A Reduced Order Plant Model for Shift Quality and Launch Strategy Improvements in Double Clutch Transmissions - A Model-Based Design Approach

摘要

In this paper a reduced order plant model for the development of control strategies, taking into account shift quality improvements in double clutch transmissions, is presented. The simulation model of the complete powertrain, the gearbox and the hydraulic actuation system incompletely built up by blocks from the standard library of MATLAB-Simulink and from the MATLAB-SimMechanics toolbox, The latter toolbox is based on Simscape, the platform product for the Simulink physical modeling family. The main advantage of SimMechanics, which is a kind of a port-based Hamiltonian physical simulation tool, is that the mechanical part of the driveline can be easily built up using inertia and joint blocks, which are connected by flexible elements. For the hydraulic part of the system, the actuation system which controls the pressure of the clutches, the shift forks and the cooling system, black-box modeling techniques are considered, using identification methods for the transfer paths of valve currents to clutches. Only if the dynamic effects due to the torque pulses from the engine, the torsional damper or the finite rotary stiffness of the driveline shafts and the differential gears are properly taken into account, the shift quality and driving comfort can be improved by means of virtual prototyping, what is the aim of this paper. In general, for the functional development and validation of the controller strategy for automatic transmissions, in order to discover errors and failures of the controlling software part, it is sufficient to have a much more simpler simulation model of the vehicle, neglecting these effects. One difficulty arises from the trade-off between the need of a fast virtual prototype running on a workstation and a plant model with sufficient accuracy of the real vehicle which requires the long-time knowledge of experts from the simulation, test and software departments. Such a model-based design approach can reduce the development of the driving strategy process like creep, launch, slip and fade control strategies and therefore shortens developing cycles from prototyping to serial production.