Robust oxygen fraction estimation for conventional and premixed charge compression ignition engines with variable valve actuation

Abstract

In-cylinder oxygen fraction serves as a critical control input to advanced combustion strategies, but is extremely difficult to measure on production engines. Fortunately, the in-cylinder oxygen levels can be estimated based on accurate estimates or measurements of the oxygen fraction in the intake and exhaust manifolds, the in-cylinder charge mass, and the residual mass. This paper outlines such a physically based, generalizable strategy to estimate the in-cylinder oxygen fraction from only production viable measurements or estimates of exhaust oxygen fraction, fresh air flow, charge flow, fuel flow, turbine flow and EGR flow. While several of these flows are accurately measured or estimated, significant errors in the turbine and EGR flows are commonly observed and can highly degrade the accuracy of any calculations which utilize these flows. An EGR flow estimator was developed to improve the accuracy of this flow measurement over the stock engine control module (ECM) method and is detailed in this paper. Furthermore, the in-cylinder oxygen estimation algorithm is developed, and proven, to be robust to turbine flow errors. Regulation of in-cylinder oxygen levels is of interest for not only in conventional combustion modes but also in advanced combustion strategies such as premixed charge compression ignition. The proposed oxygen fraction estimator is designed such that its performance and stability is ensured in both conventional and advanced combustion modes. The model-based observer estimates the oxygen fractions to be within 0.5% O2 and is shown to have exponential estimator error convergence with a time constant less than 0.05s, even with turbine flow errors of up to 25%.