A Phenomenological Combustion Model Including In-Cylinder Pollutants To Support Engine Control Optimisation Under Transient Conditions 2011-01-1837

Regulations in terms of pollutant emissions are becoming more and more constraining. The car manufacturers need to adopt a global optimisation approach of engine and exhaust after-treatment systems. An engine architecture definition coupled to an adapted control strategy seem to be an ideal way to address this issue. The problem is particularly complex, considering the trade off between the drivability which must be maintained, the reduction of the in-cylinder pollutant emissions, the reduction of the fuel consumption and the optimisation of the operating conditions to reach high conversion efficiencies via exhaust gas after-treatment systems. Sophisticated control strategies and models can only be developed with a complete understanding of the physical phenomena occurring in the combustion chamber, thanks to experimental measurements and engine system simulations. In this context, 0D predictive models of combustion and pollutant emissions, calibrated with experimental data on steady state operating conditions, are particularly interesting to perform a wide range of engine settings and configuration variations. Indeed, these variations are difficult and expensive to perform at the testbed. Thus, this paper presents the validation of a developed combustion and pollutant emissions model in several operating conditions. As a first step, the numerical results concerning the cylinder pressures and the pollutant emissions (NO, CO, unburnt hydrocarbons and soot particles) are compared to experimental measurements in the complete range of engine speeds and loads, using a single cylinder simulator. Secondly, a variation of load at a constant engine speed is simulated thanks to a four-cylinder simulator including the whole airpath system. The results are then presented focusing on the pollutant emissions predictions of the simulator. Finally, a simulation of the NEDC cycle is proposed, thanks to a modelled version of the engine control and the vehicle. These simulations are performed under Real-Time conditions thanks to the xMOD environment [1].