Co-Simulation of Multi-Domain Engine and Integrated Control Model for Transient Predictive Driving Cycle Simulation

Over the recent years, powertrain development has been characterized by a significant rise to face with the more restrictive emissions limitations. Despite the harsh criticism that the public opinion is levelling at Diesel engine, it has been involved into this rush to technological innovation. The vehicle industry is a competitive business and Original Equipment Manufacturers (OEMs) are forced to invest in innovative strategies with high complexity level and to study a wide range of working conditions, paying attention to keep the costs to a minimum. To achieve these targets, as well as accuracy, safety and repeatability of the tests, virtualization is the surely one the best ways to be followed, and modelling is an increasingly important field for research and development [2,3]. Front-loading the conventional vehicle testing to engine test benches or even to Model-in-the-Loop (MiL) and Hardware-in-the-Loop (HiL) simulations is an approach that, by using virtual components, seeks for a seamless system integration and testing process. This concept so called Road-to-Rig-to-Desktop (R2R2D) is challenged by the search for an optimum between results accuracy, calibration effort reduction and better computing power. Moreover, system optimization in transient driving condition during concept study and function development are two other advantages of R2R2D approach [4]. A Mean Value Engine Model (MVEM) and a Fast-Running Model (FRM) of a two-liter, single turbocharged Diesel engine, were the focus of the conducted research. Both engine models have been calibrated against experimental data and used to predict engine behavior over the complete operational driving cycle. The Real Time (RT) capability of both models was proved in Hardware-in-the-Loop applications [2,3,5,6]. For this thesis, the existing FRM has been optimized for MiL applications. To evaluate Engine Out (EO) NOx emissions, the FRM has been coupled with a semi-physical Simulink-based emission model. The interface between the models has been defined and the connecting points in the FRM have been optimized in order to produce consistent simulation results compared with measurement data. Lately, a generic air-path control model has been adapted for coupled simulation with the engine model. At first, open-loop approach has been followed, then a methodology has been developed for closed-loop simulations. In this way, feedback signals can compensate disturbance that may affect the system to grant the maintaining of the desired condition. The model integration has been evaluated on different simulation environments. The FRM is investigated for its potential to be more flexible with the specific use case of Cylinder Deactivation (CDA). The dynamic change of the number of cylinders firing, and the response of the air-path controller have been investigated. In particular, the behavior of the high-pressure and low-pressure exhaust gas recirculation loops and their influence on emissions and EO temperatures has been highlighted. Furthermore, a platform containing both MVEM and FRM coupled with a physics-based air-path control model has been validated and the results of the simulations have been compared against measurements data. In this heterogeneous multi-domain engine platform, the two different modeling approaches have been evaluated, based on the conducted MiL simulation, with a focus on the performance in terms of transient response and emissions of the investigated engine.