59th Annual Report on Research 2014

A set of experiments was conducted on fractured Indiana limestone. The preparation and testing procedures were carefully duplicated to ensure repeatability of the tests. First, a fracture was created on a cubic specimen of Indiana limestone by diametrical compression, similar to the Brazilian test, by applying a load to two opposite sides of the sample using steel rods. Transducers, embedded into the load platens, were attached to the sides parallel to the fracture using honey as a couplant. A constant uniaxial compression stress of 4 MPa was applied during 200 minutes to stabilize the contact between the honey and the rock. After that, the stress was maintained for four hours. Once this was done, water was permeated through the fracture for five days under a constant pressure head of 100 kPa. At the end of this period, vinegar was permeated under the same constant pressure head for six days. During the entire duration of the test, transmitted and reflected compressional and shear waves were taken together with periodic measurements of flow rates and pH of the fluid exiting the fracture. This allowed us to relate changes of seismic response with flow and pH. Prior to and after each experiment, the roughness of the fracture was obtained using a laser profilometer.

A MATLAB code was written to calculate the specific stiffness of the fracture for normal incident waves. The code computes the ratio of reflection to transmission coefficients using the amplitudes within the range of the dominant frequencies obtained from a Fast Fourier Transform of the P- and S-waves.

Figure 2 shows the normal (left) and shear (right) specific stiffness of a representative fracture during the test. The colors in the figure represent the different transducers distributed along the fracture plane. As one can see in the figure, the normal and shear specific stiffness dropped as water entered the fracture. In fact, the change of peak-to-peak amplitude of the normal and cross-transmitted incident compression and shear waves was used to detect the water front arrival time. The results demonstrate (not shown in the figure) that the arrival time of the fluid front at specific locations of the fracture was inversely proportional to the specific stiffness. In other words, fluid invaded first areas with relatively low specific stiffness and then propagated to areas with higher specific stiffness. Figure 2 also shows that water flow increased the specific stiffness of the fracture over time. At about 6,000 minutes after the start of the test, an abrupt drop in specific stiffness is observed, which corresponds to the initiation of the vinegar flow. As observed with water, the specific stiffness increased with time under all transducers, but after about three days, it stabilized under the majority of the transducers. At this time, the flow rate and pH became constant. The interpretation of these data is underway and additional tests are still necessary to understand better the interplay that exists between chemical, mechanical and seismic phenomena.

Figure 1. Transducer Layout

Figure 2. Fracture Specific Stiffness with Time. P- (left) and S- (right) Wave Transducers