Dam hidden danger continuous towed seismic rapid general investigation detection method

Abstract

The application discloses a dam hidden danger continuous dragging type earthquake rapid general investigation and detection method, which comprises the following steps: step A: collecting vibration noise data in the process of moving through gravity acceleration detectors, the vibration generated by the friction between the gravity acceleration detector base and the ground during continuous dragging movement is the vibration source; step B: extracting information from the real near-surface underground structure from the vibration noise data by using the common receiving point stacking method; step C: stacking the data of multiple detections to suppress interference noise and improve the energy of the real signal. The application breaks through the hard requirement that the conventional seismic data collection can only be collected when the detector is stationary, fully utilizes the vibration signal generated by the friction between the base of the gravity acceleration detector and the hardened road surface as the noise source when the gravity acceleration detector is moving, and collects the seismic signal in the process of dragging movement. The application of the content of the application to the rapid general investigation of dam hidden dangers can preliminarily define the approximate position of the hidden dangers.

Description

TECHNICAL FIELD

[0001] The present application relates to the dam hidden seismic detection and engineering surveying technical field, in particular to a dam hidden continuous towed seismic rapid survey detection method. BACKGROUND

[0002] 1. Research significance and research status of continuous towed seismic data acquisition:

[0003] The engineering technical field seismic data acquisition is mainly divided into active source and passive source. The active source method develops earlier, and rises with the development of seismology. Since the early application of seismology in engineering surveying field, it has been widely applied. For this reason, various types of seismic sources have been developed, such as 24-pound, 48-pound hammer, electromagnetic impact seismic source (Richter et al., 2018), mechanical impact seismic source (Yordkayhun et al., 2012), sweep frequency seismic source, downhole mechanical seismic source (Liu et al., 2013), electric spark seismic source (Casey et al., 2002) and so on. The active source data acquisition method is to set up the seismic wave detector according to the designed observation system, and to excite the seismic wave at the specified seismic source position. The active source method has high requirements for the site, and the detector needs to be laid in a certain site space. The cost of single shot excitation is high, and the data acquisition efficiency is low. In mountainous environment, the efficiency of collecting single shot data by using cone velocity type detector is 24 points per day on average, and each point is excited 2 times on average. When hardening the road surface, the average efficiency of collecting single shot data by using gravity acceleration detector is 0.5km / h. The passive source method data acquisition has low requirements for the site, and can be laid into linear or circular array to collect noise data, but the observation time of each array point is between 10min to 30min. The active source method has high detection accuracy, but has high requirements for the seismic source, detector and detection site. The passive source method has low detection accuracy and can provide macro-scale information. The site requirement is small, and the data acquisition time is long. At present, the active source and passive source methods developed at this stage are suitable for fine survey of dam hidden dangers, but they are not suitable for rapid survey requirements in the face of large-scale dam and road safety detection task requirements.

[0004] The present application aims to propose a continuous towed seismic data acquisition method, which breaks through the hard requirement that the conventional seismic data acquisition can only be collected when the detector is stationary, fully utilizes the vibration signal generated by the friction between the base of the gravity acceleration detector and the hardening road surface as the noise source, and collects the seismic signal during the towed travel of the gravity acceleration detector. The present application is applied to the rapid survey of dam hidden dangers, which can preliminarily delineate the approximate position of the hidden danger.

[0005] 2. Research significance and research status of related stacking method:

[0006] Since Aki (1957) proposed the spatial auto corelation (SPAC) technique of microseism, the SPAC technique has been widely used in the identification of repeating earthquakes (Schaff, Richards, 2004, 2011; Li al., 2007, 2011; Ma et al., 2014), the detection of aftershock events (Peng, Zhao, 2009; Wu et al., 2014), the observation and identification of low-frequency events (Obara, 2002; Shelly et al., 2007), the precise positioning of earthquakes (Waldhauser, Ellsworth, 2000; Schaff et al., 2004; Schaff, Waldhauser, 2005) and the field of engineering earthquakes. Boaga et al., (2010) applied the cross-correlation calculation method to extract the dispersion curve of surface wave from seismic noise. Mordret et al., (2010) applied the cross-correlation function interferometry method in environmental noise to monitor the time-varying characteristics of the elastic properties of Llao Pehu volcano. Zhu Liangbao et al. (2011) derived the theoretical expression of the cross-correlation function of seismic background noise (NCF) based on the elastic dynamic surface wave excitation formula, and proved that NCF is equivalent to the source excitation formula of surface wave. Pastén et al., (2016) used the cross-correlation calculation of continuous recording of seismic ambient noise to study the medium-deep structure of the Santiago Basin. Ma Tengfei (2016) proposed a new simple method suitable for multi-element cross-correlation operation of three-component seismic waveform records based on the cross-correlation formula of single-component seismic records, and verified its effect by taking the waveform of the Wenchuan JII Ms8.0 earthquake aftershock sequence in 2008 as an example. Li Shilin et al. (2016) obtained the body wave Green's function between near-aperture stations by noise cross-correlation method, which can clearly observe different seismic phases, and believe that the extracted Green's function contains propagation path information, which can be combined with tomographic imaging technology to study the structure of the earth's interior. Younes et al., (2019) studied the underground structure characteristics of the back of Morocco by cross-correlation calculation method on the data of several temporary seismic stations. Zhang Hanlan et al. (2019) used seismic interference technology to process high-speed rail seismic signals and imaging. Wang et al., (2020) used the broadband seismic noise data of several temporary and permanent arrays in Northeast China to re-study the crust and upper mantle shear velocity structure of Northeast China, mainly using the seismic correlation interference method. Yin Xingyao et al. (2020) developed a multi-channel inversion method driven by seismic data cross-correlation, introduced a local optimization operator into the penalty term of the target function to improve the influence of the poor lateral continuity of seismic data itself on the inversion result, and constructed a multi-channel seismic inversion target function that is easy to solve.Zhang et al.(2020) proposed a segmented waveform cross-correlation method, which segmented the simultaneous seismic data, extracted the travel time of the effective signal by cross-correlation method, and then used velocity tomography method to image the internal structure of the working face and the front of the cut. Mathieu et al(2021) systematically summarized the motion law of landslide body monitored by passive source cross-correlation noise and the future application prospect.

[0007] Continuous towed seismic data is similar to noise data, but the towed arrangement is collected while traveling, and the source point and receiver point are not fixed. The existing cross-correlation calculation method cannot be directly used to extract near-surface underground reflection information from noise data, and a common receiver correlation stacking calculation method suitable for continuous towed seismic data needs to be derived. The present application aims to propose a common receiver correlation stacking algorithm for extracting effective reflection signals from continuous towed seismic data, and to realize the extraction of effective reflection information from the near-surface from massive noise data. SUMMARY

[0008] In view of at least one defect of the prior art, the purpose of the present application is to provide a dam hidden danger continuous towed seismic rapid general survey detection method, which breaks through the hard requirement that conventional seismic data acquisition can only be collected when the receiver is stationary, fully utilizes the vibration signal generated by the friction between the base of the gravity acceleration receiver and the hardened road surface as a noise source when the gravity acceleration receiver is traveling, and collects seismic signals during the towed traveling process. The present application can be applied to the rapid general survey of dam hidden dangers, and can preliminarily delineate the approximate position of the hidden danger.

[0009] In order to achieve the above purpose, the present application adopts the following technical scheme: a dam hidden danger continuous towed seismic rapid general survey detection method, comprising the following steps:

[0010] Step A: simultaneously collecting vibration noise data by gravity acceleration receivers during traveling, and the vibration generated by the friction between the base of the gravity acceleration receiver and the ground during continuous towed traveling as the source;

[0011] Step B: extracting information from the real near-surface underground structure from the vibration noise data by using common receiver stacking method;

[0012] Step C: stacking the data of multiple detections to suppress interference noise and improve the energy of the real signal.

[0013] The common receiver stacking method is used in step B to extract information from the real near-surface underground structure from the vibration noise data, which includes the following contents:

[0014] D gravity acceleration detectors are simultaneously moved to collect, to constitute multiple coverage observation system, the arrangement length is K, the arrangement moving number is M times, and the arrangement post number is 1-K+M-1;Suppose that the recording time length of each time of arrangement is t1, t2,...... M-1 ,t M The total recording time of multiple coverage is The total observation field value at the position is U (x, T) = [u (x, t1), u (x, t2),...., u (x, t M )], i = 1, 2,...., M;u (x, ti) is the observation field value at x position in ti time period, i = 1, 2,...., M.

[0015] If the set window length is w, the iteration number of point stack calculation is N = T / w, and the calculation formula is derived as follows:

[0016]

[0017] Wherein, l = 1, 2,...., N, s = w*l, (w+1)*l,...., (2w-1)*l, 2*w*l;ψ a (x, t) is the function of x point position observation record U (x, s) and U (x, s) after multiplication and summation, N times iteration and normalization.

[0018] Said D is 24 or 48.

[0019] Significant effect: the application provides a dam hidden danger continuous dragging type seismic rapid general survey detection method, breaks the hard requirement that only when the detector is static can the conventional seismic data be collected, fully utilizes the vibration signal generated when the base of gravity acceleration detector rubs with hardened pavement as a noise source, and collects seismic signals in the dragging process of gravity acceleration detector. The content of the application is applied to rapid general survey of dam hidden danger, and the approximate position of hidden danger can be preliminarily circled. BRIEF DESCRIPTION OF DRAWINGS

[0020] Figure 1 It is a schematic diagram of continuous dragging type seismic observation system and common receiving point;

[0021] Figure 2 It is a picture of gravity acceleration detector, (a) and (b) are structural diagrams of gravity acceleration detector, and (c) is a real object diagram of gravity acceleration detector;

[0022] Figure 3 It is a panel schematic diagram of continuous dragging type seismic instrument host system;

[0023] Figure 4Fig. 1 is a schematic diagram of a velocity model, in which A, B and C are the positions of geophones;

[0024] Figure 5 Fig. 2 is a waveform diagram of a random noise source;

[0025] Figure 6 Fig. 3 is a schematic diagram of simulated records of the source arranged at different positions, in which (a), (b) and (c) respectively correspond to Figure 4 A, B and C in the model;

[0026] Figure 7 Fig. 4 is a schematic diagram of a simulated 8-fold stack section;

[0027] Figure 8 Fig. 5 is a schematic diagram of a multiple stack section;

[0028] Figure 9 Fig. 6 is a photograph of continuous towed seismic data acquisition, in which (a) is a line layout diagram; (b) is a towed array movement; and (c) is original seismic data;

[0029] Figure 10 Fig. 7 is a schematic diagram of (a) and (b) active source records at different positions; (c) and (d) are original continuous towed seismic records at different positions;

[0030] Figure 11 Fig. 8 is a schematic diagram of (a) a spectral analysis of original seismic records; (a)-(d) correspond to Figure 10 For comparison, the amplitude of curves (a) and (b) is enlarged by 1.0e7 times, (b) and (c) are Figure 12 surface wave spectra of (a) and (b);

[0031] Figure 12 Fig. 9 is a schematic diagram of data processing results of 6 detections after trace equalization processing;

[0032] Figure 13 Fig. 10 is a schematic diagram of (a) an inverse scattering section; (b) a surface wave apparent velocity section; (c) an active source data stack section; and (d) a stack section of 6 detections of continuous towed seismic data;

[0033] Figure 14 Fig. 11 is a flow chart of the method of the present application. DETAILED DESCRIPTION

[0034] The present application will be further described in detail below in combination with the drawings and specific examples.

[0035] As Figures 1-14As shown, the present invention relates to a detection method for a continuous dragging seismic rapid survey system for (embankment) dam (road) hidden dangers. When the required lateral resolution is 0.5m, the detection efficiency can reach 5km / h. The specific contents are: 1) The vibration generated by the friction between the base of the gravity acceleration detector and the ground is used as the earthquake source to realize the rapid acquisition of seismic data of dam hidden dangers, and solve the problem of large-scale rapid survey of dam hidden dangers (≤20m) by seismic methods; 2) The common receiving point superposition method is used to extract high-resolution superposition profiles; 3) The continuous dragging seismic acquisition method has low acquisition cost and can realize multiple rapid repeated detections on the same survey line. The superposition processing of multiple detection data can improve the resolution of the superposition profile and realize clear resolution of near-surface targets.

[0036] The present invention provides a rapid earthquake survey system for hidden dangers in dams. The main advantages are as follows: 1) Compared with the traditional seismic data acquisition method of measuring earthquake data after the detector is stationary, the continuous dragging seismic data acquisition method does not require the gravity acceleration detector to be stationary, and data acquisition can be achieved while the gravity acceleration detector is moving, which significantly improves the data acquisition efficiency; 2) Traditional seismic data acquisition methods mainly include active sources and passive sources. The active source data acquisition method has a high signal-to-noise ratio, but low efficiency, while the passive source data acquisition method is less affected by the site, but requires long-term fixed-point observation, which is time-consuming. The present invention provides a data acquisition method between the active source and the passive source. It does not require additional configuration of the seismic source during data acquisition, and uses the vibration generated by the friction between the gravity acceleration detector base and the ground as the seismic source. This device greatly simplifies the seismic data acquisition equipment for dam road hardened pavement environments, improving data acquisition efficiency. 3) A common receiving point superposition method is proposed to extract real information from near-surface underground structures from continuously dragged noise data. 4) The rapid survey system proposed by this invention can fully realize on-the-go data acquisition. When the lateral resolution of a single survey line is 0.5m, the acquisition efficiency can reach 5km / h. The efficient data acquisition method and high-precision RTK positioning system can allow multiple reciprocating surveys on the same survey line. Superimposing and imaging the superimposed profiles of multiple data on the same survey line can suppress random noise and increase the signal energy of the real target object.

[0037] 1. A detection method of a continuous dragging earthquake rapid survey system for dam road hidden dangers, comprising the following contents:

[0038] Content 1): Study the continuous drag seismic data acquisition method;

[0039] Content 2): Common receiver stacking method is used to extract information from real near-surface subsurface structures from vibration noise data;

[0040] Content 3): Superimposing multiple detection data suppresses interference noise and increases the energy of the real signal.

[0041] In the above content 1, the process of studying the continuous drag seismic data acquisition method includes the following:

[0042] The present invention proposes a continuous dragging seismic data acquisition method. The gravity acceleration geophone collects vibration and noise data simultaneously during the moving process. The continuous dragging moving data acquisition method can significantly improve the data acquisition efficiency. When the lateral resolution requirement is 0.5m, the moving speed can reach 5km / h. The continuous dragging moving data acquisition method has obvious advantages in the data acquisition efficiency of active source point measurement, multiple coverage method and passive source. The vibration generated by the friction between the base of the gravity acceleration geophone and the ground during continuous dragging is the seismic source. There is no need to configure an additional seismic source, which greatly simplifies the construction device of data acquisition and is more adaptable to the environment of rapid survey of dam hidden danger structures. The present invention aims to realize the rapid acquisition of seismic data of dam hidden dangers and solve the problem of rapid large-scale survey of shallow surface (≤20m) seismic methods.

[0043] In the above content 2, the common receiving point superposition method extracts information from the real near-surface underground structure from the vibration noise data, which mainly includes the following contents:

[0044] This invention proposes a common receiving point correlation stacking imaging method, which aims to extract the real target information from the shallow surface from the continuous drag noise data. If 24 / 48 gravity acceleration detectors move and collect data simultaneously, a multiple coverage observation system can be formed. Figure 1 As shown, the length of the array is K, the number of array moves is M, and the array pile number is 1-K+M-1. Assume that the recording time length of each array movement is t1, t2,......., t M-1 ,t M , the total recording time of multiple coverage is The total observation field value at position x is U(x,T)=[u(x,t1),u(x,t2),....,u(x,t M )], i = 1, 2, ....., M. u(x, ti) is the observed field value at position x during the time period ti, i = 1, 2, ....., M.

[0045] If the set window length is w, the number of iterations in the common receiving point superposition calculation is N = T / w, and the derived calculation formula is:

[0046]

[0047] Among them, l=1,2,...,N, s=w*l, (w+1)*l,..., (2w-1)*l, 2*w*l. ψ a(x, t) is the function of the observation record U(x, s) multiplied by U(x, s) at the position of x point, N times of iteration and normalization.

[0048] In the content 3, the data of multiple detections are superimposed to suppress interference noise and improve the energy of real signals, and the content mainly includes the following contents:

[0049] The continuous dragging data acquisition mode proposed in the content 1 can realize rapid acquisition of seismic data. Although the common receiving point correlation superposition method can extract information of target bodies from the near surface from the vibration noise data, the resolution of the superposition profile of single detection obtained is low. The low-resolution superposition profile is insufficient to distinguish small target bodies. The same line can be detected multiple times in a reciprocating manner to realize multiple coverages of the same target body.

[0050] Multiple superposition can suppress multiple waves and random interference noise. The random interference in the superposition profile obtained from the noise data of each detection is irrelevant, and after n times of superposition, the random interference is only enhanced If the effective signal is enhanced n times, the signal-to-noise ratio after superposition is increased times.

[0051] The application provides a detection method of a dam hidden danger continuous dragging seismic rapid survey system; comprising: (1) the application provides a continuous dragging seismic data acquisition mode. On the basis of the existing design of the gravity acceleration geophone, vibration noise data generated by the gravity acceleration geophone during movement is collected. The continuous dragging movement data acquisition mode can significantly improve the data acquisition efficiency, and when the lateral resolution requirement is 0.5 m, the speed of the movement type can reach 5 km / h, but the performance requirements of the seismic instrument hardware equipment are very high. When the seismic instrument is designed, long-time endurance, efficient data transmission and large-capacity data storage and the like need to be considered. The continuous dragging seismic data acquisition provides a necessary prerequisite for realizing dam hidden danger survey.

[0052] The vibration noise generated by the base of the gravity acceleration geophone and the ground friction is used as a seismic source, and the gravity acceleration geophone collects noise data, so that continuous dragging seismic data acquisition is realized, which belongs to the patent protection range.

[0053] (2) the application proposes a common receiving point correlation superposition imaging method, which extracts real target body information from the shallow surface from the continuous dragging noise data. The correlation superposition profiles of multiple detections are superimposed to suppress noise interference and improve the signal-to-noise ratio of real signals.

[0054] The continuous dragging seismic data common receiving point correlation superposition imaging method belongs to the patent protection range. The multiple superposition method belongs to mature technology and will not be repeated.

[0055] 1. Develop instruments and equipment for continuous towed seismic data acquisition

[0056] The continuous drag seismic detection system requires two key technical issues: gravity acceleration detector and main control acquisition equipment. Figure 2 The figure shows a developed gravity accelerometer. A flat base is installed at the bottom of the gravity accelerometer, making it easy to drag across the ground. The base's bottom contacts the ground, and the gravity accelerometer couples to the ground through gravity. This device has been widely used in conventional short-range precision detection and can be directly integrated into current detection systems. Specific requirements for the master acquisition device of this invention are:

[0057] (1) When the lateral resolution requirement is 0.5m, the traveling acquisition speed can reach 5km / h;

[0058] (2) The main control system is equipped with an RTK positioning system to accurately locate the position of the moving point;

[0059] (3) The main control system is stable, with functions such as long-term battery life of more than 8 hours, real-time data transmission and large-capacity data storage on the hard disk.

[0060] like Figure 3 Shown is the host system of the developed 48-channel continuous towed seismic instrument.

[0061] 2. Common-receiver correlation stack imaging of continuous drag seismic data:

[0062] In order to extract the reflection information of the real target body from the shallow surface from the continuous drag noise data, the present invention provides a noise data common receiving point correlation stacking imaging method. If 24 / 48 gravity acceleration detectors move and collect at the same time, a multiple coverage observation system can be formed. Figure 1 As shown, the length of the array is K, the number of array moves is M, and the array pile number is 1-K+M-1. Assume that the recording time length of each array movement is t1, t2,......., t M-1 ,t M , the total recording time of multiple coverage is The total observation field value at position x is U(x,T)=[u(x,t1),u(x,t2),....,u(x,t M )], i=1,2,.....,M.

[0063] If the window length is set to w, the number of iterations in the common receiving point superposition calculation is N = T / w, and the derived calculation formula is:

[0064]

[0065] where l = 1, 2,..., N, s = w*l, (w+1)*l,..., (2w-1)*l, 2*w*l. ψ a (x, t) is the function of the summation of U(x, s) multiplied by U(x, s) at x point, N iterations and normalization.

[0066] Model verification

[0067] As Figure 4 is the generated model of different diameter cavity and layered structure. From top to bottom, it can be divided into asphalt layer, backfill layer and gravel layer. The model size is 100m x 50m. The uniform grid size is 0.5m. The time iteration step is 4.0e -5 s, and the sampling length is 200ms. The numerical simulation is realized by using the finite difference elastic wave equation, and the horizontal free surface condition is added to simulate the real full wave field data. The perfectly matched layer is used as the boundary absorption condition. The source is a longitudinal wave source, which is directly loaded on the stress component. When continuous towed seismic data acquisition is carried out, the vibration generated by the gravity acceleration detector and the ground friction is the source. 24 gravity acceleration detectors are arranged in linear arrangement, and the whole arrangement moves when the arrangement moves. As Figure 5 shown in Fig. 2, 24 random noise sources with zero mean are used to simulate continuous towed seismic records as wavelets.

[0068] As Figure 6 shown in Fig. 3, the simulation records of model Figure 4 A, B and C at three different positions. The simulation records show continuous reflection axes Figure 5 , arrows). Considering the real reflection time of the wave impedance interface of the model and the actual problem of the three cavities, these reflection axes are not reflection waves from the interface and the cavity, but low-frequency strong energy noise. Because the source wavelet is random noise, the noise energy in the simulation record is strong. Affected by high-frequency noise, the single waveform is distorted and distorted.

[0069] As Figure 6 shown in Fig. 4, from left to right, the source position moves at an interval of 0.5m, and a total of 177 records are simulated. The correlation stacking algorithm is applied to realize stacking imaging, and the stacking profile as Figure 7 (a) is obtained. In each simulation, 24 noise sources are randomly generated. Repeat the simulation process, and use 24 random noise sources as wavelets to simulate continuous towed seismic records in turn, and obtain the stacking profiles Figure 7 , (b)-(h)). A total of 8 continuous towed seismic data acquisition processes are simulated, and 8 stacking profiles Figure 7 ) are obtained. The arrows A, B and C are the diffracted waves from the cavities (II, III and IV), and the energy is low, which is almost impossible to identify in some profiles Figure 7, (g) and (h)). Arrows D1, D2 and D3 are reflected waves from the interface (V), but their energy is very low and almost unrecognizable ( Figure 7 ). E and F are horizontal strong reflected waves, which belong to interference waves.

[0070] The single stacked profile has strong noise energy and low resolution, making it difficult to identify the reflection waves of the interface and the three cavities ( Figure 7 ). Overlay processing can suppress random noise and improve resolution. Figure 8 The figure shows the stacking results of 8 stacking sections calculated by conventional stacking method. Compared with the single stacking section ( Figure 7 ), the resolution of the multiple stacked sections is significantly improved, which can highlight the reflection waves of the top and bottom interfaces of the cavity body. A1 and A2 are the reflection waves of the top and bottom interfaces of cavity body II. B1 and B2 are the reflection waves of the top and bottom interfaces of cavity body III. C1 and C2 are the reflection waves of the top and bottom interfaces of cavity body IV. A3, B3 and C3 are multiple waves. Compared with the single stacked section ( Figure 7 ), Figure 8 The energy of the reflection axis from interface V is significantly enhanced (D1, D2 and D3).

[0071] Engineering Applications

[0072] (1) Data collection

[0073] The experimental area is located on an asphalt road at a school in southwest China. The surface layer is asphalt pavement, approximately 0.3m thick. The near-surface layer is backfill clay. The bottom layer is gravel. The interface between the backfill clay layer and the gravel layer is approximately 6m to 16m deep. There is a drainage pipe in the area, with a diameter of 0.5m and a buried depth of 2.0m. The designed survey line length is 70m ( Figure 9 (a)). The TS-24 continuous towed seismometer developed by ourselves is used to collect data. The distance between gravity acceleration detectors is 0.5m. Figure 9 (b) The seismic source is vibration caused by friction between the base of the gravity accelerometer and the ground. The arrangement was moved at 0.5m intervals. A seismic record was collected with each movement. The survey was repeated six times at the same location, resulting in 120 records per survey line. Each record had 24 channels, 8192 sample points per channel, and a sampling interval of 41.7μs. Figure 9 (c) is the original waveform collected on-site. To verify the accuracy of the continuous towed seismic survey, single-shot data were collected along the same survey line. The offset was 3 meters, and the shot spacing was 1 meter. The source was a 24-pound sledgehammer.

[0074] (2) Conventional overlay processing

[0075] like Figure 10(a)-(b) are single-shot records from different positions. A is the direct wave, B is the surface wave. C and D are high-frequency noise. As Figure 10 (c)-(d) are continuous towed seismic records from different positions. Because of the different surface smoothness, the force on the gravity accelerometer is different when it rubs against the surface in some local convex positions, causing distortion of the seismic record. The main signals in the distorted signals are ultra-low frequency (E) and ultra-high frequency (F1 and F2). The spectral analysis results show that Figure 10 (a)), the spectral width of the signal is 1 Hz to 2600 Hz, and the energy is stronger in the 1 Hz to 900 Hz segment, while the energy is lower in the 901 Hz to 2600 Hz segment. The spectral range of the continuous towed seismic record and the active source single-shot record is basically the same, and the signal in the 901 Hz to 2600 Hz segment is mainly noise, and the effective frequency band is in the 1 Hz to 900 Hz segment. The surface wave spectrum of the active source single-shot data shows that the dispersion characteristics of the surface wave are good, and continuous dispersion curves can be picked up, but there is a break in the 50 Hz to 100 Hz segment of the spectrum.

[0076] For the original records shown in Figure 11 (c) and (d), a series of methods are used for data processing, including band-pass filtering, correlation stacking, curvelet denoising, predictive deconvolution, automatic gain control, trace equalization, convolutional neural network, etc. The random noise of single-probe data is one of the factors affecting the resolution of the final stacked profile. The curvelet transform method can realize multi-angle and multi-scale filtering, which is very helpful for removing random noise, so the curvelet transform method is used as part of the data preprocessing in this paper.

[0077] The data processing results of the 6 probes after trace equalization are shown in Figure 12 . There are slight changes in the reflected waves at time 5 ms and 12 ms, and the lateral distance is in the 19 m-22 m region Figure 12 . The changes are most obvious in Figure 12 (a). The continuous reflected waves at time about 5 ms (B) and 12 ms (C) have obvious similarities in the 6 profiles, but the energy difference of the reflected wave C is large, and the low-energy region is not easy to identify. In the time 20 ms to 30 ms segment, there is only a continuous reflection axis (D) in Figure 12 (d). The near-surface layer of the experimental area is backfill clay, and the average P-wave velocity is estimated to be 800 m / s for time-depth conversion. Although the 6 profiles are obtained with the same parameters, because the source signal is different each time, the amplitude of the continuous reflected wave still has obvious differences, and the discontinuous reflected wave in the time 20 ms to 50 ms segment is different in the 6 profiles.

[0078] In order to verify the reliability of the continuous towed results, the data shown in Figure 10(a)-(b) are single shot records, respectively obtained by inverse scattering imaging, surface wave imaging and stack imaging methods as shown in Figure 13 (a)-(c) are corresponding profiles. The inverse scattering profile Figure 13 (a) shows strong energy anomaly bodies in lateral positions 16m to 21m. In the lateral corresponding positions, the surface wave apparent velocity profile Figure 13 (b) and the stack profile Figure 13 (c) have obvious lateral changes of velocity and reflection events. In combination with the surface wave apparent velocity profile and the stack profile, four continuous interfaces (B-E) can be identified. The depth change positions of the interfaces are 1.5m (B), 4.0m (C), 6m to 10m (D) and 8m to 14m (E). In combination with the geological data in the experimental area, it is considered that the reflection axis B is the interface between the asphalt layer and the backfill clay layer, C and D are the internal reflections of the backfill clay layer, and E is the interface between the backfill clay layer and the sand gravel layer. For the data processing results as shown in Fig. 15, the stack profile is calculated by using the conventional stack method, and the stack results of 6 times of data are shown in Figure 13 (d). Multiple stacking can highlight the true reflection interface, and the continuity of the reflection wave with a time of about 12ms is enhanced Figure 13 (d), C). In the time period of 20ms to 50ms, there are chaotic reflection waves, and it is almost impossible to identify continuous reflection events Figure 13 (d). The comparison results prove that the stack imaging results of the multiple continuous towed seismic data are close to the stack imaging results of the conventional single shot data when the time is less than 20ms, and the resolution of the stack profile of the multiple continuous towed seismic data is low and almost covered by noise when the time is in the period of 20ms to 50ms.

[0079] Finally, it should be noted that the above-mentioned is only a specific embodiment of the present application, of course, the person skilled in the art can modify and change the present application, provided that these modifications and changes are within the scope of the claims of the present application and its equivalent technology, all should be considered as the protection scope of the present application.

Claims

1. A method for rapid survey and detection of dam road hidden dangers by continuous dragging earthquake, characterized in that: The steps include: Step A: Using a gravity accelerometer to simultaneously collect vibration and noise data while traveling on a hardened road surface, the vibration generated by the friction between the base of the gravity accelerometer and the ground during continuous dragging is used as the seismic source, and the vibration signal generated by the friction between the base of the gravity accelerometer and the hardened road surface during the movement of the gravity accelerometer is used as the noise source. The gravity accelerometer collects seismic signals during the dragging process; Step B: Using the common receiver stacking method to extract information from the real near-surface underground structure from the vibration noise data; Step C: Superimpose the data of multiple detections to suppress interference noise and increase the energy of the real signal.

2. The method for rapid dam-road hidden danger continuous dragging seismic survey and detection according to claim 1 is characterized by: The step B uses a common receiving point superposition method to extract information from the actual near-surface underground structure from the vibration noise data, including the following contents: D detectors are moved and collected simultaneously to form a multiple coverage observation system. The arrangement length is K, the number of arrangement movements is M, and the arrangement stake number is 1-K+M-1. Assume that the recording time length of each movement collection is t1, t2,......., t M- 1,t M , the total recording time of multiple coverage is The total observation field value at position x is U(x,T)=[u(x,t1),u(x,t2),....,u(x,t M )], i = 1, 2, ....., M; u(x, ti) is the observed field value at position x during the time period ti; If the window length is set to w, the number of iterations in the common receiving point superposition calculation is N = Tw, and the derived calculation formula is: Among them, l=1,2,...,N, s=w*l,(w+1)*l,...,(2w-1)*l,2*w*l; ψ a (x, t) is the function obtained by multiplying the observation record U(x, s) at the point x by U(x, s), summing the results, and performing N iterations and normalization.

3. The method for rapid dam-road hidden danger continuous dragging seismic survey and detection according to claim 2 is characterized by: The D is 24 or 48.

Images