| Mobility of wheeled vehicles on deformable ground concerns a wide variety of industries, including: agriculture, forestry, mining, recreation, and military. Commercial software is available to model and simulate entire vehicle systems but lacks the ability to accurately simulate mobility on soft soil for most driving maneuvers. This thesis focuses on building a framework to extend vehicle dynamics mobility simulation on deformable terrain to allow for general driving maneuvers. First, a deformable terrain model with a stress/strain relationship informed from soil mechanics is discussed and described in the context of a three-dimensional description of the terrain surface. The second major effort is to quantify the tire/terrain contact patch, both in shape, size and the normal and shear stress distributions, which is used as the boundary condition at the surface of the terrain model. Computational bottlenecks are identified and parallel CPU and Graphics Processor Unit hardware is leveraged using OpenMP and CUDA, respectively, resulting in an order of magnitude speed-up. The new tire/terrain model is validated at two different levels: static load/deflection tests used in the deformable tire model parameter identification, and a tire testing rig in a multibody dynamics simulation engine. In-plane steady-state wheel performance is analyzed: drawbar pull, motion resistance, thrust, torque and the normal and shear stress along the centerline of the contact patch. These tests are run over a range of wheel sizes, weights, inflation pressure and slip rates on firm and loose clay loam soil. Validation is carried out using both experimental data found in the literature, as well as two-dimensional equilibrium semi-empirical methods. It is shown that the newly developed model predicts higher drawbar pull than the two-dimensional methods, but less than the experimentally reported values due to the lack of lugs in the current model. Contact stress distributions are shown to agree well with the semi-empirical methods along the tire centerline, and has a three-dimensional contact profile which agrees well with the experimental results. A dynamically driven wheel is shown to have different contact stress distributions at different wheel velocities, independent of slip, illustrating the need to consider these variables separately. |
