NEES Project Warehouse
The experiment contains the continuously recorded unprocessed data that were later sampled and used for analysis in experiments 1 through 6. The experiment set up report is available on the on the project level.
Complete - 01/09/2012
The curation dashboard shows your progress as you add required elements to your project or experiment.
|Description:||The files stored in the Unprocessed Data folder of Experiment-7, Trial-1, Rep-1 are not from a single Experiment, but consist of data recorded continuously through the duration of the testing period. Each file contains 12 channels of raw data captured by one of two nees@UCLA Kinemetrics Granite recorders using a sample rate of 1 kHz, with separate data files saved at the end of each hour of testing. Anti-aliasing filters were used, limiting the maximum usable frequency to 400 Hz. Twenty channels of data were also split and simultaneously recorded on the National Instruments (NI) dynamic signal analyzer for real-time processing as Frequency Response (Transfer) Functions, Coherence, and Spectral Densities during testing. However, such real-time processing required the specification of a reference (stimulus) channel for the transfer and coherence functions, and a sampling rate which also determined the frequency resolution. The spectral measurements along with the final of 15 time windows used in ensemble averaging were saved by the National Instruments signal analyzer, and can be found in the Corrected Data folders of Experiment-1 through Experiment-6 in both Matlab binary and CSV formats. To preserve the full time-histories without being subject to the choice of sampling parameters selected for the NI signal analyzer, the Kinemetrics Granite systems were used to continuously record time-histories at the maximum possible sample rate of 1 kHz. The resulting data in the *.OK01.txt and *.OK02.txt files can therefore be re-sampled to investigate a number of different transfer functions (e.g. pile-cap to soil, soil to soil) or time-domain analyses, giving the user the freedom to choose their own sampling parameters and techniques.
Each file from the Granite systems contains 12 channels of data in tab-delimited ASCII text format. The first column contains the sample times in Unix time format, which is the number of seconds elapsed since January 1st, 1970. The sample times can be converted to Central Daylight Time (CDT) or Greenwich Mean Time (GMT) for comparison with the test log, or zeroed to obtain elapsed seconds from the start of the data file. Columns 2-13 contain the measured data in units of bits (or "counts"), which can first be converted to Volts using the conversion factor 2-to-the-power-24 bits=40 Volts, then converted to acceleration using the conversion factors in the sensor tables located in the Experiment level documentation. The raw data will also require corrections for zero offset, and a few channels have significant 60 Hz noise which will require filtering.
The file names correspond to the time at which the data files were first created, usually at the top of the hour, with the naming conventions YYYYMMDDHHMMSS.OK01.txt and YYYYMMDDHHMMSS.OK02.txt in Greenwich Mean Time. Because individual trials lasted only a few minutes while the raw Granite data files span approximately one hour, the data from a single trial must therefore be extracted from the appropriate pair of Granite data files. For this purpose, the approximate start and stop times for each trial are listed on the test log in local Central Daylight Time (CDT=GMT-5) as well as GMT.
For ease of analysis and smaller file sizes, the Granite ASCII text data files were also converted to Matlab binary format. The ASCII and Matlab files are stored in Experiment-7/Trial-1/Rep-1/Unprocessed_Data.
|Dates:||September 28, 2010 - September 29, 2010|
|Facility:||University of California, Los Angeles, CA, United States|
No files available