Repository logo

Browsing by Author "Bindner, Joseph R., author"

Now showing 1 - 2 of 2
  • Thumbnail Image
    ItemEmbargo
    Improving soil property predictions for applications in tailings and terramechanics
    (Colorado State University. Libraries, 2024) Bindner, Joseph R., author; Scalia, Joseph, advisor; Atadero, Rebecca, advisor; Bareither, Christopher, committee member; Niemann, Jeffrey, committee member; Ham, Jay, committee member
    Soil properties are used by engineers and scientists to better understand the state and behavior of soils. For example, soil properties can be used to estimate surficial soil strength for vehicle mobility models and can be used to better understand the engineering characteristics of mine waste (tailings) stored in tailings storage facilities. Soil and tailings properties often have high spatial variability and often require high resolution data for engineering analyses. Standard laboratory procedures are commonly used to determine soil properties but are often impractical for large spatial extents. While some existing soil data products provide estimates of surficial soil properties, the fidelity of soil data products is often poorly understood and insufficient for many applications. Additionally, some field tests used to estimate soil properties, such as the cone penetration test (CPT), rely on empirical correlations that cannot be used for some soils. There remains a need for procedures which improve the speed and accuracy of soil property estimates across large spatial extents. The objectives of this study are to (i) evaluate how surficial soil moisture and soil strength vary with soil and landscape attributes across a large spatial extent, (ii) explore the use of field-based hyperspectral sensing and machine learning for the prediction of surficial soil properties across a landscape, and (iii) assess the use of laboratory hyperspectral sensing and machine learning for the prediction of tailings properties for potential application in situ via direct push methods. Soil and landscape attributes were determined at sampling locations across a semi-arid foothills region and used to assess how soil moisture and soil strength vary with soil and landscape attributes. Then, hyperspectral data were captured at select sampling locations and used to train and assess the performance of a convolutional neural network (CNN) for the predictions of soil properties. Finally, a diverse tailings-hyperspectral dataset was prepared in the lab and used to train and assess a CNN to provide proof of concepts for prediction of material properties relevant to TSF stability analyses.
  • Thumbnail Image
    ItemOpen Access
    Procedure for measurement of surficial soil strength via bevameter
    (Colorado State University. Libraries, 2020) Bindner, Joseph R., author; Scalia, Joseph, IV, advisor; Niemann, Jeffrey D., advisor; Butters, Gregory, committee member; Green, Timothy R., committee member
    Spatial prediction of moisture-variable soil strength is critical for forecasting the trafficability of vehicles across terrain. The Strength of Surface Soils (STRESS) model calculates soil strength properties as a function of soil texture from SSURGO data (or locally available data) and soil moisture from the Equilibrium Moisture from Topography, Vegetation, and Soil (EMT+VS) model. The STRESS model yields soil strength properties (friction angle and moisture-variable cohesion) that vary with soil texture and moisture conditions. However, the STRESS model is hindered by a lack of surficial soil strength data linked directly to soil texture. The objective of this study is to develop and validate a bevameter procedure to improve measurement of near-surface moisture-variable soil strength. The bevameter is a test apparatus that measures in-situ surficial soil strength properties by rotational shearing of a shear annulus under a constant normal force at a constant rate. The bevameter allows for lab or field determination of Mohr-Coulomb surficial soil strength properties at a given moisture content in a manner that approximates how vehicles interact with surficial soils. Experimental variables evaluated include the shearing surface (grousers, sandpaper, or bonded angular sand) and the use of interior and exterior annular surcharge weights to minimize slip sinkage of the shear annulus. Based on the results of this study, a bevameter procedure is recommended that uses a coarse sandpaper as the shear interface with an internal and external surcharge of 2 kPa during shear testing. Using the revised bevameter procedure for field testing, the performance of predicted moisture-variable soil strength by the STRESS model is evaluated. Field validation illustrates the need to develop surficial-soil specific pedotransfer functions for use in the STRESS model.