Date of Award

2019

Document Type

Open Access Master's Thesis

Degree Name

Master of Science in Mechanical Engineering (MS)

Administrative Home Department

Department of Mechanical Engineering-Engineering Mechanics

Advisor 1

Mahdi Shahbakhti

Advisor 2

Darrell Robinette

Committee Member 1

Jason Blough

Abstract

Vehicle drivability is an important factor which more and more customers have started assessing before buying a vehicle. Customers carry out this assessment based on both vehicle reviews/ratings and based on the test drives. One of common maneuver which a customers perform during the test drive is sudden accelerator pedal tip-in or tip-out to accelerate or coast the vehicle. Clunk and shuffle are the phenomena that usually occur during this scenario causing driver discomfort. The clunk and shuffle are caused by the backlash and compliance physical properties of the driveline. Consequently, control strategy needs to be developed which can provide a fast driveline response without clunk and shuffle. One major input to develop a control strategy is the knowledge of the vehicle states and parameters based on the available measurements, which is the major focus of this work.

This work begins with a discussion of various existing estimation strategies that have been used to estimate the states of vehicle along with their merits and demerits. Then a full order model, developed in the previous works, is validated for a locked torque converter case along with its reduced order model which is used for estimator development. The error in the simulated shuffle frequency for the full order model and reduced order model is less than 1%. The reduced order model is then used to develop an observable state space model to estimate the backlash state and size of the model. The estimators developed are validated and the robustness analysis is done for different scenarios of torque inputs, delays and sampling times. It is found that the sampling time of the estimators and measurement inputs significantly effect the estimates of lash traversal time and backlash size with a mean error of 9% in lash traversal time estimate and 2% error in lash size for 10 ms sampling time. Furthermore, the estimators are found to be more robust to the variations in the wheel speed measurements as compared to variations in the engine speed measurements.

Available for download on Thursday, April 30, 2020

Share

COinS