In turbo systems, exhaust gas energy is used to compress intake air, and the amount of air flowing into the combustion chamber is controlled by the boost pressure. The boost pressure is regulated by the wastegate valve, which controls the intake air supplied to the engine by the compressor. However, the turbocharger compressor has a hardware limit on the number of rotations due to its high temperature and speed. In conventional turbo control, the limit is indirectly controlled by the measurable boost pressure, rather than the hardware limit value of the turbo speed. As a result, in low-pressure conditions, such as high altitudes, more turbo revolutions are required for the same target boost pressure, which increases the calibration labor for each environmental condition and causes engine power loss. This paper proposes improving engine output by predicting turbo speed accurately using sensors and model values available in EMS, along with boost pressure. Model predictive control (MPC) is used to perform an optimal control of the wastegate opening to prevent the maximum turbo speed from being exceeded during transients by limiting the conventional target boost pressure through a reduction in the turbo speed limit value margin. In conventional turbocharger boost pressure control, the pilot control determines the feedback WG opening quantity based on the equivalent relationship between the compressor in a steady state and the turbine power, and the boost pressure is directly feedback controlled via the WG opening quantity through a PID controller to improve real-time boost pressure tracking control performance. In this paper's Virtual Turbospeed Control (VTC) utilizing Model Predictive Control, the highly nonlinear relationship between boost pressure and WG opening quantity is transformed into a linear model based on energy variation using turbospeed. The WG opening quantity is adjusted according to the objective function defined by considering the control and constraint conditions for the allowable turbospeed range for the target boost pressure of the turbocharger with explicit consideration of constraints. Therefore, the MPC controller provides better performance than conventional turbo control algorithms and is robust to changes in operating conditions and various types of disturbances. To calculate the QP solution for such an objective function, implementation of an optimization solver is required in ECU. In the first project, the 1.6-liter Turbo engine was selected, and QP solver algorithm implementation was carried out in collaboration with Garrett Motion for ECU mass production logic development. Task division optimization design was achieved considering CPU computational load, execution time, and memory for integrated application of the VTC logic into our own EMS mass-produced ECU. As a result, VTC logic of this paper was first applied to i20N project (SOP '21.4), securing the potential for increasing engine output by reducing the turbocharger speed limit margin, and enhancing the Over-speed protection function through MPC application.

Mr. Minkyu Han, Senior Research Engineer, Hyundai Motors