Radial Hydraulic Fracturing Experiment: 3 Cycles of Fracture Propagation, Arrest, and Closure in Molasse de Villarlod Sandstone - Sample M05

Overview

This dataset encompasses detailed measurements from a lab-scale radial hydraulic fracturing experiment conducted on a cubic sample of Molasse de Villarlod sandstone, designated as Sample M05. The sandstone, sourced from a quarry in Fribourg, Switzerland, is known for its porosity (18.1%) and permeability, making it an ideal material for studying hydraulic fracture processes. The primary focus of the experiment was to observe and analyze the propagation, arrest, and closure of hydraulic fractures under controlled triaxial stress conditions.

Experimental Setup

The experiment was conducted on a cubic sandstone sample with dimensions of 25 × 25 × 25 cm. The sample was placed in a truetriaxial frame that applied confining stresses in all three principal directions:

Vertical Confining Stress: 7 MPa

Horizontal Confining Stress: 18 MPa

A viscous glucose fluid (65 Pa.s) containing a UV additive was used as the fracturing fluid. This fluid was injected through a 1/8'' high-pressure tube cemented with epoxy into a centrally drilled hole within the sample. An axisymmetric notch was created at the injection point to facilitate fracture initiation and promote the planarity of the fracture.

The experiment was designed to simulate three cycles of fracture initiation, propagation, arrest, and closure. The closure of the fracture was occured by the leakoff of the fracturing fluid into the surrounding porous medium.

Acoustic Monitoring

To capture the dynamics of fracture propagation and closure, the experiment employed both passive and active acoustic monitoring systems:

Passive Acoustic Monitoring:

Sensors: 16 Vallen VS150-M passive piezoelectric sensors were used to capture Acoustic Emissions (AEs) within the frequency range of 50 kHz to 600 kHz.

Signal Processing: Continuous signal analysis and denoising were performed on the captured AE data. The STA/LTA algorithm was applied to the denoised signal to identify potential p-wave arrivals, providing insights into the fracture mechanics.

Active Acoustic Monitoring:

Transducers: A total of 64 piezoelectric transducers were integrated into the loading platens, with 32 acting as sources and 32 as receivers. The array included 10 shear-wave and 54 longitudinal-wave transducers.

Signal Generation and Acquisition: A Ricker excitation signal with a central frequency adjustable between 300 and 750 kHz was generated and amplified using a high-power amplifier. The signal was routed to one of the 32 source transducers via a multiplexer, and the resulting signals were recorded simultaneously by the 32 receiver transducers at a sampling frequency of 50 MHz. Each source was excited 50 times to improve the signal-to-noise ratio, with the complete acquisition sequence taking approximately 2.5 seconds.

Additional Measurements

In addition to acoustic monitoring, several other key measurements were recorded during the experiment:

Fluid Injection Parameters: The pressure and rate of fluid injection were continuously monitored.

Flat-Jack and Piston Parameters: The pressures and volumes exerted by each pair of flat-jacks were recorded at a frequency of 1 Hz.

Fracture Opening Measurement: An eddy current sensor, an electromagnetic inductive device, was placed inside the wellbore at the notch/inlet to directly measure the fracture opening.

All measurements were synchronized using a dedicated LabView application to ensure consistency across the dataset.

Conclusion

This dataset provides a comprehensive view of the hydraulic fracturing behavior of Molasse de Villarlod sandstone under controlled laboratory conditions, with a focus on the processes of fracture propagation, arrest, and closure. The dataset includes raw and processed acoustic data, fluid injection metrics, and direct observations of fracture opening. It is an invaluable resource for researchers and engineers studying hydraulic fracturing, rock mechanics, and related fields. The data is suitable for detailed analysis and modeling of fracture mechanics in porous, permeable sandstones.

Processing code

Follow the URL repositories below to access to the codes for processing these dataset.

https://github.com/GeoEnergyLab-EPFL/ActiveAcoustiX

https://github.com/GeoEnergyLab-EPFL/FracLowRate

Contact and Support

Email:

Brice Lecampion: [e-mail]

Mohsen Talebkeikhah: [e-mail]

Technical info

Dataset Structure

The dataset is organized into several files, each containing specific types of data recorded during the hydraulic fracturing experiment on the Molasse de Villarlod sandstone sample (Sample M05). The files are named based on the date of the experiment and the exact time when the data was recorded. Below is a detailed description of each file:

Folder: 23-12-06

This folder is named after the date of the experiment (December 6, 2023) and contains all the relevant data files from that day.

Files within the 23-12-06 Folder:

183454.bin

Description: This binary file contains the active acoustic data recorded during the experiment. The data captures the signals from the 64 piezoelectric transducers used in the active acoustic monitoring setup.

183454.ini

Description: This file contains low-rate data recorded during the experiment. It serves as the raw input that needs to be converted into a readable format for analysis.

Note: This file is converted to 183454.xlsx for easier data access and analysis.

183454.xlsx

Description: This Excel file is the converted and readable version of the low-rate data initially recorded in 183454.ini. It contains processed data from the experiment, including various measurements and observations.

183454.json

Description: This JSON file holds comprehensive metadata about the experiment. It includes information such as the type and dimensions of the rock sample, details about the fracturing fluid, the applied stresses, and the locations and properties of both active and passive sensors.

183454.txt

Description: This text file contains the timing information for the active shots and the passive acoustic emission (AE) events. It is crucial for correlating the acoustic data with the exact moments of fracturing and AE occurrences.

Opening_Measurement_Output.xlsx

Description: This Excel file provides data on the fracture opening measurements over time. It includes the exact times and corresponding fracture opening values, offering insight into the dynamics of fracture closure and propagation.

Passive_catalog-AE-count-M05.csv

Description: This CSV file contains a catalog of the number of Acoustic Emissions (AEs) detected per real-time during the experiment. It represents the results of the passive acoustic monitoring data.

Note: Due to the large size of the passive data (several terabytes), the full dataset is not included in this upload. Interested parties can contact the research team for access to the complete passive data.

Access to Full Data

For full access to the passive acoustic data, which is too large to upload (several terabytes), please contact the research team directly. The catalog provided (Passive_catalog-AE-count-M05.csv) gives a summary of the passive AE events, but the raw data requires special handling and storage due to its size.