Abstract
Keywords - Air fuel ratio estimation, Kalman filtering, Engine control, Real time system, Testbed implementation
Abstract - As the complexity of diesel engines is increasing (post treatment systems, HCCI, ), the need to control accurately the combustion and hence the operating conditions inside the cylinder is becoming more and more important. In this context, the AFR control on diesel engines is becoming crucial. The dispersion between the injectors and between the air masses of each cylinder results in significant air fuel ratios (AFR) differences between the cylinders. In some operating conditions (NOx trap regeneration, ), the discrepancy of AFR between the cylinders can lead to an increase of the engines pollutant emission. The individual measurement of AFR in each cylinder can help manage these phases and improve the global functioning of the engine. For several reasons (cost effectiveness, space, sensor characteristics, ), this control must be based on the measurement given by the UEGO sensor located downstream the turbine.
The work presented in this paper describes the development and the implementation in a rapid prototyping management system of an estimator reconstructing the AFR of each cylinder from the measurement of the UEGO sensor. The estimator is based on a physical model of the exhaust, representing the dynamics of the gas flow from the outlet of the cylinders down to the lambda sensor. From this model, it is then possible, using Kalman filter techniques, to deconvolve the AFR measurement downstream the turbine and estimate the AFR of each cylinder.
The estimator was tested on an engine testbench, proving the feasibility of the solution described. The results show a good correlation between the estimation and a UEGO sensor located in cylinder 1 exhaust pipe.