查看更多>>摘要：The application of effective exploration techniques in mechanized tunneling is crucial in order to obtain a knowledge of the subsoil prior to drilling. We present a three-staged method, which seeks to image the excavation environment in detail, also reducing the computational demand of common exploration techniques. The algorithm combines two approaches: supervised machine learning and full waveform inversion. Firstly, the machine learning algorithm is applied on data sets of measured pore water pressures and ground settlements during tunnel propagation, making a primary prediction of geological changes ahead of the boring machine. Secondly, seismic measurements are acquired for full waveform inversion based on parameter identification. This method incorporates the primary predictions from the supervised machine learning in the form of a parametrization of the position, shape and material properties of the disturbance. Thirdly, the subsurface model gained out of the second stage is utilized as a starting model for a second full waveform inversion using the adjoint method, providing an even more detailed image of the subsurface. The exploration algorithm is tested on a synthetic shallow tunnel environment with an unknown obstacle ahead of the tunnel boring machine. It is shown that the algorithm finds the unknown obstacle with improved accuracy.
查看更多>>摘要：Seismic imaging using ambient noise has been widely used to image subsurface structures. The spatial auto correlation (SPAC) method using ambient noise recorded by seismic arrays can image subsurface velocity structures at various scales. In addition, an approximation of zero-offset reflection responses of the structure can be achieved by station autocorrelations of the ambient noise. In this study, to detect an abandoned mine tunnel that is expected to have low velocity anomalies, we jointly apply the SPAC method and station autocorrelation method along linear arrays on the surface above the tunnel. One hundred and thirty-six nodal seismometers were deployed linearly along three lines with station intervals of 1.5 and 5 m, respectively. The ambient seismic noise was continuously recorded for two hours by these stations. Through the SPAC analysis, the Rayleigh wave dispersion curves for the frequency band of 3 Hz to 15 Hz were extracted from the ambient noise, from which the 2D near-surface shear-wave velocity models were estimated. A low velocity zone is obvious at a depth of about 60-100 m in the velocity profile along line 2. In addition, clear diffraction signals can be seen at similar positions on station autocorrelation profile along line 3, which is very close to line 2. This anomalous area is interpreted to be the abandoned tunnel, which is further verified by drilling. Our study shows that the combination of SPAC and station autocorrelation methods can be effective in near surface exploration.
查看更多>>摘要：At present, there are many deblending algorithms to deal with the separation of the blended data. However, the convergence speed for most algorithms is relatively slow, which is difficult to implement in industrial production. Thus, we propose an accelerated joint iterative method (AJIM) to separate the blended data based on the iterative shrinkage thresholding algorithm (ISTA) and the acceleration linearized Bregman method (ALBM) in the curvelet transform domain. In conventional ISTA, the updated deblended result of each iteration is obtained by thresholding curvelet coefficients and the gradient term. It is stable but always needs a lot of iterations to obtain a good performance. ALBM is proposed based on LBM via an acceleration factor at each iteration, and the acceleration factor can increase the proportion of unthresholded curvelet coefficient, which further accelerates the convergence. This method converges more faster than LBM and ISTA at the beginning, but it can cause artifacts at the late iterations because some unthresholded coefficients contain some wrong information. Based on advantages of ISTA and ALBM, the weight parameters with linear and exponential functions are proposed to control the contribution of ISTA and ALBM in the proposed AJIM. Therefore, AJIM can greatly speed up the deblending efficiency at the beginning and improve the deblending accuracy in the late stage. At last, AJIM can obtain the same deblending effect as ISTA, but only a few iterations are used. Synthetic and field data tests are used to demonstrate that AJIM-based deblending can effectively and quickly separate the blended data.
查看更多>>摘要：To overcome the limitations including the topography in survey area and the complex geometry of the target on conventional transient electromagnetic method (TEM) interpretation, we extend in this paper a workflow for processing semi-airborne TEM data. Based on the digital topography data, geologic sections and well logs, a 3D geologic model was built for numerical simulation. After mapping the geologic model to resistivity mesh, a mimetic finite volume with Krylov subspace method was used for calculating the influence coefficient of topography. This processing workflow was applied on a semi-airborne TEM survey for determining the thickness of cover layer on a hillside in Yunnan, China. The thickness of the cover layer inferred from the processed TEM data is consistent with the local geological sections. This workflow can not only process measured data, but also perform numerical simulation on the target area with complex topography and stratum structure.
查看更多>>摘要：Historic landfills that were constructed to standards that would not meet current regulations represent environmental and human-health risks. The adequate characterization of legacy landfills is paramount to manage the risks they pose effectively. In this study, the non-invasive geophysical technique of electrical resistivity tomography has been used to characterize two sites that have been identified as posing risks to the local environment. The sites are significantly different in the type of waste present, moisture content, type of bedrock and shape and distribution of the waste; although they have in common that neither site is lined.The results of the resistivity imaging have allowed defining the boundaries of the landfill and the depth of the waste. Different types of waste and their distribution have been identified. Additionally, a relationship between the resistivity distribution and the release of residual gas from the landfill has been described; where the presence of saturated clay in the waste prevents the release of said gasses. This is be associated with locations with higher resistivity and lower clay content. That relationship is not observed when the clay is dry. When interpreting the results, the heterogeneous nature of the waste can lead to misinterpretations due to resistivity overlap between the bedrock and the waste materials. Therefore, comparison of the resistivity models with direct information like borehole logs can significantly improve the reliability of the interpretations. However, the resistivity survey should predate the installation of bores to identify the most suitable locations for them.
查看更多>>摘要：The spectral induced polarization (IP) technique is widely used in prospecting and applied geophysics and allows remote assessment of porosity, mineralogical composition, textural and structural features of rocks, and chemical composition of the saturating fluid. The nature of the IP phenomenon in rocks is still not fully understood; therefore, much attention is paid to laboratory IP measurements on rock/ore samples carried out in the time (TDIP) and frequency (FDIP) domains. The intensity of the polarization effect can be low in rocks. Reliable IP measurements on rock samples in a broad time or frequency domain require tailored high-precision equipment. This paper presents the first experience of using the multifunctional electrochemical P-40 x potentiostat-galvanostat instrument for high-precision laboratory measurement of spectral induced polarization (SIP) on rock samples. The advanced capabilities of the P-40 x allowed us to implement a fully automated TDIP and FDIP measurement technology. The difference between the P-40x and specialized equipment for IP measurements in the time domain lies in recording of TDIP signals (full wave) of current and voltage without processing. A code has been developed for processing the observed TDIP signals. It suppresses harmonic and non-harmonic interference signals and the drift of potential electrodes and calculates IP parameters. The code and proposed algorithm for processing TDIP signals were successfully tested on synthetic signals contaminated by noise components of various origins and data from laboratory IP measurements on rock samples. We also compared the TDIP and FDIP measurements obtained with the P-40 x and specialized equipment AIE-2 (TDIP) and SIP Fuchs III (FDIP), which have been used for high-precision laboratory measurements of IP for many years. We run our experiments with such test materials as RC circuit, ore carbonatite samples, and sand. Here we demonstrate that the IP data obtained with the P-40 x and specialized equipment are identical.
查看更多>>摘要：The limited bandwidth of a single source normally restricts the imaging resolution for the subsurface target. Multi-frequency acquisition is a promising way to enhance the resolution with reasonable data fusion. An Ormsby wavelet based theoretical study confirms the significance of both the low and high frequencies for improving the resolution by means of peak duration and side-lobe oscillation level. To avoid over-boosting the noise during the data fusion, we propose a joint deconvolution approach to associate the single-frequency data with the target function directly. A weighted matched filter is accordingly derived based on the signal-to-noise ratio of the multi-frequency data. Error analysis shows that the joint weighted deconvolution based data fusion achieves better noise control compared to the direct-summation based deconvolution. Both wedge model based synthetic result and ground penetrating radar acquired field data example confirm the effectiveness of the proposed data fusion approach on resolution enhancement.
查看更多>>摘要：The anelasticity and anisotropy widely exist in real subsurface media. Strong anelasticity will lead to phase dispersion and amplitude attenuation during seismic wave propagation. Anisotropy cause seismic waves to have obviously different kinematic characteristics from that of isotropy. For seismic migration, ignoring the anelasticity and anisotropy of subsurface media will significantly reduce the quality of migration images, even cause imaging failure. We propose a Q-compensated least-squares reverse time migration (Q-LSRTM) in tilted transversely isotropic (TTI) media to correct these effects. According to the Born approximation of seismic wave equation, a linearized visco-acoustic TTI pure qP-wave modeling operator is derived using a new visco-acoustic TTI wave equation for one standard linear solid (SLS) model, which can deal with the anelasticity and can simulate pure qP-wave steadily in attenuating anisotropic media without qSV-wave artifacts. Then, the corresponding adjoint equation is formulated using the adjoint-state method to calculate the gradient sensitivity kernel for the visco-acoustic TTI media. Because of the least-squares inversion, the Q compensation can be achieved during the iterations, so that the over-amplification of noises can be avoided naturally. In addition, compared with conventional LSRTM, the proposed method compensates for the anelasticity and corrects the anisotropy, so as to produce images with better spatial resolution and amplitude fidelity. Numerical examples demonstrate the feasibility and advantages of the proposed method for the data including strong attenuation effects over conventional LSRTM.
查看更多>>摘要：Electrical resistivity tomography (ERT) is rightly considered an effective tool for studying subsurface, but some particular geological conditions can lead to results that are difficult to explain. Carbon-rich rocks undoubtedly belong to such cases because of their very high resistivity contrast compared to the host rock. Shungite-bearing rocks have total organic carbon (TOC) in the range of 5 to 95 weight percentage (wt%), and such a range ac-counts for a wide variety of electrical rock properties. We used ERT at three shungite deposits to study their internal structure and to evaluate the reliability of the resulting geoelectrical models. The study of carbon-rich rocks (shungites) of the Zaonega Formation in the Onega basin is interesting in the Precambrian carbon accu-mulation and the exploration of shungite deposits. As a result, we clarified the main features of geological sections, such as topsoil base, primary shungite-rich successions, individual zones of metasomatic rocks, and gabbro intrusions. Drilling data and geological reference sections show the reliability of the ERT cross-sections and accuracy of main geological boundaries with an average discrepancy of 1-2 m. Laboratory studies of drill cores samples revealed the nonlinear relationship between resistivity and TOC, which, in the future, may facilitate identifying rocks with relatively low carbon. However, it was impossible to quantitatively assess TOC by ERT models due to the resolution-dependent correlation loss. Comparison of synthetic model with natural field one shows that the significant considerable contrast of resistivity for geological layers about 1:1000 ratio and more produces lower sensitivity and higher uncertainty parts of the inverse models. At the same time, despite high RMS errors, all obtained models retain their geological validity. Our research results demonstrate the ef-ficiency of ERT for mineral exploration of shungites and reveal its internal limitations when the survey these specific highly conductive carbon-rich rocks.
查看更多>>摘要：This work aims at understanding the potential of ambient vibration monitoring in deriving information about material properties change related to landslide seasonal behaviour and resonance directivity in landslides. The case study is the Pietrafitta landslide in southern Italy, known for its effects on the SS87 National Road and for being influenced by artificial vibrations. The field experiment consisted in four two-day-long measurement cycles, completed between April and November 2016, by using two broad-band three components velocimetric stations: one station at the landslide's toe and a reference station outside the landslide. Obtained data were analysed in terms of root mean square, spectral amplitude, horizontal-to-vertical spectral ratio and directionality effects. Results indicated the existence of amplitude daily variations due to vehicular traffic for both the stations, and long-time variations in the amplitudes and frequency content at the landslide site. Polarization and directional resonance showed a daily variation, with the most coherent results in the range of 20 degrees - 60 degrees at the night hours, in the absence of traffic. This azimuthal trend was transverse to the sliding direction. The most marked resonance effect was found in the April recordings. The observed horizontal-to-vertical spectral ratio curves changed over time, due to variation in shear stiffness related to seasonal consolidation. Results from the horizontal-to-vertical spectral ratio inversion suggested a decrease from April to October 2016 in the Vp / Vs ratio, and in the velocity contrast at the base of the sediments consistent with common decrease in landslide activity due to summer drought.
NSTL
Elsevier