# Comprehensive Database of Correlated Wind Velocities for Structural Analysis and Risk Assessment Mohanad Khazaali¹, Vasileios Christou², Paolo Bocchini³ ¹ Research Assistant, Department of Civil and Environmental Engineering, ATLSS Engineering Research Center, Lehigh University. Email: maa314@lehigh.edu ² Formerly Postdoctoral Research Associate, Department of Civil and Environmental Engineering, ATLSS Engineering Research Center, Lehigh University. Email: christouv@lehigh.edu ³ Associate Professor and Director of Graduate Programs, Department of Civil and Environmental Engineering, ATLSS Engineering Research Center, Lehigh University. Email: paolo.bocchini@lehigh.edu Corresponding Author --- ## Introduction and goals This is a comprehensive repository of spatially and temporally correlated wind velocity samples. The samples are classified based on the maximum sustained wind speed (maximum 1-min average) and the 3-sec gust wind speed (maximum 3-sec average), which are the intensity measures commonly used for structural analysis. The repository consists of 58,000 wind samples with 1-hour duration, as well as 110,847 samples with 2-min and 87,257 wind samples with 10-min durations. This database is meant to facilitate researchers and engineers in readily selecting and using wind velocity time histories for many applications, such as the analysis of tall structures located in open terrain areas. ## How to Cite **Dataset:** Khazaali, M., Christou, V., Bocchini, P. (2020). Comprehensive Database of Correlated Wind Velocities for Structural Analysis and Risk Assessment. Available at https://doi.org/10.6084/m9.figshare.12965270 **Companion paper:** Khazaali, M., Christou, V., Bocchini, P. (2020). Classification Protocol and Comprehensive Database of Correlated Wind Velocities for Structural Analysis and Risk Assessment. Under review. ## Description of the Repository All files in the repository are provided in comma separate value (CSV) format. The files can be opened with any text editor (such as Notepad) and any application for spreadsheets (Such as MS Excel, Apple Numbers, or Google Sheets). Each CSV file contains one sample. The data in each file is organized in 13 columns and a variable number of rows. The first row of each file contains the column labels, indicating the variable contained in that column and its units. Each other row is associated with a specific instant in time. The first column includes the time stamps; the other 12 columns correspond to the 12 components of the wind velocity vector field, named V1 to V12. The 12 components are associated with 12 heights, ranginging from 5 meters to 60 meters over the ground, with a resolution of 5 meters. The velocity at intermediate heights can be obtained by interpolation. Cell (r+1,c+1) contains the wind velocity of component c at the time indicated in cell (r+1,1), for a specific sample. All data in the repository follows the Interational System of units. Hence, the unit of time is seconds (s) and the unit of velocity is meters per second (m/s). The wind repository comprises five main folders, described in the following subsections. ### 1-hour long samples, organized based on mean wind speed The folder called “1-hour samples organized by mean wind speed” comprises 58 sub-folders associated with mean wind speeds ranging between 18 m/s and 75 m/s with increments of 1 m/s. Each sub-folder contains 1,000 files in CSV format, each with 13 columns (time instants plus 12 components), as explained formerly. The sub-folder name indicates the mean wind speed of the samples that it contains. For example, the sub-folder named “Mean18_1hr” contains 1,000 samples generated with mean wind speed of 18 m/s at 10 m over the ground. Each sample within the sub-folder is named based on its mean wind speed and the associated wind intensities. For instance, “Sample00001_1hr_mean18_MSWS20_GWS26.csv” indicates that this file contains the first sample, which has 1-hour duration, mean equal to 18 m/s, and it happens to have a maximum sustained wind speed in the range 20-21 m/s and 3-sec gust wind speed in the range 26-27 m/s. In total, 58,000 wind samples are stored in this folder and occupy 332 GB in size. ### 2-minute long samples, organized by maximum sustained wind speed The folder named “2-min samples organized by MSWS” contains wind samples with 2-min duration, where the 1-min window including the maximum sustained wind is placed at the center of the 2-min. In total, 66 sub-folders with maximum sustained wind speed ranging from 19-20 m/s to 84-85 m/s are listed. For example, the sub-folder named “MSWS19-20_2min” includes samples with 2-min duration and maximum sustained wind speed that is greater than 19 m/s and smaller than or equal to 20 m/s. In general, the samples in each sub-folder have different mean wind speeds and the file names indicates this. The number of samples in each sub-folder varies, and each wind sample file is named with the same convention explained previously. For example, the file named “Sample00046_2min_aroundMSWS_mean18_MSWS20_GWS26.csv” contains sample generated with mean wind speed equal to 18 m/s, with a maximum sustained wind speed in the interval 20-21 m/s and gust wind speed in the interval 26-27 m/s. In total, 55,816 wind samples are stored in this folder (the remaining 2,184 wind samples are not included because their maximum 1-min averages fall in the first or last 30 s, as explained in the companion paper) and occupies 11 GB in size. ### 10-minute long samples, organized by maximum sustained wind speed The wind samples with 10-min duration and classified based on their maximum sustained wind speeds are organized in a similar way. The corresponding folder is named “10-min samples organized by MSWS” and it contains 40,838 samples organized in 66 sub-folders, which occupy 40 GB. ### 2-minute long samples, organized by 3-sec gust wind speed The folder named “2-min samples organized by GWS” contains 2-min long samples, where the 3-sec wind time history including the 3-sec gust wind speed is centered in the 2-min. In total, 90 sub-folders with 3-sec gust wind speed ranging from 24-25 m/s to 113-114 m/s are included. For example, the sub-folder named “GWS24_25_2min” includes samples with 2-min duration and with 3-sec gust wind speed that is greater than 24 m/s and smaller than or equal to 25 m/s. The naming convention follows the general scheme. For instance, the file named “Sample00046_2min_aroundGWS_mean18_MSWS20_GWS26.csv” has original mean equal to 18 m/s, with 1-min maximum sustained wind speed in the interval 20-21 m/s and 3-sec gust wind in the interval 26-27 m/s. In total, 55,031 wind samples are stored in this folder (the remaining 2,968 wind samples are discarded for the reason explained in the companion paper) and occupy 11 GB in size. ### 10-minute long samples, organized by 3-sec gust wind speed The same organization protocol is followed for the wind samples with 10-min duration and classified based on their 3-sec gust wind speeds. In this case, the folder is named “10-min samples organized by GWS” and it contains 43,419 samples organized in 90 sub-folders, which occupy 42 GB. ## Acknowlegments This repository is one of the outcomes of the “Probabilistic Resilience Assessment of Interdependent Systems (PRAISys)” project (www.praisys.org). The support from the National Science Foundation through grant CMMI-1541177 and from the Pennsylvania Department of Community & Economic Development (DCED) through grant PIT-19-02 is gratefully acknowledged. The support of the Department of Civil and Environmental Engineering and the ATLSS Engineering Research Center at Lehigh University is gratefully acknowledged too. The authors would also like to thank Mr. Liyang Ma for providing thoughtful feedback and assistance throughout this work. The opinions and conclusions are based on the authors perspectives and do not necessarily reflect the views of the sponsoring organizations.