Experimental Evaluation of a Distributed Fiber Optic Sensor for Mining Application
Mollahasani Madjdabadi, Behrad
MetadataShow full item record
Triggered remote seismic events have been widely studied in the earthquake engineering context where various possible explanations have been provided, including directivity of dynamic stresses, a critically stressed environment, the presence of hydrothermal geological environments at remote distances, and so on. Similar events have been observed in underground mining regions; however, they have rarely been studied in terms of the underlying mechanisms such as the presence of faults of marginal stability, increases in the stress gradient between mined out regions as a result of connective fractures, unclamping effect on geological features such as dikes or joint swarms, and so on. This research was triggered in part by the hypothesis that remote seismic events in mines could be triggered when gravity-driven displacements are transferred to distances far from active mining (10’s to 100’s of metres). Accordingly, the thesis focuses on experimental research on a novel deformation sensing sensor for future verification of this assumption. A secondary focus is mathematical modeling to help understand the deformation mechanisms and magnitudes that may take place in a jointed rock mass. Distributed Brillouin sensing systems (DBSs) have found growing applications in engineering and are attracting attention in the field of underground structures including mining. The capability for continuous measurements of strain over large distances makes DBSs a promising monitoring approach for understanding deformation field evolution within a rock mass, particularly when the sensor is installed away from “excavation damaged zones” (EDZ). A purpose-built fiber optic sensing cable, a vital component of DBSs, was assessed in laboratory conditions to establish the capability and limitations of this technology to monitor deformation fields over large distances. A test program was performed to observe DBSs response to various perturbations including axial and shear strain resulting from joint movements. These tests included assessments of the strain-free cable response and the application of extensional and lateral displacement to various sensing cable lengths (strained lengths from 1 m down to 1 cm). Furthermore, tests were done to evaluate the time-dependent behavior of the cable and to observe the effect of strain transfer using a soft host material (i.e., a soft grout) under lateral displacement. The noise level of the DBSs range was ±77 µε, determined through repeated measurements on an unstrained cable. Stretching test results showed a linear correlation between the applied strain and the Brillouin frequency shift change for all strained lengths above half the spatial resolution of the DBSs. However, for strained lengths shorter than half the spatial resolution, no strain response was measurable and this is due to the applied internal signal processing of the DBSs to detect peak Brillouin gain spectrum and noise level. The stability with time of the measurements was excellent for test periods up to 15 hours. Lateral displacement test results showed a less consistent response compared to extension tests for a given applied displacement. The Brillouin frequency shift change is linearly correlated with the applied displacement in tension but it shows a parabolic variation with lateral displacement. Moreover, the registered frequency response (correlated with strain) of the system decreased significantly when the sensing cable was embedded in a sand-filled tube compared with direct cable displacement. A comprehensive laboratory scale testing program was undertaken to study the response of the system to different loading paths in time and space domains. Purely extensional displacement fields were applied to demonstrate that the system could produce repeatable displacement responses for three different configurations of distributed strained patterns. A borehole installation method was developed by testing the sensing cable’s response while embedded in mortar beams. When the cable is directly embedded in the mortar, uncontrolled self-debonding happens that introduces uncertainties in the measurements. This limitation was overcome by anchoring debonded sections of the sensing cable at regular spacing. This arrangement produced consistent strain patterns for each strained interval. It was shown that the performance of the debonded sections changes for longer anchor spacing and for closely spaced joints where more than one joint crosses the debonded interval. The influence of borehole diameter and strength of the filling material were evaluated for their possible effects on the strain transfer process to the sensing cable. With the anchored arrangement of debonded cables, these properties of the grout did not have a measurable effect on the results, and as long as the tensile strength of the grout is low enough to break at the joint locations, the strain transfer performance from the rockmass to the sensing cable was excellent. A study was devoted to understanding such a deformation monitoring system under various shear displacement conditions. These included the difference in response of the system in direct shear compared to tests performed in direct tension. The system response was evaluated for various strained lengths as well as distances over which the bending strains are acting (kink lengths). The latter was found to be an important factor influencing monitoring results. In addition, the system behavior under shear displacement where the sensor is inclined with respect to the joint strike was evaluated to understand the effect of a combined extension and shear displacement. The effect of the existence of two joints over the strained lengths was assessed in both direct and inclined shear. A new relation was established between the registered Brillouin frequency shift change and all contributing components of deformation when the sensor is elongated while under shear displacement. The testing program shows that Distributed Brillouin Sensing (DBS) technology has promise for detecting deformations over long distances. Not only strain localization occurring at pre-existing discontinuities or at developing cracks can be detected by this sensor, but also strain levels well below the typical damage initiation threshold (~0.1%) for hard brittle rocks are above the basic noise level of the system. However, the sensing element is quite fragile when under shear displacement and can easily break at small shear displacements. Therefore, it is better to have an idea of the dominant deformation mechanism in the rock mass before the installation of the sensor. The sensor would be much more durable where the rock mass experiences less shearing. Mathematical simulations of a 2D rock mass were carried out using the distinct element method. Two major parameters including interlocking degree and pre-existing conditions such as mined-out zones at higher levels were studied. Different rock mass models with varying block sizes, joint set orientations, and joint persistency were built to study the effect of interlock on the displacement pattern away from mining. In general, displacement as large as 5 cm could travel distances as far as 500 m away from the active mining zone. The exact displacement pattern is largely controlled by the characteristics of the joints sets. However, the transfer of large displacements was limited to distances of the size of the mining boundary, where rock mass interlocking promote arching. Furthermore, with non-persistent joint sets, a few shear slip events were noticed at higher levels whereas more remote joints did not show slip. All slip events were close to the mining boundary. Remote shear slip events, could not be generated by changing parameters representing the degree of interlock in the rock mass. When a backfilled old mine was added to the middle height of the model, some 500 m away from active mining, results showed that a large number of joints around the old mining zone slipped due to displacements induced by the distant deeper mining. It was found that the pre-existing excavation and the mine extraction strategy is a critical factor for providing conditions under which such slip events at remote distances occur from active mining.