Model Predictive Control of an Automotive Driveline for Optimal Torque Delivery with Minimal Oscillations during Torque Converter Slipping Conditions

During certain driving scenarios, low-speed engine vibrations get propagated to the driveline and affect the drivability of a vehicle. To reduce the impact of these vibrations, a locked torque converter lockup clutch (TCC) is allowed to temporarily slip to increase the damping in the driveline. However, the initial slow dynamics of the fluid path of the torque converter cause the vehicle to feel sluggish. In this article, we design a model predictive controller (MPC) that optimally controls the torque request from the actuator (i.e., engine or e-motor) and the lockup clutch capacity for reducing this sluggishness. The study is conducted for a light-duty vehicle and uses an experimentally validated, detailed full-order model (FOM) for developing and validating a computationally efficient, reduced-order driveline model (ROM). The ROM includes the nonlinear dynamics of the torque converter’s hydraulic coupling, and the drivetrain response of the ROM is within 2.4% of the FOM’s response. The designed controller makes use of an electronic control unit (ECU)-estimated reference turbine torque command, measured actuator, turbine, and wheel speed signals as inputs and provides actuator torque command and TCC capacity command as outputs. We validate the controller’s performance using model-in-the-loop (MIL) and processor-in-the-loop (PIL) experiments. The results show the designed controller overcomes the torque lag at the propeller shaft by 83%. We also verify the robustness of the designed controller to various cases of torque converter’s transient fluid dynamics and multiple clutch slip initiations. We observe that the controller provides the desired response of the drivetrain with a maximum error of 4.1% in the delivered propeller shaft torque when compared with the locked TCC response.