Data Driven Estimation of Exhaust Manifold Pressure by Use of In-cylinder Pressure Information 2013-01-1749

Although the application of cylinder pressure sensors to gain insight into the combustion process is not a novel topic itself, the recent availability of inexpensive in-cylinder pressure sensors has again prompted an upcoming interest for the utilization of the cylinder pressure signal within engine control and monitoring. Besides the use of the in-cylinder pressure signal for combustion analysis and control the information can also be used to determine related quantities in the exhaust or intake manifold. Within this work two different methods to estimate the pressure inside the exhaust manifold are proposed and compared. In contrary to first principle based approaches, which may require time extensive parameterization, alternative data driven approaches were pursued. In the first method a Principle Component Analysis (PCA) is applied to extract the cylinder pressure information and combined with a polynomial model approach. In the second method an approach which uses the full pressure trace over the crank angle, also in combination with a polynomial model, is applied.

Experimental validation on a Common Rail Diesel engine in different operating conditions led to promising results, e.g. with the PCA approach the exhaust manifold pressure estimation achieved an average error of less than 1% over a wide operating range.