Date of Award


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

Committee Member 2

Jeremy Worm


Drivability is an important metric during the development of an automobile. Calibration engineers spend a significant amount of time trying to improve the drivability of vehicles for various driving conditions. With an increase in the available computational power in an automobile, novel model-based methods are being implemented for further improving the drivability, while reducing calibration time and effort. Phenomenon known as clunk and shuffle, which are caused due to backlash and compliance in the driveline, are a major cause of issues related to drivability and noise, vibration and harshness (NVH) during tip-in and tip-out scenarios.

This thesis focuses on developing a high-fidelity, control-oriented vehicle driveline model, which can be used for developing systems, to improve the drivability of a vehicle, during tip-in and tip-out events. A first principle physics-based model is developed, which includes the engine as a torque generator, backlash elements as discontinuities, and driveshafts as compliant elements. Experimental validation results showed that the accuracy of the developed model, in representing shuffle oscillation frequency, during the tip-in scenarios, with locked torque converter clutch, is approximately 99 %.

A parametric analysis is performed to characterize the behavior of the model during different input conditions, and to study the effect of backlash size, and driveshaft compliance on the response of the driveline. Based on the observations from the parametric analysis, the high-fidelity model is later condensed into a reduced-order model, and comparative analysis is carried out between two reduced-order model (ROM) designs. The comparative results between the full-order model and ROM show that the ROM with separate tire parameters is better in predicting the frequency and amplitude of shuffle oscillations during tip-in events.