Data Associated with A CPT Database for Multiple Thin-Layer Correction Procedure Development

------------------------------------------------------------------------------

CONTENTS AND METHODOLOGY

CPT Database

This cone penetrometer test (CPT) dataset comprises laboratory-based and numerically-generated data collected in highly interlayered soil profiles. Both the measured tip resistance in the interlayered profile and the true tip resistance (that would be measured in the profile absent of multiple thin-layer effects) are provided for 49 CPT profiles. All tip resistance values have been normalized for overburden pressure and represent q_c1n values.

The CPT data is presented in two .csv files. The file "qt.csv" contains true tip resistance, q^t, associated with each of the 49 profiles. The file "qm.csv" contains the measured tip resistance, q^m, associated with each of the 49 profiles.

Each .csv file contains 50 rows. The first column of each file contains a "profile category" number. This identifies CPTs with the same soil profile geometry (i.e., CPTs with the same soil profile geometry are assigned the same profile category number). The first row of each file contains the depths associated with each tip resistance value. The first cell in the first row is "NAN" so that depths align with tip resistance values in the underlying rows.  

Rows 2 through 20 contain laboratory data. The laboratory data are derived from a series of calibration chamber tests performed at Deltares (details of the experiments are described in De Lange 2018). Details about how the laboratory data were processed to create the q^m-q^t profiles in this dataset are provided in Yost et al. (2022) and Yost (2022).

Rows 21 through 50 contain numerical data. The numerical data were generated using a 2D axisymmetric implementation of the Material Point Method with a slightly modified version of the open source software Anura3D (anura3d.com). Complete details regarding the calibration and validation of the numerical model used to generate this data are provided in Yost et al. (2022) and Yost (2022).

Jupyter Notebook

The Jupyter notebook "CPT-database-companion-notebook.ipynb" reads the data and performs several operations intended to prepare the data for statistical learning applications, including:

• A script to read data in and initialize depth, q^m, q^t, and profile category variables
• A script to visualize the contents of the database
• A framework to parse the database into separate training and test datasets, grouping profiles with same profile category number together 
• A script to visualize the parsed training and testing data
• An example of how to manipulate data with subsampling and filtering to change the size of the depth interval
• An example of “chunking” profiles to generate more data to use in training and test datasets
• A framework for how to add data to the existing database
• An example of using a profile from the database to assess a simple blurring filter

------------------------------------------------------------------------------

MANUSCRIPTS ACCOMPANYING THIS DATASET

Yost, K.M., Yerro, A., Martin, E. Green, R.A. (in preparation). "A  CPT Database for Multiple Thin-Layer Correction Procedure Development." In preparation for submission to Earthquake Spectra. 

Yost, K.M. (2022). "Improving CPT-Based Earthquake Liquefaction Hazard Assessment at Challenging Soil Sites." Virginia Tech, Blacksburg, VA. PhD Dissertation.

------------------------------------------------------------------------------ 

Data provided in Version 1 correspond to procedures presented in the Yost (2022) dissertation chapter. Version 2 is consistent with the procedures described in the Yost et al. (in review) manuscript and differs slightly in the following ways:

• A bug was corrected in the computation of soil unit weights, impacting normalization of all tip resistances to qc1n values.
• Different data was selected to calibrate relationships between tip resistance and relative density (for sand) and undrained shear strength (for clays) used to determine qt for the laboratory portion of the database.

-----------------------------------------------------------------------------------

REFERENCES

De Lange, D. A. (2018). “CPT in Thinly Layered Soils.” In: van Elk, J., Doornhof, D., eds., Delft, the Netherlands.

Yost, K. M., Yerro, A., Green, R. A., Martin, E., and Cooper, J. (2022). “MPM Modeling of Cone Penetrometer Testing for Multiple Thin-Layer Effects in Complex Soil Stratigraphy.” Journal of Geotechnical and Geoenvironmental Engineering, 148(2), 04021189.