Optimization of a high performance engine GDI Wet System and its control via virtual analysis and experimental tests

This thesis describes a combined 1-D numerical and experimental analysis of a GDI injection system. The aim is to design the GDI injection system to have the minimum injection variability between injectors. This is possible analysing the rail pressure waves that affect the injections. Thanks to a MATLAB numerical code the results coming from a first calibration attempt were compared with experimental data to verify the reliability of the acquired results. They were calculated after specific tests made on the hydraulic test bench developed at Lamborghini. Different configurations were studied changing the system geometry, such as rail diameter, high pressure pipe length and inlet position inside the rail, flow-restrictor diameter at the end of high pressure pipe. Eight different configurations were analysed and the one that showed the lowest CoV of injection was proposed as a final design. Then a 1-D numerical analysis of the GDI injection system was developed on the chosen configuration with the aim of predicting pressure waves propagation phenomena and the injected mass flow rate. The focus of the 1-D analysis is to verify through the comparison between simulated values and experimental ones if the model predicts accurately the physics of the system, in order to use it on a wider range of operating points. The following step consisted in controlling the GDI pump in MIL environment through a cosimulation between high pressure system model built in GT-ISE and control model made with Simulink. The high pressure control model was developed and validated. The aim was to have a closed loop control of the rail pressure using the same ECU control strategy. The main task of the control is to actuate the angle, respect to pump TDC, the MSV valve had to be closed to make the actual pressure follow the target one. The control model and the simulation in MIL environment are very useful since they help on the precalibration of ECU functions, reducing testing activities.