Joško Petrić

Modeling of Wet Clutch Engagement Including a Thorough Experimental Validation

A detailed experimental validation has been carried out to point to limitations of static wet-clutch friction model for typical clutch engagement transients. The model accuracy can be increased by incorporating the fluid film dynamics, as done in the lumped-parameter dynamic clutch model developed at the University of Purdue. That model is extended herein in order to increase its accuracy especially in the case of grooved clutches. The extensions include a description of clutch actuator dynamics and introduction of an empirical scaling factor for the fluid film thickness state equation. More rigorous treatment of fluid dynamics for the grooved clutch is also presented.

Recent Advances in Control-Oriented Modeling of Automotive Power Train Dynamics

This paper presents a survey of the recent research results of the authors in the field of modeling of automotive power train systems and components. The goal of the research is to propose simple and accurate power train models for controller design and to propose computationally efficient simulations. The modeling includes typical power train components such as electronic throttle, SI engine, torque converter, planetary gear set, wet clutch, differential, half shaft, and tire. Experimental model validation results are presented

Bond graph modelling and power flow analysis of series–parallel HEV transmissions

The bond graph method is used to model kinematics and dynamics of various one-mode and two-mode series–parallel configurations of Hybrid Electric Vehicle (HEV) transmissions. The bond graph method is also used to analyse the power flow of the considered transmissions in electric variator and full-hybrid operating modes. The analysis is illustrated by numerical static curves related to the maximum transmission output torque for a wide range of vehicle velocities.

Bond graph modeling of series-parallel hybrid electric vehicle power train dynamics

The bond graph method is used to model dynamics of various one-mode and two-mode series-parallel configurations of hybrid electric vehicle power trains. Proper handling of causalities results in minimum-realization models convenient for various simulation, analysis, and control design studies. The same, unified model structure can be used to describe the dynamics of all considered hybrid power trains. Several alternative variants of the power train dynamics model are also derived for various structures of driveline submodel.

A Control-Oriented Polytropic Model of SI Engine Intake Manifold

The paper presents experimental results which show significant changes of the intake manifold air temperature during fast tip-in/tip-out engine transients. An adequate two-state polytropic manifold model is developed and experimentally validated. An emphasis is on the derivation and parameterization of a time-variant structure of the heat transfer coefficient. The polytropic manifold model is extended to a three-state form for the more general case of different heat transfer properties for the manifold plenum and runners. An influence of the engine back flow on the runner thermal transients is observed. A simple extension of the three-state model with the back flow effect is proposed.Copyright

Bond Graph Modeling and Analysis of Series-Parallel Hybrid Electric Vehicle Transmissions

The bond graph method is used to model kinematics of various one-mode and two-mode series-parallel configurations of hybrid electric vehicle transmissions. Based on the derived speed and torque equations, a comparative analysis of hybrid transmissions steady-state behaviors is conducted. An example of control-oriented bond graph modeling of hybrid ransmission dynamics is presented, as well.

Pneumatic Inverted Wedge

Abstract The paper describes the inverted wedge actuated by the pneumatic cylinder. It can be considered as a two degrees-of-freedom planar robot, which is controlled by a single control input (movement of the slider) in order to keep the frame of the wedge in balance. Only one measured variable (angle of the frame) was anticipated, hence an observer was necessary to control this model using state controller. The nonminimum-phase characteristic and pneumatics make control of described system even more challenging. Thus, this model offers great possibilities as an experimental device in control education.

Optimization of control variables of a series-parallel hybrid electric power train

A back propagation through time-like conjugate gradient-based algorithm is used for off-line optimization of a series-parallel hybrid power train control variables over standardized certification driving cycles. The aim is to minimize the fuel consumption subject to battery state-of-charge constraints. The optimization is carried out using a simplified power train model in order to find optimal engine speed and torque time responses. The optimization results are presented for three characteristic certification driving cycles.

A pneumatically actuated ball and beam system

A ball and beam balancing mechanism actuated by a pneumatic rotary drive is presented in this paper. A laboratory model of the pneumatically actuated ball and beam system has been developed within student projects during courses related to mechatronics and automatic control. A mathematical model of the nonlinear system is developed using Lagranges equations. The model is linearized at the equilibrium point, and the linear model has been implemented for the design of control algorithms. The cascade proportional derivative (PD) controller and the optimal linear quadratic regulator have been applied and experimentally verified. The utilization of a pneumatic actuator introduces additional dynamic terms and nonlinear effects, and that makes the balancing mechanism even more difficult to control. The methods discussed in this paper have good potential application to other control techniques and offer great possibilities for giving students a better understanding of the theoretical and practical aspects of nonlinear system control.

Design of EKF-based SoC estimator for an ultracapacitor module

This paper presents the design of ultracapacitor module state-of-charge (SoC) estimator based on an Extended Kalman Filter (EKF). A lumped-parameter resistive-capacitive (RC) electrical circuit representation of ultracapacitor module has been used for EKF design, wherein non-linear capacitance effects have been included through SoC-dependent capacitance term, while cell balancing effects were modeled by a second-order disturbance model. The proposed EKF-based SoC estimator has been compared to an LTI model-based Kalman filter as SoC estimator, and verified by means of computer simulations and experimentally on the dedicated ultracapacitor test setup.