| The construction center of tunnel engineering in China has been moving towards the western region with complex geological environments and conditions,and tunnel construction is constantly facing challenges caused by complex geological conditions.Faults,as common hazard-causing structures,often cause geological disasters such as water and mud inrush,collapse,and large deformation of the surrounding rock due to the fractured and low strength of the surrounding rock and the development of fissures in the fault influence zone.Accurate and timely identification of faults in the excavation process of deep-buried tunnels is of great significance to the safe construction of tunnels and the prevention and control of geological hazards.However,the existing geological analysis methods mainly rely on the experience of geologists to make the qualitative judgment,which is highly subjective and prone to misjudgment and omission,and it is difficult to satisfy the needs of accurate and rapid fault identification in present-day tunneling projects.This study proposes a comprehensive idea of fault identification based on multivariate spectral mineral anomaly analysis.Based on the cross-disciplinary knowledge of tunnel engineering,mineralogy,engineering geology,tectonic geology,spectroscopy,and other disciplines,the study focuses on the theoretical and technological difficulties of "quantitative identification of faults in tunnel engineering" and focuses on the following research aspects:(1)Through 220 cases of clay mineral anomalies in fault zones and analysis of typical engineering cases,the geological factors and geological processes affecting clay mineral anomalies in fault zones have been systematically revealed.Based on the mode of formation of clay minerals in the fault zones,we proposed four types and eight types of clay mineral anomaly patterns in fault zones,i.e.,the felsic anomaly category(illite type and kaolinite type),mafic anomaly category(chlorite type),ultramafic anomaly category(smectite type,serpentine+talc type,vermiculite type,and palygorskite+sepiolite type),and argillaceous anomaly category(mixed-layer type).Typical case studies of kaolinite-type and montmorillonite-type,chlorite-type,vermiculite-type,and chlorite-type mineral anomalies were carried out by comparatively analyzing the microstructural,geochemical,and mineralogical characteristics of the protoliths and fault rocks,taking the central Shanxi Yellow Diversion Project and a water diversion project in Xinjiang as examples.The results further verified the scientific validity of the proposed anomalous patterns of clay minerals in fault zones.(2)The mineral symbiotic assemblages of the anomalous patterns of minerals in fault zones were constructed.The mixing and proportioning experiments of the mineral powders were carried out,and the mixing spectra of the felsic anomaly,mafic anomaly,and ultramafic anomaly patterns were tested.The visible-near infrared(350 nm～2500 nm)spectral variability of the mineral anomaly patterns in fault zones was investigated,and the influence of the mixed spectral variability on the identification of mineral assemblages and the abundance estimation was analyzed.Visible-near infrared spectral variability characteristics of mineral anomaly patterns in fault zones can be classified into three categories:no interference,superimposed interference,and annihilation.Albite and k-feldspar with "no significant spectral absorption features" increase the error of abundance estimation,and the unmixing accuracy is also related to whether the absorption features are superimposed.The unmixing accuracy is related to the preprocessing method.However,simultaneously,it is highly controlled by the characteristics of the mineral mixing spectra.The unmixing accuracy of mixing spectra of different mineral combinations with the same preprocessing method is different.The CR-1st D preprocessing method is advantageous in linearly unmixing quaternary mineral mixtures and can better attenuate the unmixing failure problem.(3)The mid-infrared(2500 nm～5370 nm)mixed spectral variability of mineral anomaly patterns in fault zones is investigated.The superposition of reflectance peaks of similar minerals after mixing and the shift of peaks controlled by multiple minerals are the most critical variability features of the mid-infrared mixing spectra.As the visible-near infrared spectra,the accuracy of mid-infrared unmixing is related to the pre-processing method.However,simultaneously,it is highly controlled by the characteristics of the mineral mixing spectra.Since minerals such as feldspar and quartz have diagnostic features in the mid-infrared waveband,the identification of minerals such as feldspar and quartz using the mid-infrared waveband has an advantage,which makes up for the shortcomings of mineral identification in the visible-near-infrared waveband.However,the substantial interference of mid-infrared in the vibrational band of 8μm～12μm causes the variability of the spectral features of mineral mixtures,which leads to a better unmixing effect of the near-infrared band than that of the mid-infrared band in the case of the complex composition of end-member minerals.There is no linear relationship between the height or area of reflectance peaks and the mineral content of the mid-infrared spectral,and it is not easy to estimate the mineral abundance by using the relative intensity of the reflectance peaks in the mid-infrared waveband,as in the case of the near-infrared waveband.(4)Laser Raman(300cm-1～3200cm-1)mixed spectral variability of mineral anomaly patterns in fault zones was investigated.The identification of minerals by laser Raman spectra is related to the crystal structure and chemical composition:the best results are for carbonate minerals,followed by nesosilicate,chain silicate,and tectosilicate,and the worst results are for layered silicates.Unlike the visible-near infrared mixed spectral variability of mineral anomaly patterns in fault zones,in addition to superposition interference and annihilation,the laser Raman spectra have offset interference,which is mainly manifested in the fact that the presence of one mineral makes the Raman peaks of the other mineral shifted to the direction of longwave or shortwave.The pervasive noise,offset interference,and crystallographic orientation problems in laser Raman spectroscopy have led to significant challenges in mineral abundance estimation based on spectral unmixing.(5)The visible-near-infrared,mid-infrared,and laser Raman spectral criteria for identifying anomalous mineral assemblages in fault zones are summarised based on the mixed spectral variability of the mineral anomaly patterns in fault zones.A fault identification method is proposed based on multispectral mineral anomaly analysis,which has been successfully validated and applied in a water diversion tunnel in Xinjiang.The calc-alkaline and relatively Mg-rich biotite gneisses and schists were subject to chloritization under the fault stress,Fe-and Mg-rich hydrothermal fluids.The assemblage of Mg-rich clay minerals montmorillonite+saponite+talc indicates that the mineral transformation pattern in the F4 is Mg-flux transformation,and the mineral anomalous pattern is chlorite type.The quantitative difference between the anomalous enrichment of chlorite and the anomalous reduction of mica at different mileage along the tunnel identifies the internal structure of the F4. |
PDF Full Text Request