| As a non-invasive,direct and quantitative geophysical method for groundwater detection,surface magnetic resonance has the advantages of high work efficiency and unique interpretation.It is widely used in areas such as water finding in arid areas and regional hydrogeological surveys in China.There are also preliminary applications in early warning of tunnel mine water disaster.At present,advanced exploration of tunnel water hazards can only realize one-dimensional depth imaging,and it is difficult to accurately reflect the spatial distribution characteristics of non-layered water-containing structures such as aquifers,faults,and karst caves.It is easy to cause hidden safety problems due to the problem of positioning blind spots.Therefore,a data preprocessing method suitable for multi-peak noise removal needs to be researched to improve the signal-to-noise ratio.However,the narrow space of the tunnel limits the size and laying method of the detection coils,and the two-dimensional and three-dimensional array detection methods used on the ground are difficult to apply.At the same time,the amount of spike noise caused by the transient pulses caused by large-scale instruments commonly used in tunnels is much more than that of surface magnetic resonance detection.Traditional data processing methods cannot satisfy the quality requirements of magnetic resonance signals for advanced detection in tunnel space.Therefore,a data preprocessing method suitable for multi-peak noise removal needs to be researched to improve the signal-to-noise ratio.In addition,the magnetic anomaly brought by the metal support frame and large electrical equipment in the tunnel space causes the Larmor frequency changing with space,causing the simultaneous occurrence of magnetic resonance and off-resonance phenomena,which seriously affects the accuracy of the inversion interpretation.Existing imaging methods cannot meet the requirements of high-resolution two-dimensional imaging of water hazards in a narrow detectionspace,strong noise interference,and complex environments.Therefore,in order to realize high-precision imaging of tunnel water hazards,it is of great significance to study a high-efficiency,high-resolution and high-precision 2D magnetic resonance detection method.This article is funded by the National Major Scientific Instrument Special Project “Development of Disasters Water Source(Tunnel or Mine)Direct Detection Instruments and Equipment”.To address the above issues,this paper conducts research on the detection of hidden dangers of water disasters in tunnels.In order to meet the requirements of tunnel detection space and efficiency,A rotation detection method for tunnel water hazards was proposed for the first time.Multiple groups of magnetic resonance signals were obtained by scanning the tunneling face by changing the angle of the antenna.Based on the characteristics of the rotation detection to acquire the tunnel signal characteristics,a collaborative filtering method was proposed to suppress the spike noise interference.The calibration method ensures that the rotation detection method works under the condition that the magnetic field is stable,and finally realizes two-dimensional imaging of the water content and relaxation time of the hidden danger of tunnel water disasters.The main research contents of this article are as follows:(1)In view of the narrow tunnel space,the current research on magnetic resonance advanced detection mainly focuses on one-dimensional depth imaging in front of the tunnel face.It is difficult to perform full-space detection of the tunnel including the top and bottom walls,while the ground two-dimensional array detection can't be used.We proposed a tunnel magnetic resonance rotation detection method for the first time.By adjusting the antenna angle to measure in a fan shape,multiple sets of magnetic resonance signals were obtained.A two-dimensional forward modeling based on the rotation detection method was established,and the rotation was derived.Detect the expression of magnetic resonance response signals,calculate two-dimensional sensitivity kernel functions with different rotation angles,analyze the response characteristics of three aquifers in three typical tunnel spaces,including aqueducts,faults and karst caves,and verify the effectiveness of the magnetic resonance detection method for tunnels.(2)Aiming at the problem that the spike noise seriously affects the quality of the magnetic resonance signal obtained by the rotation detection method,a noise suppression method based on cooperative filtering is proposed.Firstly,the characteristics of the two-dimensional transform domain of tunnel magnetic resonance data are analyzed in depth.The Wiener filter coefficients are calculated by using the data without spike noise,and the two-dimensional transform domain collaborative filtering of the data with spike noise is realized.This method does not need to replace or eliminate the spike of data,so it does not lose the magnetic resonance signal while suppressing the spike noise.The effects of different peak noise amplitude,number and Gaussian noise level on the noise suppression effect are studied,and the effectiveness of the algorithm is verified.Furthermore,by comparing the peak noise suppression effect of wavelet transform,the superiority of the algorithm is verified,and the quality requirements of magnetic resonance signals obtained by rotation detection are met.(3)Aiming at the problem of composite resonance caused by metal interference,which caused rotation detection to be unable to work in non-uniform magnetic fields,a general forward model suitable for magnetic resonance,off-resonance and composite resonance was established,and the time domain expression of magnetic resonance signal under metal interference was derived.A self-correcting inversion method for frequency curve is proposed,and an objective function is established with the frequency offset as an unknown quantity.Through the iterative correction of the frequency offset,the water content,and the relaxation time,the accuracy of magnetic resonance inversion under the interference of metal is improved.The average error of the frequency model correction results of the simulation model is0.2 Hz,which meets the requirements of practical applications,and provides astrong guarantee for the quality of the water disasters hidden danger data for tunnel advanced detection.(4)Aiming at the problem that one-dimensional imaging cannot be located and quantitatively evaluated,a two-dimensional rotation detection method based on complex envelope inversion is proposed,and a two-dimensional inversion objective function for tunnel space is established.The stability and resolution of the inversion results were improved by introducing Tikhonov regularization factor and smoothing constraints.Through the simulation of the water channel model,the two-dimensional distribution images of water content and relaxation time*2T at different noise levels and rotation times are compared,and the imaging results of the minimum size,maximum distance,minimum water content,and maximum number of water bodies of the water channel are discussed.The imaging results of fault models with different slopes and different karst cave models were analyzed,and the effectiveness of the two-dimensional imaging method of rotation detection was verified.The precise positioning and high-precision imaging of hidden dangers of tunnel water disasters were realized.The innovations are as follows:(1)Aiming at the exiting measurement methods have blind space due to the narrow tunnel space,a tunnel magnetic resonance rotation detection method is proposed.The sector scan is performed by rotating the antenna,and multiple sets of magnetic resonance signals obtained from the measurement are used to invert the differences to achieve water disasters.The two-dimensional distribution imaging of hidden water content and relaxation time*2T solves the problems that conventional one-dimensional magnetic resonance detection of tunnels cannot accurately locate the hidden water disasters location and quantitatively assess the spatial distribution of them.(2)Aiming at the problem of magnetic anomalies caused by metal brackets in the tunnel space,which causes deviations in the magnetic resonance signals of therotation detection,a self-correction method of metal interference based on the frequency curve is proposed,and a general forward model suitable for magnetic resonance,off-resonance and compound resonance is established.The self-correction of the metal interference is realized by the frequency curve estimation,and the problems of the magnetic resonance and off-resonance caused by the spatial variation of the Lamer frequency,and the single emission frequency,which both caused by the metal interference are solved.(3)Aiming at the non-layered and non-uniform distribution of water hazards in tunnel space,existing imaging methods cannot reflect the spatial distribution of water hazards.A two-dimensional imaging method for tunnel water disasters hidden danger detection was proposed.By analyzing the magnetic resonance imaging resolution of tunnel rotation detection,different water channel models and complex water disasters hidden faults and karst cave models were designed.Two-dimensional high-precision imaging of hidden dangers of water disasters solves the problems of quantitative detection and early warning of hidden dangers of water disasters in tunnels.Finally,according to the above key technologies,using the JLMRS-IV instrument independently developed by Jilin University,a verification test of the rotation detection method was performed at West Lake,Jilin University Central Campus,and the advanced detection experiment was carried out at the Maijishan Tunnel on the Baolan Railway Passenger Dedicated Line.The rotation detection simulation experiment was performed on the measured data,and the water-containing structure in front of the tunnel face was obtained.The results show that the rotation detection method can obtain accurate two-dimensional imaging results of water content and relaxation time in a strong noise interference environment.It is proved that the tunnel magnetic resonance rotation detection method proposed in this paper has certain practical value.The research in this article will provide a strong guarantee for the safety of underground engineeringconstruction and further broaden the application field of magnetic resonance technology. |
PDF Full Text Request