Control of an electromechanical clutch actuator by a parallel Adaptive Feedforward and Bang-Bang controller: Simulation and Experimental results

Abstract Vehicle’s powertrain performance and comfort are largely depending on the clutch control strategy for an Automated Manual Transmission (AMT). The aim of the clutch control strategy is to ensure the smooth running of clutch operational cases: a comfortable clutch launch (vehicle takes off smoothly without jerk), a fast upshift/downshift (gear ratio changes) and a fast clutch opening. In most industrial cases, regardless of clutch actuation technology, clutch control is managed by clutch pressure control. However, the clutch pressure control is a challenge regarding clutch non-linearities and time-varying parameters. In this paper, a parallel adaptive feedforward and bang-bang controller is proposed in order to control the clutch pressure with an electromechanical clutch actuator. In this system, a control issue comes from potential time-varying parameters but the main challenge comes from the hysteretic behavior of the system due to dry friction in the actuator assembly. An analytic model of the clutch and its electromechanical actuator including dry friction has been constructed and a prototype has been designed and integrated on a test bench. The parallel adaptive feedforward and bang-bang controller architecture and algorithms are developed. For three critical clutch operational cases, simulations and experiments have been run. Finally, in spite of time-varying parameters and a high hysteretic system behavior, simulation and experimental control results highlight that the controller allows a precise tracking of the pressure reference and a fast time response.