Mine excavation seismic signal imaging method
By performing optimization, segmentation, exploration source conversion processing and full waveform inversion imaging of the mine seismic signal, the problems of inaccurate signal processing and low imaging accuracy in the prior art are solved, and high resolution and high accuracy are achieved The imaging effect meets the geological guarantee needs of intelligent tunnel excavation.
Patent Information
- Application Number
- CN202211255751.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2022-10-13
- Publication Date
- 2024-11-29
- Estimated Expiration
- 2042-10-13
AI Technical Summary
The existing technology is difficult to achieve high-resolution processing and high-precision imaging of mine seismic signals, resulting in the failure to meet the geological guarantee requirements for intelligent tunnel boring.
By obtaining the continuous seismic signal, performing preferential and segmented, performing exploration source conversion processing and processing results in segments, performing pre-processing, and using a full waveform inversion method for roadway advance detection imaging.
It realizes high-resolution conversion and utilization of seismic signals along the excavation, improves the high-precision imaging capability of abnormal geological structures ahead of the excavation, and meets the requirements of intelligent excavation geological guarantee for tunnels.
Smart Images

Figure CN115453624B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of mine geophysical exploration data processing and imaging technology, and more particularly to a method for converting and processing explorable sources of mine excavation seismic signals and imaging them. Background Art
[0002] As one of the two core links of coal mine production, the demand for intelligent tunneling is extremely urgent. However, during tunneling, disasters such as coal and gas outbursts and water inrush seriously threaten the safety of tunneling production and the personal safety of miners. Geological support technology is the basis for the safety of intelligent coal production. It is the basic data source for geological prediction, disturbance perception and risk assessment before, during and after tunneling construction. It is the prerequisite for the implementation of all key technologies for intelligent tunneling. Among them, integrating the detection instrument with the tunneling machinery and carrying out on-the-spot detection is one of the most effective ways to ensure the rapid tunneling. It can achieve parallel exploration and precise disaster management, and can give full play to the production efficiency of tunneling machinery. It is the most urgently needed guarantee technology for the rapid intelligent tunneling development in the future. The mine on-the-spot seismic exploration technology has attracted extensive attention from scholars in the industry for its advanced advantages of safety, greenness, and compliance with the scientific concept of intelligent and unmanned mining of coal mines. However, the technology is still in the experimental research stage, and many key basic research has not been broken through.
[0003] Data processing and imaging are one of the most critical links in geophysical exploration technology. Since the excitation mode, wavelet form, energy, frequency and other aspects of the tunnel boring machine seismic source are quite different from conventional seismic sources such as explosives, conventional cognition and data processing technology can no longer achieve results. In addition, the current interpretation results of mine seismic advance detection imaging are mostly "arc drawing", with many false abnormal interfaces and low imaging accuracy, which makes it difficult to meet the geological support needs of intelligent tunneling.
[0004] Therefore, how to provide a high-resolution processing and high-precision imaging method suitable for mine excavation seismic signals is an urgent problem to be solved by those skilled in the art. Summary of the invention
[0005] In view of this, the present invention provides a method for imaging mine seismic signals while excavating, an electronic device and a computer-readable medium, which effectively solve the problem that seismic signals while excavating cannot be directly used for data imaging and interpretation, and realize high-resolution conversion and utilization of explorable sources of seismic signals while excavating and high-precision imaging of abnormal geological structures ahead of excavation.
[0006] In order to achieve the above object, the present invention adopts the following technical solution:
[0007] A method for imaging seismic signals while digging a mine, comprising:
[0008] S1: Acquire continuous excavation seismic signals, and optimize and segment the continuous excavation seismic signals;
[0009] S2: Perform exploratory source conversion processing on the selected and segmented seismic signals during excavation and perform segmented stacking of the processing results;
[0010] S3: pre-processing the exploration source conversion processing and segment stacking results of the excavation seismic signal;
[0011] S4: Perform full waveform inversion imaging of elastic waves for advance detection in tunnel excavation on the preprocessed data.
[0012] Preferably, S2 specifically includes:
[0013] S201: Acquire a reference signal x(t) and a received signal y(t) from the optimized and segmented excavation seismic signals;
[0014] S202: Perform seismic wavelet compression and multi-source stacking feature interference removal on the reference signal x(t) and the received signal y(t) respectively:
[0015]
[0016] In the formula, is the time delay in the process of removing interference from multi-source superposition characteristics, t is the seismic signal time, X(t) and Y(t) are the transition signals in the process of transforming the explorable sources corresponding to x(t) and y(t), is the impulse deconvolution factor under the optimal Wiener filtering constraint corresponding to x(t) and y(t);
[0017] S203: Extract effective signals from the transition signals X(t) and Y(t):
[0018]
[0019] Where τ is the time delay of the effective signal extraction step, t is the seismic signal time, The signal converted into the explorable source;
[0020] S204: Repeat S201-S203 until all segmented excavation seismic signals are processed and the signals converted from the explorable sources are vertically stacked.
[0021] Preferably, S4 specifically includes:
[0022] S401: single-scale inversion is performed on the preprocessed data using a tunneling elastic wave full waveform inversion method based on total variation regularization constraints;
[0023] S402: Perform velocity structure correction on the single-scale inversion result;
[0024] S403: Using the result after wave velocity structure correction as the initial model of the next scale to continue full waveform inversion;
[0025] S404: Repeat S401 to S403 until all scale inversions are completed to obtain the tunnel advance detection imaging results.
[0026] Preferably, S401 specifically includes:
[0027] S4011: Calculate the objective function value and initial gradient direction:
[0028]
[0029] Where E(m) is the objective function, is the gradient direction of the model update, d obs =G obs *w obs (t) and d cal =G cal *w cal (t) are observed data and simulated data, respectively, d obs ref =G obs ref *w obs (t) and d cal ref =G cal ref *w cal (t) are the reference traces selected from the observed data and the simulated data, R′ is the source of the receiving channel other than the reference channel, and R″ is the source of the receiving channel. is the cross-correlation operation; * is the convolution operation, m represents the P-wave and S-wave velocity models and density model; G cal is the simulated data Green’s function, G cal ref is the Green's function of the reference track selected for the simulation data, G obs is the observed data Green’s function, G obs ref is the Green's function of the reference channel selected for the observed data, w cal (t) is the simulated data seismic wavelet, w obs (t) is the observed seismic wavelet; s is the number of earthquake sources, and r is the number of detection points; is the Fréchet derivative;
[0030] S4012: Based on the objective function value and the initial gradient direction, calculate the step length α according to the parabola fitting method and the L-BFGS method l-1and search direction δm l-1 ;
[0031] S4013: Based on step size α l-1 and search direction δm l-1 Calculate the total variation regularization constraint update term L(p N ,q N );
[0032] S4014: According to formula m l =(m l-1 +α l-1 δm l-1 )-λ·L(p N ,q N ) Iteratively update the model, λ is the weight coefficient of the regularization term, m l represents the model parameters of the lth iteration, m l-1 represents the model parameters of the l-1th iteration, and the single-scale inversion result of the full waveform inversion method of elastic wave in tunnel excavation based on total variation regularization constraints is obtained.
[0033] Preferably, S402 specifically includes:
[0034] S4021: Perform multi-parameter weighted constraint structure correction processing on the single-scale inversion results:
[0035]
[0036] In the formula, Δm i,j ε is the single iteration model update amount, is the model update amount after multi-parameter weighted constraint structure correction, ε=V p ,V s , Den represents the medium longitudinal wave velocity, shear wave velocity, and density physical parameters, nx is the number of horizontal grids in the model, nz is the number of vertical grids in the model, and m i,j ε is a multi-parameter initial model, i and j represent the positions of the grid nodes in the z and x directions respectively;
[0037] S4022: Perform one-dimensional velocity profile spatial structure correction and smoothing constraint on the multi-parameter weighted constraint structure correction processing result.
[0038] Preferably, S4022 specifically includes:
[0039] Step a: Set the constraints of the spatial structure correction of the one-dimensional wave velocity profile, including the lower limit value of the spatial correction, the range of the structural correction area, and the distance between the areas;
[0040] Step b: Select a vertical axis coordinate y i, extract the one-dimensional wave velocity profile along the tunnel axis x is the horizontal axis coordinate;
[0041] Step c: Calculate the one-dimensional wave velocity profile The corresponding model update amount With slope
[0042] Step d: Update the model below the spatial correction lower limit value Suppress it and change it directly to 0 to get the corrected model update amount
[0043] Step e: Update the model according to the correction Determine each grid point on the one-dimensional wave velocity profile one by one:
[0044] When there is a positive update area on both sides of the negative update area, or when there is a negative update area on both sides of the positive update area, and when the area range and the distance between areas meet the structural correction area range and the distance between areas, the model update amount in the positive area on both sides of the negative area or the negative area on both sides of the positive area is suppressed to 1 / 5 of the original model update value, and the corrected model update amount is obtained.
[0045] Step f: Update the model according to the correction With slope Determine each grid point on the one-dimensional wave velocity profile one by one:
[0046] When there are two points between the three adjacent points where the slope sign changes, the model update amount of the grid points before and after the two points changes in sign, and there is no point with a slope of 0 between the two points, the model update amount of the grid points in the area between the three adjacent points where the slope sign changes is corrected to the average of the model update amounts of the first slope sign change point and the third slope sign change point, and the corrected model update amount is obtained
[0047] Step g: Update the model according to the correction With slope Determine each grid point on the one-dimensional wave velocity profile one by one:
[0048] When the model update signs of all grid points in the range between two non-adjacent points where the slope sign changes do not change, the model update of the grid points in the area between the two slope sign change points is corrected to the average of the model update of all grid points in the area, and the corrected model update is obtained.
[0049] Step h: Update the model according to the correction Calculate the corrected one-dimensional velocity profile
[0050] Step i: Get the next vertical axis grid y i+1 , repeat steps b to h until the correction of the one-dimensional velocity profiles of all grids is completed, and the result after velocity structure correction is obtained.
[0051] Preferably, S1 specifically includes:
[0052] S101: Acquire continuous excavation seismic signals, calculate average amplitude energy of the continuous excavation seismic signals, and remove excavation seismic signals with amplitude energy lower than the average;
[0053] S102: Segmentally process the excavation seismic signal selected in S101 according to a certain time window.
[0054] Preferably, S3 specifically includes:
[0055] S301: The single shot record for exploration is shifted back half a period on the time axis as a whole, and waveform correction is performed to obtain a conventional single shot record form;
[0056] S302: Perform trigger delay correction according to the physical property conditions of the detection medium.
[0057] A computer-readable medium stores instructions, and when the instructions are executed on the readable medium, the readable medium executes the method for converting and processing explorable sources of mine excavation seismic signals and for imaging the same.
[0058] An electronic device includes a memory and a processor, wherein the memory stores a computer program that can be run on the processor, and is characterized in that when the processor executes the computer program, the explorable source conversion processing and imaging method of mine excavation seismic signals is implemented.
[0059] The present invention has the following advantages:
[0060] (1) The present invention realizes high-resolution utilization of explorable sources of excavation seismic signals;
[0061] (2) The present invention solves the problem of inaccurate amplitude and phase information of seismic signals caused by the conversion process of excavation seismic signals to explorable sources, and eliminates the impact of the conversion results on inversion imaging;
[0062] (3) The present invention solves the problem that convolution and cross-correlation operations introduce abnormal noise and increase in inversion nonlinearity during inversion imaging;
[0063] (4) The present invention gives full play to the advantages of seismic detection while tunneling, which can improve the net production efficiency of tunneling by more than 10%, solve the problems of small amount of detection data and strong multi-solution of waveform inversion in tunnel advance detection, and improve the imaging accuracy;
[0064] (5) The present invention realizes high-precision imaging of abnormal geological structures ahead of tunnel excavation. The imaging results can accurately determine the location and occurrence of geological anomalies, and can meet the geological guarantee requirements for intelligent tunnel excavation. BRIEF DESCRIPTION OF THE DRAWINGS
[0065] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying creative work.
[0066] Figure 1 The accompanying drawing is a flow chart of a method for imaging seismic signals during mine excavation provided by the present invention.
[0067] Figure 2 The accompanying drawings are the seismic observation system for tunneling provided by the present invention, wherein: 1. the detector is arranged in the middle of the tunnel wall; 2. the detector is coupled with the support anchor; 3. the tunnel; 4. the detector; 5. the fault; 6. the coal seam; 7. the cutting head of the longitudinal cantilever type tunneling machine.
[0068] Figure 3 The attached figure is a diagram of the original mine excavation seismic signal provided by the present invention, wherein Figure 3 (a) represents the X component of the original mine excavation seismic signal, Figure 3 (b) represents the Z component of the original mine excavation seismic signal.
[0069] Figure 4 The attached figure shows the result of the processing of the original excavation seismic signal to the exploration source provided by the present invention, wherein Figure 4 (a) shows the result of the conversion of the X component of the original mine excavation seismic signal into the explorable source. Figure 4 (b) shows the result of the conversion of the Z component of the original mine excavation seismic signal into an explorable source.
[0070] Figure 5 The accompanying figure shows the comparison between the full waveform inversion result of the present invention and the measured profile, wherein Figure 5 (a) is the longitudinal wave inversion result, Figure 5 (b) is the shear wave inversion result. Figure 5 (c) is the density inversion result. DETAILED DESCRIPTION
[0071] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0072] The embodiment of the present invention discloses a method for imaging seismic signals while digging a mine. Figure 1 As shown, including:
[0073] S1: Acquire continuous excavation seismic signals, and optimize and segment the continuous excavation seismic signals;
[0074] S2: Perform exploratory source conversion processing on the selected and segmented seismic signals during excavation and perform segmented stacking of the processing results;
[0075] S3: pre-processing the exploration source conversion processing and segment stacking results of the excavation seismic signal;
[0076] S4: Perform full waveform inversion imaging of elastic waves for advance detection in tunnel excavation on the preprocessed data.
[0077] In this embodiment, S1 specifically includes:
[0078] S101: High-frequency continuous dynamic acquisition of excavation seismic signals to obtain continuous excavation seismic signals, such as Figure 2 As shown:
[0079] (1) The seismic observation system is arranged linearly, with 24 detectors placed in the middle of the tunnel wall at a certain interval. The detectors are fixed on the anchor rods supporting the tunnel wall.
[0080] (2) Sampling parameter setting: The sampling parameters of the drilling seismic exploration technology are consistent with those of the conventional active source detection technology, and the sampling frequency is generally 10kHz to 1kHz.
[0081] The high-frequency, multi-coverage, high-precision advance seismic detection technology provided by the present invention is designed to use 24 detection points and 6 shot points to perform full waveform inversion imaging once a day, and then perform high-frequency detection for many consecutive days. The imaging results are multi-covered and comprehensively identified for anomalies to improve the detection accuracy.
[0082] The tunneling seismic signal mainly comes from the strong vibration signal generated when the cutting head cuts the coal seam. However, during the operation of the tunneling machine, in addition to cutting the coal seam, there are also states such as idling and stopping. At this time, the collected data is invalid data. Among them, when the tunneling machine cuts the coal seam normally, it will generate a large excitation energy, which is several times that of idling and stopping. Therefore, the present invention optimizes the tunneling seismic signal according to the average amplitude energy of the collected signal, and eliminates the tunneling seismic signal with an amplitude energy lower than the average.
[0083] S102: The excavation seismic signals selected in S101 are processed in sections according to a certain time window, and then the final results are superimposed. The selection of the time window is determined according to the actual situation, and the present invention does not limit it.
[0084] In this embodiment, S2 specifically includes:
[0085] S201: Acquire a reference signal x(t) and a received signal y(t) from the optimized and segmented excavation seismic signals;
[0086] S202: Perform seismic wavelet compression and multi-source stacking feature interference removal on the reference signal x(t) and the received signal y(t) respectively:
[0087]
[0088] In the formula, is the time delay in the process of removing interference from multi-source superposition characteristics, t is the seismic signal time, X(t) and Y(t) are the transition signals in the process of transforming the explorable sources corresponding to x(t) and y(t), is the impulse deconvolution factor under the optimal Wiener filtering constraint corresponding to x(t) and y(t);
[0089] S203: Extract effective signals from the transition signals X(t) and Y(t):
[0090]
[0091] Right now
[0092] Where τ is the time delay of the signal extraction step, t is the seismic signal time, The signal converted into the explorable source;
[0093] Furthermore, due to the complexity of the seismic wave field during excavation, the signal-to-noise ratio of the seismic record may vary due to different field conditions. When the resolution of the algorithm processing result is lower than the preset ideal threshold, S202 is executed again to obtain a better processing result.
[0094] S204: Repeat S201-S203 until all segmented excavation seismic signal processing is completed, and vertically stack the signals converted from the explorable sources to obtain the final high-resolution results.
[0095] In this embodiment, S3 specifically includes:
[0096] S301: Since the processing algorithm for converting the excavation seismic signal into an explorable source is mainly based on the principle of seismic interference, the exploratory single shot record is shifted back half a period on the time axis as a whole, and waveform correction is performed to obtain a conventional single shot record form;
[0097] S302: The single-shot recording form obtained after processing is zero-offset seismic data with the reference seismic trace as the virtual source, and trigger delay correction is performed according to the physical property conditions of the detection medium before imaging.
[0098] In this embodiment, S4 specifically includes:
[0099] S401: Perform single-scale inversion on the preprocessed data using a full waveform inversion method for elastic waves in tunneling under total variation regularization constraints;
[0100] S402: Perform velocity structure correction on the single-scale inversion result;
[0101] S403: Using the result after wave velocity structure correction as the initial model of the next scale to continue full waveform inversion;
[0102] S404: Repeat S401 to S403 until all scale inversions are completed to obtain the tunnel advance detection imaging results.
[0103] S401 specifically includes:
[0104] S4011: Calculate the objective function value and initial gradient direction:
[0105]
[0106] Where E(m) is the objective function, is the gradient direction of the model update, d obs =G obs *w obs (t) and d cal =G cal *w cal (t) are observed data and simulated data, respectively, d obs ref =G obs ref *w obs (t) and d cal ref=G cal ref *w cal (t) are the reference traces selected from the observed data and the simulated data, δR = d cal *d obs ref -d obs *d cal ref , R′ is the source of the receiving channel other than the reference channel, and R″ is the source of the receiving channel. is the cross-correlation operation; * is the convolution operation, m represents the P-wave and S-wave velocity models and density model; G cal is the simulated data Green’s function, G cal ref is the Green's function of the reference track selected for the simulation data, G obs is the observed data Green’s function, G obs ref is the Green's function of the reference channel selected for the observed data, w cal (t) is the simulated data seismic wavelet, w obs (t) is the observed seismic wavelet; s is the number of earthquake sources, and r is the number of detection points; is the Fréchet derivative;
[0107] S4012: Based on the objective function value and the initial gradient direction, calculate the step length α according to the parabola fitting method and the L-BFGS method l-1 and search direction δm l-1 ;
[0108] S4013: Based on step size α l-1 and search direction δm l-1 Calculate the total variation regularization constraint update term L(p N ,q N ); the specific process is:
[0109]
[0110] Where p and q are regularization constraint operators, N is the maximum number of calculation stacking times of the total variation regularization constraint update term in a single iteration of full waveform inversion, i = 1…nz, j = 1…nx, nx is the number of horizontal grids in the model, and nz is the number of vertical grids in the model.
[0111] and pass and Obtained from multiple superposition calculations of e=1…N.
[0112] Where:
[0113] Where:
[0114] Where:
[0115] Where:
[0116] Where:
[0117] In the formula, let is a zero matrix, h is a regularized update operator, and C and D are regularized iteration operators.
[0118] S4014: According to formula m l =(m l-1 +α l-1 δm l-1 )-λ·L(p N ,q N ) Iteratively update the model, λ is the weight coefficient of the regularization term, m l represents the model parameters of the lth iteration, m l-1 represents the model parameters of the l-1th iteration, and the single-scale inversion result of the convolution-type wavelet-independent tunnel elastic wave full waveform inversion method under the total variation regularization constraint is obtained.
[0119] S402 specifically includes:
[0120] S4021: Perform multi-parameter weighted constraint structure correction processing on the single-scale inversion results:
[0121]
[0122] In the formula, Δm i,j ε is the single iteration model update amount, is the model update amount after multi-parameter weighted constraint structure correction, ε=V p ,V s , Den represents the medium longitudinal wave velocity, shear wave velocity, and density physical parameters, nx is the number of horizontal grids in the model, nz is the number of vertical grids in the model, and m i,j ε is a multi-parameter initial model, i and j represent the positions of the grid nodes in the z and x directions respectively;
[0123] S4022: Perform one-dimensional velocity profile spatial structure correction and smoothing constraint on the multi-parameter weighted constraint structure correction processing result, specifically:
[0124] Step a: Set the constraints of the spatial structure correction of the one-dimensional wave velocity profile, including the lower limit value of the spatial correction, the range of the structural correction area, and the distance between the areas;
[0125] Step b: Select a vertical axis coordinate y i , extract the one-dimensional wave velocity profile along the tunnel axis x is the horizontal axis coordinate;
[0126] Step c: Calculate the one-dimensional wave velocity profile The corresponding model update amount With slope
[0127] Step d: Update the model below the spatial correction lower limit value Suppress it and change it directly to 0 to get the corrected model update amount
[0128] Step e: Update the model according to the correction Determine each grid point on the one-dimensional wave velocity profile one by one:
[0129] When there is a positive update area on both sides of the negative update area, or when there is a negative update area on both sides of the positive update area, and when the area range and the distance between areas meet the structural correction area range and the distance between areas, the model update amount in the positive area on both sides of the negative area or the negative area on both sides of the positive area is suppressed to 1 / 5 of the original model update value, and the corrected model update amount is obtained.
[0130] Step f: Update the model according to the correction With slope Determine each grid point on the one-dimensional wave velocity profile one by one:
[0131] When there are two points between the three adjacent points where the slope sign changes, the model update amount of the grid points before and after the two points changes in sign, and there is no point with a slope of 0 between the two points, the model update amount of the grid points in the area between the three adjacent points where the slope sign changes is corrected to the average of the model update amounts of the first slope sign change point and the third slope sign change point, and the corrected model update amount is obtained
[0132] Step g: Update the model according to the correction With slope Determine each grid point on the one-dimensional wave velocity profile one by one:
[0133] When the model update signs of all grid points in the range between two non-adjacent points where the slope sign changes do not change, the model update of the grid points in the area between the two slope sign change points is corrected to the average of the model update of all grid points in the area, and the corrected model update is obtained.
[0134] Step h: Update the model according to the correction Calculate the corrected one-dimensional velocity profile
[0135] Step i: Get the next vertical axis grid y i+1 , repeat steps b to h until the correction of the one-dimensional velocity profiles of all grids is completed, and the result after velocity structure correction is obtained.
[0136] The structural correction based on multi-parameter weighted constraints and the one-dimensional velocity profile spatial structure correction and smoothness constraint method cannot directly participate in every iteration of full waveform inversion. They need to be combined with multi-scale inversion to perform structural correction on the results after single-scale inversion, and use the structurally corrected model as the initial model of the next scale to continue full waveform inversion. The total variation regularization constraint method participates in every iteration of full waveform inversion.
[0137] The convolution-type full-waveform inversion method that does not rely on wavelets solves the problem of inaccurate amplitude and phase information of seismic signals caused by the conversion processing algorithm of the excavation seismic signal explorable source, and eliminates the influence of the conversion result on the inversion imaging. The method solves the problem that the inversion imaging method introduces some noise in the convolution and cross-correlation operation process, and causes the nonlinearity of the full-waveform inversion to increase, thereby increasing the sensitivity to the existence of local minima, and solves the problem that the observation system of the tunnel seismic advance detection mode is highly restrictive, which leads to the enhancement of the multi-solution of the full-waveform inversion and the increase of the calculation error of the physical property parameters, thereby improving the inversion accuracy.
[0138] In order to verify the effectiveness and efficiency of the method for converting and imaging the explorable source of seismic signals during mine excavation, the scheme proposed in the present invention is used to Figure 3 As shown in the figure, the excavation seismic signal can be converted into the explorable source and the full waveform inversion imaging result is obtained. Figures 4-5 shown.
[0139] Figure 3 After the acquisition of the seismic signal during excavation was completed, the test tunnel continued to be excavated, and the full waveform inversion imaging results were compared with the actual exposure results of the tunnel. Figure 5Results shown. After tunnel excavation, no abnormal geological body was found in front of the excavation, and the exposure results were consistent with the full waveform inversion imaging results; in addition, the abnormal area ① of the coal seam roof in the full waveform inversion imaging results was close to the head position. It was analyzed that the rock layer of the coal seam roof near the head position was subjected to external force due to the construction of the tunnel boring machine, resulting in rock mass crushing; the above results verify the effectiveness and efficiency of the method of the present invention.
[0140] In this specification, each embodiment is described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the embodiments can be referred to each other. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part.
[0141] The above description of the disclosed embodiments enables one skilled in the art to implement or use the present invention. Various modifications to these embodiments will be apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.
Claims
1. A method for imaging seismic signals during mine excavation, characterized in that: include: S1: Acquire continuous excavation seismic signals, and optimize and segment the continuous excavation seismic signals; S2: Perform exploratory source conversion processing on the selected and segmented seismic signals during excavation and perform segmented stacking of the processing results; S3: pre-processing the exploration source conversion processing and segment stacking results of the excavation seismic signal; S4: Perform full waveform inversion imaging of elastic waves for advance detection of tunneling on the pre-processed data; S4 specifically includes: S401: single-scale inversion is performed on the preprocessed data using a tunneling elastic wave full waveform inversion method based on total variation regularization constraints; S402: Perform velocity structure correction on the single-scale inversion result; S403: Using the result after wave velocity structure correction as the initial model of the next scale to continue full waveform inversion; S404: Repeat S401 to S403 until all scale inversions are completed to obtain the tunnel advance detection imaging results.
2. A mine excavation seismic signal imaging method according to claim 1, characterized in that: S2 specifically includes: S201: Acquire a reference signal x(t) and a received signal y(t) from the optimized and segmented excavation seismic signals; S202: Perform seismic wavelet compression and multi-source stacking feature interference removal on the reference signal x(t) and the received signal y(t) respectively: In the formula, is the time delay in the process of removing interference from multi-source superposition characteristics, t is the seismic signal time, X(t) and Y(t) are the transition signals in the process of transforming the explorable sources corresponding to x(t) and y(t), is the impulse deconvolution factor under the optimal Wiener filtering constraint corresponding to x(t) and y(t); S203: Extract effective signals from the transition signals X(t) and Y(t): In the formula, τ is the time delay of the effective signal extraction step, t is the seismic signal time, The signal converted into the explorable source; S204: Repeat S201-S203 until all segmented excavation seismic signals are processed and the signals converted from the explorable sources are vertically stacked.
3. The method for imaging seismic signals while digging a mine according to claim 1, characterized in that: S401 specifically includes: S4011: Calculate the objective function value and initial gradient direction: Where E(m) is the objective function, is the gradient direction of the model update, d obs =G obs *w obs (t) and d cal =G cal *w cal (t) are observed data and simulated data, respectively, d obs ref =G obs ref *w obs (t) and d cal ref =G cal ref *w cal (t) are the reference traces selected from the observed data and the simulated data, δR = d cal *d obs ref -d obs *d cal ref , R′ is the source of the receiving channel other than the reference channel, and R″ is the source of the receiving channel. is the cross-correlation operation; * is the convolution operation, m represents the P-wave and S-wave velocity models and density model; G cal is the simulated data Green’s function, G cal ref is the Green's function of the reference track selected for the simulation data, G obs is the observed data Green’s function, G obs ref is the Green's function of the reference channel selected for the observed data, w cal (t) is the simulated data seismic wavelet, w obs (t) is the observed seismic wavelet; s is the number of earthquake sources, and r is the number of detection points; is the Fréchet derivative; S4012: Based on the objective function value and the initial gradient direction, calculate the step length α according to the parabola fitting method and the L-BFGS method l-1 and search direction δm l-1 ; S4013: Based on step size α l-1 and search direction δm l-1 Calculate the total variation regularization constraint update term L(p N ,q N ); S4014: According to formula m l =(m l-1 +α l-1 δm l-1 )-λ·L(p N ,q N ) Iteratively update the model, λ is the weight coefficient of the regularization term, m l represents the model parameters of the lth iteration, m l-1 represents the model parameters of the l-1th iteration, and the single-scale inversion result of the full waveform inversion method of elastic wave in tunnel excavation based on total variation regularization constraints is obtained.
4. The method for imaging seismic signals while digging a mine according to claim 1, characterized in that: S402 specifically includes: S4021: Perform multi-parameter weighted constraint structure correction processing on the single-scale inversion results: In the formula, Δm i,j ε is the single iteration model update amount, Δm′ i,j ε is the model update amount after multi-parameter weighted constraint structure correction, ε=V p ,V s , Den represents the medium longitudinal wave velocity, shear wave velocity, and density physical parameters, nx is the number of horizontal grids in the model, nz is the number of vertical grids in the model, and m i,j ε is a multi-parameter initial model, i and j represent the positions of the grid nodes in the z and x directions respectively; S4022: Perform one-dimensional velocity profile spatial structure correction and smoothing constraint on the multi-parameter weighted constraint structure correction processing result.
5. A mine excavation seismic signal imaging method according to claim 4, characterized in that: S4022 specifically includes: Step a: Set the constraints of the spatial structure correction of the one-dimensional wave velocity profile, including the lower limit value of the spatial correction, the range of the structural correction area, and the distance between the areas; Step b: Select a vertical axis coordinate y i , extract the one-dimensional wave velocity profile along the tunnel axis x is the horizontal axis coordinate; Step c: Calculate the one-dimensional wave velocity profile The corresponding model update amount With slope Step d: Update the model below the spatial correction lower limit value Suppress it and change it directly to 0 to get the corrected model update amount Step e: Update the model according to the correction Determine each grid point on the one-dimensional wave velocity profile one by one: When there is a positive update area on both sides of the negative update area, or when there is a negative update area on both sides of the positive update area, and when the area range and the distance between areas meet the structural correction area range and the distance between areas, the model update amount in the positive area on both sides of the negative area or the negative area on both sides of the positive area is suppressed to 1 / 5 of the original model update value, and the corrected model update amount is obtained. Step f: Update the model according to the correction With slope Determine each grid point on the one-dimensional wave velocity profile one by one: When there are two points between the three adjacent points where the slope sign changes, the model update amount of the grid points before and after the two points changes in sign, and there is no point with a slope of 0 between the two points, the model update amount of the grid points in the area between the three adjacent points where the slope sign changes is corrected to the average of the model update amounts of the first slope sign change point and the third slope sign change point, and the corrected model update amount is obtained Step g: Update the model according to the correction With slope Determine each grid point on the one-dimensional wave velocity profile one by one: When the model update signs of all grid points between two non-adjacent points where the slope sign changes do not change, the model update of the grid points in the area between the two slope sign change points is corrected to the average of the model update of all grid points in the area, and the corrected model update is obtained. Step h: Update the model according to the correction Calculate the corrected one-dimensional velocity profile Step i: Get the next vertical axis grid y i+1 , repeat steps b to h until the correction of the one-dimensional velocity profiles of all grids is completed, and the result after velocity structure correction is obtained.
6. The method for imaging seismic signals while digging a mine according to claim 1, characterized in that: S1 specifically includes: S101: Acquire continuous excavation seismic signals, calculate average amplitude energy of the continuous excavation seismic signals, and remove excavation seismic signals with amplitude energy lower than the average; S102: Segmentally process the excavation seismic signal selected in S101 according to a certain time window.
7. The method for imaging seismic signals during mine excavation according to claim 1, characterized in that: S3 specifically includes: S301: The single shot record for exploration is shifted back half a period on the time axis as a whole, and waveform correction is performed to obtain a conventional single shot record form; S302: Perform trigger delay correction according to the physical property conditions of the detection medium.
8. A computer readable medium storing instructions, characterized in that: When the instructions are executed on the readable medium, the readable medium is caused to execute the method according to any one of claims 1 to 7.
9. An electronic device comprising a memory and a processor, wherein the memory stores a computer program that can be run on the processor, characterized in that: When the processor executes the computer program, the method according to any one of claims 1 to 7 is implemented.
Citation Information
Patent Citations
Method for detecting geological structures of coal mine excavation roadway in various directions by using excavation signals
CN106772557A
Multipoint geostatistical pre-stack inversion method based on updated theory of permanence of probability ratio
WO2021147529A1
