Nonlinear model reference observer design for feedback control of a low temperature combustion diesel engine

A nonlinear observer design is proposed for fuel delivery control of a low temperature combustion (LTC) diesel engine. First, a nonlinear engine model, combining the continuous time gas exchange and the event-based closed cycle processes, is presented for the diesel engine. Secondly, an optimization routine based on the extremum seeking scheme is implemented for online observer calibration on a 10 engine rotations basis to follow the engine sensor outputs. The extremum seeking algorithm efficiently minimizes the observer error, demonstrated by simulation results with embedded engine test data. Finally, a feedback control architecture is proposed based on the observer trajectory as the engine feedback.