This paper proposes a model-based gain scheduling strategy of a Skogestad internal model control (SIMC)-based boost pressure controller for passenger car diesel engines. This gain scheduling strategy is proposed with a new scheduling variable to handle the nonlinear variable geometric turbocharger (VGT) plant characteristics. The scheduling variable is derived from the pressure ratio between the exhaust and intake manifolds and the exhaust air-to-fuel ratio to estimate the static gain of the VGT plant, which varies widely with change in the engine operating conditions. The proposed static gain model was designed with the scheduling variable, engine speed, and fuel injection quantity. Compared to the steady-state experimental data, the static gain model showed an R-squared value of 0.91. The boost pressure controller had the proportional-integral (PI) structure to allow for online calibration, and the PI gains were determined using the SIMC method. The proposed static gain model for the VGT plant was integrated into the SIMC control structure to obtain the appropriate control gains under wide engine operating area. The proposed control algorithm was compared with a fixed gain boost pressure controller through various step tests of the desired boost pressure. The fixed gain controller showed a large overshoot of 64% when the exhaust gas recirculation (EGR) operating condition was changed. In contrast, the proposed gain scheduled boost pressure controller reduced the overshoot to 12%. The model-based gain scheduling strategy successfully adjusted the control gains to achieve consistent control performance under various engine operating conditions.
Composite Adaptive Internal Model Control and Its Application to Boost Pressure Control of a Turbocharged Gasoline Engine
