Multidisciplinary Simulation Model for the Balancing of Powertrain Combustion, Control and Components for Optimal Fuel Consumption, Emissions, Cost and Performance for a Diesel Engine Powered Passenger Car 2012-01-1572

Passenger cars equipped with diesel engines will meet challenging emission legislation for the coming decade, with introduction of Euro6 and Euro7, which comprises reduced NO_X emissions and possibly new driving cycles including off-cycle limits. The technology measures to meet these legislative limits comprise a broad spectrum of engine and aftertreatment, i.e., engine measures such as improved fuel injection with respect to mass and timing, improved exhaust gas recirculation, improved warm-up and reduced friction, as well as aftertreatment measures such as selective catalytic reduction and lean NO_X trap in combination with diesel particulate filter, and the thereby associated engine control. The resulting technology matrix is therefore large, and calls for a multidisciplinary simulation approach for appropriate selection and optimization of technology and control with the objectives and constraints of emissions, fuel consumption, performance and cost.

The idea behind multidisciplinary simulation is to include all subcomponents of the powertrain into a complete system simulation model, in order to study the influence of and interaction between subsystems on emissions, fuel consumption and cost of the complete powertrain. This approach requires simplified models on a subsystem level for reasonable simulation times on a system level. The subcomponents models mainly consist of a vehicle and road load model, engine model with tunable EGR and aftertreatment models for the appropriate system to be considered. The complete system model, which is programmed in MATLAB Simulink, with variability in component choice and engine calibration, is then coupled to an external optimization routine in order to find the optimal combination in terms of the objectives stated above.

The core of system simulation model consists of a semi-empirical engine model based on engine bench test data from a design of experiment, where appropriate engine operating parameters are varied to span the possible engine operating states we expect. From this, a mathematical model is derived where the input is defined as operating state such as engine speed, load, EGR-ratio, coolant temperature, intake temperature, boost pressure, etc., and the results are fuel consumption, emissions, exhaust temperature and mass flow, etc. Furthermore, the model includes full physical models of the aftertreatment system including temperature distribution along the exhaust line. The simplified nature of the models results in fast execution times of roughly 1/100 of real time.