Dynamic calibration of fuelling in the PFI SI engine

This paper presents the experimental validation of a system identification method for the systematic calibration of a fuelling and speed controller strategy for a production PFI SI engine. The technique offers a time-efficient ‘one-shot’ alternative to traditional fuelling map calibration methods since it is based entirely on rapidly obtained dynamically perturbed input–output data.Non-linear black-box parameter identification directly produces a dynamic inverse multivariable NARMA feedforward controller and linearising feedback compensator. The technique is experimentally demonstrated on a 4-cylinder 1.6 l engine over the idle region in combination with a robust H∞H∞ multivariable feedback design.