Identification of control-relevant diesel engine models using a local linear parametric approach

Control is essential to meet future emission requirements in combustion engines. Accurate models are required to design a controller that achieves robust performance over a range of operating conditions. The aim of this paper is to develop a non-parametric and parametric identification procedure that is specifically tailored towards high performance diesel engine control, while minimizing measurement time. First, a non-parametric identification is proposed where the inputs are excited with multisines. A Local Rational Method (LRM) is employed to obtain multivariable Frequency Response Functions (FRFs) in a single experiment. Secondly, a parametric identification procedure uses the non-parametric estimates to obtain control-relevant parametric models. The identification procedure is demonstrated using a modern Heavy-Duty Diesel (HDD) engine providing highly accurate low order parametric models for a 2x2 plant using just 300s of measurement time at an engine operating point.