Abstract
Unmanned vehicles and robotic systems are finding their way to the fields of modern agriculture. The company, Agro Intelligence ApS is enabling unmanned agricultural field operations by researching and developing an autonomous robotic vehicle designed for arable farming.
The performance of the vehicle is affected by the environment where it operates. By considering unmanned agricultural operations on a wide geographical scale, various combinations of soils and tools are to be considered in the design and operation of such vehicle. This requires an efficient method for computing loads and evaluating site-specific performance. This is addressed throughout the thesis using physics-based modeling and simulation of the soil-machine interaction.
The interaction between machine and soil is described by three perspectives that define the soil--machine interaction system consisting of soil, tool and machine dynamics. This study initiates from the first perspective, the soil, by the development of an approach to visually evaluating the soil surface micro-topography due to a tillage operation. To understand the subsurface soil disturbance, a study of physics-based modeling of tillage and soil--tool interaction is presented. This is described via the second perspective, the soil--tool interaction. The third perspective is the machine dynamics where the two former mentioned perspectives are consolidated in simulating the dynamics of an unmanned robotic system in three levels of fidelity.
To model and simulate the interaction between soil and the machine, an open-source multi-physics code 'Chrono' is utilized as the underlying simulation environment. The tools and the vehicle are modeled as a multi-body dynamics system of rigid bodies and the soil is modeled using the Discrete Element Method (DEM).
Configuring numerical models of soil such that the model reflects different soil types and soil-mechanical properties are challenging. A method is developed for identifying engineering parameters based on the soil shear strength properties. This method is tested in three soil types by computing the soil-reaction forces of a rectangular blade-sweep. Additionally, a study of DEM-based soil aggregate fracture is presented and the results are compared to previously published data.
Besides modeling the robotic system, the maximum drawbar pull was measured experimentally and a Functional Mock-up Unit of the vehicle dynamics model was developed for conducting co-simulation via the Functional Mock-up Interface.
Each of the three perspectives of the soil--machine interaction system is described in a dedicated chapter. Finally, the findings of this Industrial PhD project are summed up in a conclusion and suggestions for future work and perspectives are proposed.
The performance of the vehicle is affected by the environment where it operates. By considering unmanned agricultural operations on a wide geographical scale, various combinations of soils and tools are to be considered in the design and operation of such vehicle. This requires an efficient method for computing loads and evaluating site-specific performance. This is addressed throughout the thesis using physics-based modeling and simulation of the soil-machine interaction.
The interaction between machine and soil is described by three perspectives that define the soil--machine interaction system consisting of soil, tool and machine dynamics. This study initiates from the first perspective, the soil, by the development of an approach to visually evaluating the soil surface micro-topography due to a tillage operation. To understand the subsurface soil disturbance, a study of physics-based modeling of tillage and soil--tool interaction is presented. This is described via the second perspective, the soil--tool interaction. The third perspective is the machine dynamics where the two former mentioned perspectives are consolidated in simulating the dynamics of an unmanned robotic system in three levels of fidelity.
To model and simulate the interaction between soil and the machine, an open-source multi-physics code 'Chrono' is utilized as the underlying simulation environment. The tools and the vehicle are modeled as a multi-body dynamics system of rigid bodies and the soil is modeled using the Discrete Element Method (DEM).
Configuring numerical models of soil such that the model reflects different soil types and soil-mechanical properties are challenging. A method is developed for identifying engineering parameters based on the soil shear strength properties. This method is tested in three soil types by computing the soil-reaction forces of a rectangular blade-sweep. Additionally, a study of DEM-based soil aggregate fracture is presented and the results are compared to previously published data.
Besides modeling the robotic system, the maximum drawbar pull was measured experimentally and a Functional Mock-up Unit of the vehicle dynamics model was developed for conducting co-simulation via the Functional Mock-up Interface.
Each of the three perspectives of the soil--machine interaction system is described in a dedicated chapter. Finally, the findings of this Industrial PhD project are summed up in a conclusion and suggestions for future work and perspectives are proposed.
Bidragets oversatte titel | Fysikbaseret modellering og simulering af jord-maskine interaktion |
---|---|
Originalsprog | Engelsk |
Forlag | Århus Universitet |
---|---|
Status | Udgivet - aug. 2021 |