Nonlinear System Identification of the Filling Phase of a Wet Clutch System

The work presented illustrates how the choice of input perturbation signal and experimental design improves the derived model of the dynamics of a wet clutch system. The relationship between the applied current signal and resulting pressure in the filling phase of the clutch is established based on bandlimited periodic signals applied at different current operating points and signals approximating an ideal filing signal. A polynomial nonlinear state space model (PNLSS) is estimated and validated over a range of measurements and yields better fits over a linear model, while the performance of either model depends on the perturbation signal used for model estimation. A wet clutch is a mechanical device that transmits torque from an input axis to an output axis via fluid friction. An electro-hydraulic pressu regulated proportional valve regulates the pressure inside the clutch which brings about the engagement of the piston with the friction plates. A model describing the relation between the current applied to the valve and the resulting pressure during the filling stage of the clutch is required to bring about a smooth engagement using either iterative learning control or model predictive control.