Date of Award

5-2013

Document Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

Legacy Department

Mechanical Engineering

Advisor

Biggers, Sherrill B.

Committee Member

Joseph , Paul F.

Committee Member

Summers , Joshua D.

Committee Member

Li , Gang

Committee Member

Ravichandran , Nadarajah

Abstract

The problem of wheel performance on deformable soil has been studied for many years, but prior to the rise of computational mechanics, such investigations have been limited to development of analytical and empirical models, as well as experimental research. Such models have merit but are necessarily highly idealized and are limited in their applications. Today, many computational models have been implemented for a wide variety of wheel/soil applications.
For the specific case of sandy (i.e. non-cohesive) soils, in terms of the soil's physics the Discrete Element Method (DEM) provides arguably the most realistic model. In DEM, each element represents a single grain of soil (ideally), or may represent a group of soil particles moving together if necessary. A survey of the literature quickly reveals that DEM is computationally intensive and that a great deal of computational effort is normally spent calibrating the DEM model parameters to the desired characteristics of the soil of interest. The goal of this research was to develop and validate an approach to calibration that would require fewer resources, leaving more resources available for solving the problems of interest. This goal was realized through the collection of data over a range of values for each of five simulation parameters using two-dimensional simulations of the Direct Shear Test. By statistical processes the data was used to develop a set of equations that estimate the properties of interest for the simulated soil (small- and large-strain friction angles), based on the simulation parameters. The equations were used to calibrate a two-dimensional rigid wheel/soil simulation of the Wheel Endurance and Sand Traction Merry-Go-Round System (WEST-MGRS). The calibrated model was found to accurately predict the relative performance between a variety of configurations of grousers on actual wheels operating in sand using WEST-MGRS. Therefore, this research shows that the model can be used as a tool to compare the tractive performance of potential designs.
A question that must be answered regarding experiments and simulations of the wheel/soil problem is the question of soil dimensions. Whether simulated or experimental, the system must use a soil container large enough to approximate a semi-infinite soil domain. This research expanded on previous work that had proposed a method for sizing soil dimensions in a dynamic 3-D finite element model. With minor modifications, the method was found to be effective for a wide range of wheel loads and geometries, as well as soil types.

Share

COinS