Simulation of CO and NO emissions in a SI engine using a 0D coherent flame model coupled with a tabulated chemistry approach

Abstract

Environmental issues stimulate the elaboration of new powertrain systems and fuels for transport as an essential priority to decrease air pollution and green house gases emissions. Developments ranging from architecture definition to engine control and calibration are today increasingly performed using complete vehicle simulators running close to real times. The challenge for engineers is therefore to develop models able to accurately reproduce the engine response without altering the CPU efficiency of the simulator. For this purpose, 0-dimensional models are commonly used to describe combustion processes in engine combustion chambers. This paper extends a 0-dimensional coherent flame model (CFM), called CFM1D, to incorporate chemical effects related to the fuel composition and thermodynamic conditions at low computational costs. Improvements are carried out integrating the NO relaxation approach (NORA) based on a priori homogeneous reactor computations and initially developed for 3D simulations to describe post-oxidation processes in the burnt gases. In this work, this method is extended to the modeling of CO production and oxidation leading to the CORA (CO Relaxation Approach) model. Both NO and CO reaction rates are therefore written as linear relaxations towards their equilibrium mass fraction values Ykeq (where k stands for NO or CO) within a characteristic time τk. In this approach, Ykeq and τk are tabulated as functions of equivalence ratio, fresh gases dilution rate by burnt gases, pressure and enthalpy. The resulting new model, called CFM1D-TC (CFM1D-Tabulated Chemistry), is then used to perform simulations on a large range of operating conditions of a spark-ignition engine burning methane-air-diluent mixtures. Comparisons are made with experiments and with simulations performed using a classical reduced chemical scheme in the burnt gases. The achieved results evidence the interest in terms of accuracy and CPU-efficiency of this new approach to describe post-flame processes in spark ignition engines.