions (RDE) are becoming increasingly important and make it worthwhile to examine longterm aging phenomena in direct injection engines. This paper proposes an advanced control framework to account for aging effects occurring in the injection nozzles and ensure consistent engine behavior over an extended operating time. Based on the previously identified aging condition, a nonlinear model predictive controller (NMPC) was designed to optimize future engine performance and emissions. The underlying model was derived from experimental data of a single-cylinder diesel engine, and the numerical effort was significantly reduced using a neural network approximation. Simulated experiments with the NMPC demonstrated the great potential in controlling exhaust emissions and improving long-term engine performance.