Date of Award

2022

Document Type

Open Access Dissertation

Degree Name

Doctor of Philosophy in Mechanical Engineering-Engineering Mechanics (PhD)

Administrative Home Department

Department of Mechanical Engineering-Engineering Mechanics

Advisor 1

Mahdi Shahbakhti

Advisor 2

Darrell Robinette

Committee Member 1

Wayne Weaver

Committee Member 2

Tan Chen

Abstract

The presence of backlash and compliance in automotive drivelines can lead to undesirable NVH phenomena known as clunk and shuffle, respectively, which severely deteriorates the drivability of the vehicle. Since the backlash and compliance are design choices for automotive drivelines, they cannot be completely eliminated. Therefore, torque shaping control systems are used to appropriately modulate the torque commands sent to automotive actuators under different operating conditions in order to reduce clunk and shuffle. Model-based controllers provide good control performance without the need for cumbersome calibrations and long development time. Consequently, this dissertation centers on the design and real-time implementation of an optimal, model-based torque shaping control system for mitigating clunk and shuffle from the driveline.

In this PhD dissertation, Kalman filter-based estimation algorithms and soft landing reference governor-based control algorithms are designed to provide the shaped torque commands to the actuators. A high-fidelity plant model and control-oriented models are developed for use in the model-based torque shaping controller, and they are validated using experimental data. Both the high-fidelity plant model and the control-oriented model capture the frequency and phase of the shuffle oscillations with an average error of less than 10%.

For effective performance of the torque shaping controller up-to-date backlash position and backlash size information is needed, and estimators are designed using Kalman filter techniques which make use of readily available driveline sensors to provide this information. The backlash position estimator is shown to be accurate in estimating the plant backlash position with a delay of up to 2 sample periods. The backlash size estimator is demonstrated to estimate the plant backlash size in various operating conditions with an error of less than 10%.

The torque shaping shuffle and clunk control algorithms are designed using a pre-compensator and lead compensator-based feedback controller, and a reference governor-based optimal controller which work using information from existing driveline sensors, and provide a smooth and connected driving experience. This model-based shuffle and clunk controller reduces the number of calibrations by more than 90% compared to a rule-based controller, and is easily implementable on embedded processors. Finally, the impact of model uncertainties on the performance of the shuffle and clunk controller is analyzed, and the controller is shown to be robust in multiple use cases. Additionally, the torque shaping controller is integrated with two separate model-based, optimal controllers. One of the controllers reduces torque lag caused during the transition of the torque converter clutch from locked to slipping condition from 13.5% to 2.1%.

Share

COinS