16:15 〜 16:30
[SSS10-26] Validating strain measurement using Fiber Bragg Grating sensors embedded in a shallow thin slit
キーワード：FBG sensor、Strain measurement
The spatially dense measurement of the strain on the two-dimensional fault surface is desired to clarify the localized aseismic slip and the earthquake nucleation process, which has been missed in laboratory experiments with the traditional strain measurement installed only on the side surfaces of the rock specimen. The optical fiber containing the Fiber Bragg Grating (FBG) sensors is one of the best solutions for this, as the slits used to embed the optical fibers into the fault plane are extremely thin and will have minimal influence on the frictional slip of the fault. Additionally, the spacing of the imprinted FBG sensors on a single fiber can be as small as 10 mm, enabling high-resolution strain measurements. In this study, we conducted a preparatory test using a rock specimen with the slits mimicking the simulated 6-meter-long fault with the embedded FBG array used for planned large-scale rock friction experiments. The test aimed to evaluate the performance of the embedded FBGs by comparing them with the conventional strain gauges installed on the specimen surfaces.
We used a 55 x 37 x 600 mm metagabbro specimen, the same type of rock as the planned large-scale experiments with Young's modulus of 103 GPa. We made two parallel slits in the longitudinal (z) direction with widths and depths of 1 mm and 1.5 mm, respectively, which were also designed same as the simulated fault. We embedded the optical fibers containing the FBG sensor (Technica Optical Components, LLC.) with a spacing of 200 mm in the z-direction by filling the adhesive (EP-340, Kyowa Electronic Instruments Co., Ltd.) into the slits. Three optical fibers were embedded into the same slit such that the FBGs were placed at the identical measurement location to evaluate the coupling effect. We also installed two FBGs on the opposite surface of the specimen and the four semiconductor strain gauges (KSP-2-120-E4, Kyowa Electronic Instruments Co., Ltd.) on both sides of the surface near the FBGs to use as a reference to evaluate the FBGs. We used the HYPERION si255 interrogator (Luna Innovations Inc.) to measure the strain with the FBG sensors at 5KHz sampling.
We placed the rock specimen vertically in the uniaxial loading apparatus and applied a force of 0.4, 1, 2, 3, and 4 kN for 10 seconds to induce the axial deformation of the rock specimen. The loading cycle was repeated three times. We could measure the stain with the FBG in the range between 2-20 με, consistent with the strain expected from the applied force and Young's modulus of the rock specimen. The three collocated FBGs showed equivalent values, indicating the coupling effect did not affect the strain measurements. However, the systematic discrepancy between the FBGs and the strain gauges remained, likely caused by the bending of the rock specimen caused by the spherical platen inserted between the specimen and the load cell. Therefore, we corrected the FBG measurements by considering the bending moment in the two transverse directions. The correction reduced the discrepancy between the FBGs and the references, although the magnitude of strain with the FBGs was still consistently overestimated by 10-20 % from the reference. We suggest that the discrepancy may be caused by the effect of slit deformation on the strain measurement, which was ignored in the present study for the sake of simplicity, and the variation in the strain sensitivity of the FBGs.
Overall, the embedded FBGs are sensitive enough to measure the strain change of 2-20 με, which can monitor the migration of the local slip front on the simulated fault.
We used a 55 x 37 x 600 mm metagabbro specimen, the same type of rock as the planned large-scale experiments with Young's modulus of 103 GPa. We made two parallel slits in the longitudinal (z) direction with widths and depths of 1 mm and 1.5 mm, respectively, which were also designed same as the simulated fault. We embedded the optical fibers containing the FBG sensor (Technica Optical Components, LLC.) with a spacing of 200 mm in the z-direction by filling the adhesive (EP-340, Kyowa Electronic Instruments Co., Ltd.) into the slits. Three optical fibers were embedded into the same slit such that the FBGs were placed at the identical measurement location to evaluate the coupling effect. We also installed two FBGs on the opposite surface of the specimen and the four semiconductor strain gauges (KSP-2-120-E4, Kyowa Electronic Instruments Co., Ltd.) on both sides of the surface near the FBGs to use as a reference to evaluate the FBGs. We used the HYPERION si255 interrogator (Luna Innovations Inc.) to measure the strain with the FBG sensors at 5KHz sampling.
We placed the rock specimen vertically in the uniaxial loading apparatus and applied a force of 0.4, 1, 2, 3, and 4 kN for 10 seconds to induce the axial deformation of the rock specimen. The loading cycle was repeated three times. We could measure the stain with the FBG in the range between 2-20 με, consistent with the strain expected from the applied force and Young's modulus of the rock specimen. The three collocated FBGs showed equivalent values, indicating the coupling effect did not affect the strain measurements. However, the systematic discrepancy between the FBGs and the strain gauges remained, likely caused by the bending of the rock specimen caused by the spherical platen inserted between the specimen and the load cell. Therefore, we corrected the FBG measurements by considering the bending moment in the two transverse directions. The correction reduced the discrepancy between the FBGs and the references, although the magnitude of strain with the FBGs was still consistently overestimated by 10-20 % from the reference. We suggest that the discrepancy may be caused by the effect of slit deformation on the strain measurement, which was ignored in the present study for the sake of simplicity, and the variation in the strain sensitivity of the FBGs.
Overall, the embedded FBGs are sensitive enough to measure the strain change of 2-20 με, which can monitor the migration of the local slip front on the simulated fault.