A system and method for monitoring seepage channels of reservoir rockfill dams using microseismic signals

Abstract

The present invention belongs to the technical field of reservoir safety detection, and provides a system and method for monitoring the seepage channel of a reservoir rockfill dam using microseismic signals. The system includes a fixing device, a signal acquisition device, and a cloud data analysis system; the fixing device is used to install the signal acquisition device on the dam body; the signal acquisition device is used to collect microseismic signals of rockfill material rupture, and send the microseismic signals to the cloud data analysis system; the cloud data analysis system is used to obtain whether the rockfill dam body is ruptured based on the received microseismic signals and the information vector machine model for identifying the type of rockfill dam rupture; based on the received microseismic signals and the spatial distribution map of the source position, the distribution of the source position in the spatial distribution map of the source position is analyzed to determine whether there is a potential seepage channel in the local area. The present invention can track the rupture of rockfill materials and its evolution process caused by seepage inside the rockfill dam body of the upper reservoir of the pumped storage power station.

Description

Technical Field

[0001] The invention belongs to the technical field of reservoir safety detection, and in particular relates to a system and method for monitoring a seepage channel of a reservoir rockfill dam by using microseismic signals. Background Art

[0002] The statements in this section merely provide background information related to the present invention and do not necessarily constitute prior art.

[0003] Dam body seepage monitoring often uses piezometers, which are generally installed on a few cross sections of the dam body. Considering cost factors, the number of piezometers installed is very limited. Therefore, piezometer monitoring can generally only obtain the seepage pressure at a small number of discrete points inside the dam body, and cannot directly monitor the seepage pressure and its changing process in all areas of the dam body. Therefore, it is difficult to accurately monitor or identify three-dimensional seepage channels with complex development paths that may appear inside the dam body. In addition, the signal transmission of conventional piezometers adopts wired transmission, and the cost of laying and managing transmission lines is often high. Summary of the invention

[0004] In order to solve the above problems, the present invention proposes a system and method for monitoring the seepage channel of a reservoir rockfill dam using microseismic signals. The present invention is a non-destructive and efficient monitoring method for a rockfill dam body, which can identify and track in real time the rupture of rockfill materials and their evolution caused by seepage inside the rockfill dam body of a pumped-storage power station reservoir, thereby effectively monitoring the development process of three-dimensional seepage channels inside the rockfill dam body of the pumped-storage power station reservoir.

[0005] According to some embodiments, the present invention adopts the following technical solutions:

[0006] In a first aspect, the present invention provides a system for monitoring seepage channels of reservoir rockfill dams using microseismic signals.

[0007] A system for monitoring seepage channels of reservoir rockfill dams using microseismic signals, comprising: a fixing device, a signal acquisition device, and a cloud data analysis system;

[0008] The fixing device comprises a fixing plate, a fixing rod and screws, the fixing plate is connected to the signal acquisition device through the screws, through holes are provided at both ends of the fixing plate, and the fixing rod passes through the through holes, so that the signal acquisition device is installed on the dam body;

[0009] The signal acquisition device includes a microseismic signal sensor for acquiring microseismic signals of rockfill material fracture and sending the microseismic signals to a cloud data analysis system;

[0010] The cloud data analysis system is used to determine whether the rockfill dam body is broken based on the received microseismic signals and the information vector machine model for identifying the type of rockfill dam body rupture; or to analyze the distribution of the source position in the spatial distribution map of the source position based on the received microseismic signals and the spatial distribution map of the source position to determine whether there is a potential seepage channel in the local area.

[0011] Furthermore, the fixing device also includes an expansion screw. The fixing rod is a multi-diameter hollow rod. A hole smaller than the diameter of the fixing rod is pre-drilled in the surface concrete panel, and the fixing rod passes through the fixing plate starting from the top direction, and the expansion screw is inserted into the hole. The fixing nut is screwed into the outer plate surface of the fixing plate at the top end of the fixing rod, so that the expansion screw inside the hole begins to expand and forms a firm whole with the dam body around the hole wall, so that the signal acquisition device can be installed on the dam body.

[0012] Furthermore, the fixing device also includes a signal transmission protection tube, and the microseismic signal sensor is arranged inside the signal transmission protection tube.

[0013] Furthermore, the signal acquisition device also includes an FGPA microcontroller and a power supply. The FGPA microcontroller is connected to the microseismic signal sensor through a signal transmission line. The FGPA microcontroller is connected to the power supply, high-speed AD converter, storage, clock and communication components. The power supply supplies power to the microseismic signal sensor, FGPA microcontroller, high-speed AD converter, storage, clock and communication components.

[0014] Furthermore, the cloud data analysis system includes a central processing unit, which is connected to the storage, multimedia components, I / O interfaces and communication components.

[0015] Furthermore, the analyzing the distribution of the earthquake source positions in the earthquake source position spatial distribution map to determine whether there is a potential seepage channel in the local area specifically includes:

[0016] When the hypocenter locations of seepage microseismic events are distributed and aggregated in a strip-like or surface-like manner in one or some local areas of the monitoring area, there are potential seepage channels in the corresponding local areas; if the hypocenter locations of microseismic events are discretely distributed in one or some local areas of the monitoring area without aggregation, there are no potential seepage channels in the corresponding local areas.

[0017] Furthermore, the information vector machine model for identifying the type of rockfill dam rupture includes an IVM binary classification model.

[0018] In a second aspect, the present invention provides a method for monitoring the seepage channel of a reservoir rockfill dam using microseismic signals.

[0019] A method for monitoring a seepage channel of a reservoir rockfill dam using microseismic signals, characterized in that the system for monitoring a seepage channel of a reservoir rockfill dam using microseismic signals according to the first aspect is adopted, comprising:

[0020] Obtain microseismic signals of rockfill material rupture;

[0021] Decomposing and reconstructing the microseismic signal by wavelet transform to obtain a feature vector;

[0022] Based on the feature vector, using the trained information vector machine model for identifying the type of rockfill dam rupture, it is determined whether the rockfill dam is ruptured;

[0023] The process of training the information vector machine model for identifying the rupture type of the rockfill dam body comprises: obtaining microseismic signals of the rupture of the rockfill body material under the action of non-seepage and seepage; using wavelet transform to decompose and reconstruct the microseismic signals of the rupture of the rockfill body material under the action of non-seepage and seepage, and constructing a training set; using the training set to train the information vector machine model for identifying the rupture type of the rockfill dam body until the accuracy of the prediction result reaches a set threshold, and a trained information vector machine model for identifying the rupture type of the rockfill dam body is obtained.

[0024] In a third aspect, the present invention provides a method for monitoring the seepage channel of a reservoir rockfill dam using microseismic signals.

[0025] A method for monitoring a seepage channel of a reservoir rockfill dam using microseismic signals, using the system for monitoring a seepage channel of a reservoir rockfill dam using microseismic signals as described in the first aspect, comprising:

[0026] Obtain microseismic signals and spatial distribution maps of earthquake source locations;

[0027] Based on the microseismic signals and the spatial distribution map of the earthquake source positions, the distribution of the earthquake source positions in the spatial distribution map of the earthquake source positions is analyzed to determine whether there are potential seepage channels in the local area.

[0028] Furthermore, the acquisition of the spatial distribution map of the earthquake source position includes:

[0029] Assume that the coordinates of the epicenter of the microseismic event are (x k ,y k , z k ), the time when the microseism occurs is t k , define t ki is the starting time of the elastic wave generated by the microseismic event received by the i-th sensor. According to the distance between the source of the microseismic event and each sensor, and the relationship between speed and time, the following six equations are listed according to the two-point distance formula:

[0030] (x k -x 1 )2 +(y k -y 1 ) 2 +(z k -z 1 ) 2 =v k 2 (t k1 -t k ) 2

[0031] (x k -x 2 ) 2 +(y k -y 2 ) 2 +(z k -z 2 ) 2 =v k 2 (t k2 -t k ) 2

[0032] (x k -x 3 ) 2 +(y k -y 3 ) 2 +(z k -z 3 ) 2 =v k 2 (t k3 -t k ) 2

[0033] (x k -x 4 ) 2 +(y k -y 4 ) 2 +(z k -z 4 ) 2 =v k 2 (t k4 -t k ) 2

[0034] (x k -x 5 ) 2 +(y k -y 5 ) 2 +(z k-z 5 ) 2 =v k 2 (t k5 -t k ) 2

[0035] (x k -x 6 ) 2 +(y k -y 6 ) 2 +(z k -z 6 ) 2 =v k 2 (t k6 -t k ) 2

[0036] By combining the above six equations and substituting the average equivalent wave velocity v of the rock mass, the coordinates of each sensor, and the value of the starting time when each sensor receives the elastic wave generated by the microseismic event, the coordinates of the epicenter of the microseismic event (x k ,y k , z k ) and the time t at which the microseism occurs k ;

[0037] According to the coordinates of the epicenter of the microseismic event (x k ,y k , z k ) and the time t at which the microseism occurs k , draw a spatial distribution map of the earthquake source location.

[0038] Compared with the prior art, the present invention has the following beneficial effects:

[0039] 1. The present invention provides a system for monitoring the seepage channel of a reservoir rockfill dam using microseismic signals, which includes a fixing device, a signal acquisition device, and a cloud data analysis system. The fixing device uses a nut to fix and remove the fixing plate, solving the installation and recovery problems of the device; a highly sensitive microseismic sensor is used, and the installation process is simple and convenient, which is suitable for the daily regulation of the operation mode of the upper reservoir of the pumped storage power station; by integrating the fixing device, microseismic sensor, wireless transmitter, and power supply, the wired transmission of the microseismic signal is transformed into wireless transmission, making the arrangement position of the sensor more flexible and the system functional completeness good; the knob-type signal transmission protection cylinder has friction teeth added on the inside, which increases the load-bearing capacity of the device and enhances the coupling effect between the microseismic equipment and the hole wall of the fixing device; the cloud data analysis system can perform cloud storage, cloud processing, cloud analysis, and cloud early warning of data, and can monitor the potential seepage channel of the rockfill dam body in real time. Compared with the traditional automatic seepage monitoring system, the system has stronger perception ability and higher automation and intelligence levels.

[0040] 2. The present invention provides a method for monitoring the seepage channel of a reservoir rockfill dam using microseismic signals. The method is a non-destructive and efficient monitoring method for a rockfill dam body, and can identify and track in real time the rupture of rockfill materials and their evolution caused by seepage inside the rockfill dam body of a pumped-storage power station reservoir, thereby effectively monitoring the development process of three-dimensional seepage channels inside the rockfill dam body of the pumped-storage power station reservoir. The method solves the technical bottleneck problem that it is difficult to accurately obtain the three-dimensional seepage channels inside the dam using existing traditional methods such as dam seepage pressure monitoring methods and dam seepage pressure and water level observations, and provides a practical technical means for effectively monitoring the seepage channels inside the dam body. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] The accompanying drawings in the specification, which constitute a part of the present invention, are used to provide a further understanding of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute improper limitations on the present invention.

[0042] Figure 1 It is a structural schematic diagram of a system for monitoring the seepage channel of a reservoir rockfill dam using microseismic signals shown in the present invention;

[0043] Figure 2 It is a flow chart of a method for monitoring the seepage channel of a reservoir rockfill dam using microseismic signals according to the present invention;

[0044] FIG3( a ) is a waveform diagram of a seepage rupture microseismic signal shown in the present invention;

[0045] FIG3( b ) is a waveform diagram of a non-seepage fracture microseismic signal shown in the present invention;

[0046] Figure 4This is a rendering of the present invention showing the classification of seepage rupture and non-seepage rupture of rockfill dam body through binary classification of IVM model;

[0047] Figure 5 is a schematic diagram of the spatial distribution of microseismic location events shown in the present invention;

[0048] Figure 6 is a schematic diagram of a signal acquisition device shown in the present invention;

[0049] Figure 7 is a schematic diagram of a cloud data analysis system shown in the present invention;

[0050] FIG8( a ) is a fifth microseismic event-seepage rupture effect diagram shown in the present invention;

[0051] FIG8( b ) is a diagram showing the effect of the 107th microseismic event-seepage rupture shown in the present invention;

[0052] FIG8( c ) is a diagram showing the 103rd microseismic event-non-seepage rupture effect of the present invention;

[0053] FIG8( d ) is a diagram showing the 220th microseismic event-non-seepage rupture effect of the present invention;

[0054] Among them, A, fixing device, A1, fixing rod, A2, expansion screw, A3, fixing plate, A4, fixing nut, A5, screw, A6, signal transmission protection tube, B, signal acquisition device, B1, microseismic signal sensor, B2, power supply, B3, high-speed AD converter, B4, storage, B5, clock, B6, FGPA microcontroller, B7, communication component, B8, signal transmission line, C, cloud data analysis system, C1, storage, C2, multimedia component, C3, I / O interface, C4, communication component, C5, central processing unit. DETAILED DESCRIPTION

[0055] The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0056] It should be noted that the following detailed descriptions are exemplary and are intended to provide further explanation of the present invention. Unless otherwise specified, all technical and scientific terms used herein have the same meanings as those commonly understood by those skilled in the art to which the present invention belongs.

[0057] It should be noted that the terms used herein are only for describing specific embodiments and are not intended to limit exemplary embodiments according to the present invention. As used herein, unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. In addition, it should be understood that when the terms "comprising" and / or "including" are used in this specification, it indicates the presence of features, steps, operations, devices, components and / or combinations thereof.

[0058] In the present invention, the orientation or position relationship indicated by terms such as "end", "top", "side", "bottom", etc. is based on the orientation or position relationship shown in the drawings. They are relational words determined only for the convenience of describing the structural relationship of the various parts or elements of the present invention. They do not specifically refer to any part or element in the present invention and cannot be understood as a limitation on the present invention.

[0059] In the present invention, terms such as "connected" and "connection" should be understood in a broad sense, indicating that the connection can be fixed, integral or detachable; it can be directly connected or indirectly connected through an intermediate medium. Relevant scientific research or technical personnel in this field can determine the specific meaning of the above terms in the present invention according to specific circumstances, and they should not be understood as limiting the present invention.

[0060] Microseismic signals are tiny vibration signals generated by the fracture of brittle materials. They are effective physical information that can reasonably characterize the fracture process of an object. Microseismic monitoring technology is an important means of monitoring the stability of surrounding rock in deep engineering projects. It can conduct real-time dynamic monitoring of the entire process of rock mass development from microfracture to instability and destruction. The present invention introduces microseismic monitoring technology into the monitoring of seepage channels of rockfill dams, and proposes a system and method for monitoring the seepage channels of reservoir rockfill dams using microseismic signals, providing an efficient and reliable technical means for the safety monitoring of pumped storage power stations.

[0061] Embodiment 1

[0062] This embodiment provides a system for monitoring seepage channels of a reservoir rockfill dam using microseismic signals.

[0063] Attached Figure 1 This is a schematic diagram of the structure of a system for monitoring the seepage channel of a reservoir rockfill dam using microseismic signals in Embodiment 1 of the present invention. The system includes: a fixing device A, a signal acquisition device B, and a cloud data analysis system C.

[0064] The fixing device A is used to install the signal collection device B on the dam body, and includes a fixing rod A1, an expansion screw A2, a fixing plate A3, a fixing nut A4, a screw A5 and a signal transmission protection tube A6.

[0065] The fixing rod A1 is a multi-diameter hollow rod, and a hole slightly smaller than the diameter of the fixing rod A1 is pre-drilled on the surface concrete panel, and the fixing rod A1 passes through the fixing plate A3 starting from the top direction, and the expansion screw A2 with the key fixing pivot function of the overall component is inserted into the preset hole, and then the fixing nut A4 is screwed into the outer plate surface of the fixing plate A3 at the top of the fixing rod A1, so that the expansion screw A2 inside the hole begins to expand and forms a firm whole with the dam body around the hole wall, thereby preliminarily constructing a reliable connection body consisting of the rockfill dam panel and the fixing device.

[0066] The signal acquisition device B is then preliminarily fixed at the longitudinal axis hole of the fixing plate A3, and the screw A5 is screwed in from the bottom end of the fixing plate A3 to stably install the signal acquisition device B, thereby constructing a good vibration system composed of the rockfill dam panel, the fixing device A and the signal acquisition device B, which can ensure the effective transmission between the microseismic signal generated inside the rockfill dam body and the microseismic sensor B1, so that the front-end equipment of the system for monitoring the seepage channel of the rockfill dam in the upper reservoir of the pumped-storage power station using microseismic signals can obtain effective microseismic signal data.

[0067] As one or more implementation modes, the signal acquisition device B can be fixed as a whole on the outside / top of the fixing plate A3 by utilizing the prefabricated inner knob thread structure of the signal transmission protection tube A6 and the prefabricated knob tooth structure of the fixing plate A3. This structure further improves the conduction quality of the microseismic signal on the fixing device; and the outer side of the signal transmission protection tube A6 is coated with an insulating and waterproof protective layer, which can greatly improve the durability and signal acquisition stability of the microseismic signal sensor B1 in the sensor acquisition device B. The signal transmission line B8 on the top of the microseismic signal sensor B1 is led out along the top of the sensor protection tube A6 and connected to the signal acquisition device B.

[0068] As one or more embodiments, the fixing rod A1 is a multi-diameter hollow rod, and its aperture type can be selected according to the suitability of on-site drilling; the diameter of the hollow rod is larger than the aperture of the expansion screw A2, and the thickness of the through hole on the fixing plate A4 is slightly smaller than the thickness of the solid part of the expansion screw A2.

[0069] As one or more implementation modes, the signal transmission tube A6 includes a low-noise cable, including a skin, a conductive layer, an insulation body, a core wire and a waterproof material, which can effectively prevent the noise phenomenon caused by the change of capacitance due to bending through the core wire; optionally, the signal transmission tube A6 material can be PVC, PU, ​​Si or FPM.

[0070] As one or more implementation modes, the wiring method in the signal transmission tube A6 is as follows: first, in an axial wiring segment, the signal transmission tube A6 is wired with the axis of the microseismic signal sensor B1 and the signal acquisition device A as a reference, while ensuring that the signal transmission tube A6 in this wiring segment does not bend; the signal transmission tube A6 is introduced along the through-hole direction to the external signal input end of the signal analysis device B.

[0071] The signal acquisition device B is used to collect microseismic signals generated by hydraulic fracturing in the dam body in real time, and transmit them to the cloud data analysis system C through 5G wireless. It includes a microseismic signal sensor B1, a power supply B2, a high-speed AD converter B3, a storage B4, a clock B5, a FGPA microcontroller B6, a communication component B7 and a signal transmission line B8.

[0072] As one or more embodiments, the microseismic signal sensor B1 is a high-sensitivity piezoelectric microseismic signal sensor. It includes a housing, a piezoelectric element, and a mass block structure. By collecting the inertial force acting on the piezoelectric element and the microseismic signal sensor, the corresponding voltage value can be calculated through Newton's second law; the microseismic signal sensor B1 also includes a built-in preamplifier, which can convert the sensor's signal into a low-impedance voltage signal, and no longer requires an external charge amplifier, which reduces the space occupied by the signal acquisition device A component and the overall space of the dam monitoring device is flexibly allocated to ensure its accurate operation; optionally, according to the sensor structure and feature classification, the microseismic signal sensor B1 can be a compression type, shear type, or flexure type sensor, and each sensor has certain differences in the measurable acceleration size, measurement sensitivity, and resonant frequency, and the selection should be based on the specific situation.

[0073] As one or more implementation modes, the connection mode between the microseismic signal sensor B1 and the signal transmission tube A6 can be BNC, TNC, Miniature or C29-104P type. If the interface mode of the microseismic signal sensor B1 and the signal transmission tube A6 does not match, the two can also be effectively connected through a terminal converter.

[0074] The cloud data analysis system C is used to receive the microseismic signal data transmitted by the signal acquisition device B, and includes a storage C1, a multimedia component C2, an I / O interface C3, a communication component C4 and a central processing unit C5.

[0075] As one or more implementation modes, the storage C1 may include one or more computer program products, and the computer program products may include various forms of computer-readable storage media, such as volatile memory and / or non-volatile memory; the computer-readable storage medium stores the above-mentioned AD conversion, signal conditioning, IVM learning model and a series of computer programs, which can be called and executed by the processor C1.

[0076] As one or more embodiments, the multimedia component C2 can be used to receive instructions input by the user, and can also output various information (such as text data, images or sounds) to the outside (such as the user), and can include one or more of buttons, displays, speakers, etc.; the display can display the monitored potential seepage channels.

[0077] As one or more implementation modes, the communication component C4 may be used to provide the signal analysis device B with a signal that other components are operating normally.

[0078] As one or more implementations, the processor C5 may be a central processing unit (CPU) or other forms of processing units having data processing capabilities and / or instruction execution capabilities, and may control other components of the signal analysis device B to perform desired functions.

[0079] Embodiment 2

[0080] This embodiment provides a method for monitoring the seepage channel of a reservoir rockfill dam using microseismic signals.

[0081] This embodiment uses conventional fracture mechanics tests and seepage failure mechanics tests of rockfill rock samples as data samples. After screening, integrating, and eliminating erroneous data, and then filtering and denoising the data, 50 data that meet the requirements of model construction are obtained. See the flow chart for details. Figure 2 Since the microseismic signal waveform data of each test case basically contains the whole process of rockfill crack incubation, in order to make full use of the data, this example selects the data and decomposes 50 large sample data into 546 training samples with representative and evolution characteristics.

[0082] The specific steps of this embodiment are as follows:

[0083] Step S2-1: Through conventional mechanical tests and seepage failure tests on rockfill rock samples, microseismic signals of rockfill material fracture under non-seepage and seepage effects are obtained respectively.

[0084] Through uniaxial compression failure test, direct shear failure test and seepage failure test of rockfill, seepage type and non-seepage type microseismic signals are collected respectively and used as training samples of the recognition model, such as Figure 3(a)-Figure 3(b) shown.

[0085] Step S2-2: Decompose and reconstruct the signal through wavelet transform, and use it as the feature vector of the sample in the recognition model.

[0086] The wavelet analysis method is used to perform five-layer decomposition and reconstruction of the microseismic signals of the seepage type and non-seepage type, and the sample entropy of the fifth-layer approximate signal is calculated, which is used together with the amplitude of the microseismic signal as the characteristic index of the sample in the identification model. The wavelet transform method can refine the signal at multiple scales through the telescopic and translation operations. Compared with the Fourier transform, the wavelet transform can realize the localized decomposition of time and frequency.

[0087] Wavelet transform can realize the time subdivision of the signal at high frequencies and the frequency subdivision at low frequencies, which has unique advantages over Fourier transform. For any function f(t)∈L2(R), the discrete wavelet transform calculation formula is as follows, where a is the scaling factor, b is the translation factor, and ψ(t)∈L2(R) represents a wavelet basis for signal change.

[0088]

[0089] Step S2-3: Using the training samples, an information vector machine model for identifying the rupture type of the rockfill dam is established.

[0090] The Information Vector Machine (IVM) maps the feature vectors of training samples to a high-dimensional feature space and searches for an optimal classification hyperplane in this space to maximize the segmentation interval between the two types of samples.

[0091] This embodiment uses the binary classification of the IVM model to achieve the classification of seepage rupture and non-seepage rupture of the rockfill dam body. Figure 4 .

[0092] Preferably, the IVM two-classification model is based on Bayesian statistical learning theory and kernel method, combined with the assumed density approximation ADF and the sample selection method based on information entropy theory, and uses ADP recursively approximates the posterior distribution after each additional sample (information vector), and obtains the approximate likelihood distribution to ensure the traceability and tractability of the algorithm. At the same time, in the recursive approximation process, the sample selection method based on information entropy theory is used to screen the most informative samples in order to replace the learning of the original training sample set, and the method of recursively updating and recording intermediate variables is used to cleverly avoid the repeated processing of the covariance matrix K (kernel matrix), so as to achieve the purpose of significantly reducing the time and space complexity of the algorithm. The following is a brief description of some key steps in the learning and prediction process of the IVM classification model.

[0093] The meanings of some symbols in the formulas in this section are as follows: p represents probability distribution, q represents approximate probability distribution, N represents Gaussian distribution, X and y represent the input feature vector set (matrix) and output target vector of the training sample, respectively. * and * Respectively represent the input feature vector and output target (unknown to be predicted) of the sample to be predicted, μ * and σ * They represent the predicted mean and predicted variance respectively, f represents the latent variable set, m represents the likelihood substitution variable, B or β represents the noise distribution variance, μ represents the Gaussian distribution mean, K or ∑ represents the Gaussian distribution covariance matrix, θ represents the covariance function hyperparameter vector, and the explanation of other symbols can be obtained in other places in this embodiment.

[0094] During the learning process, the IVM binary classification model maintains two sample index sets I and J, where I is the valid set and J is the set to be selected. Initially, I = Φ, J = {1, 2, ..., n}, and at any time, I∩J = Φ, I∪J = {1, 2, ..., N} (assuming that d information vectors are selected from N training samples). The information vector is obtained in a continuous, online learning-like manner: First, the ADF is applied to approximate the posterior and likelihood distributions when there are i information vectors, i.e., Ii (for the case of Gaussian distribution, the approximate solution is consistent with the exact solution):

[0095]

[0096] After that, the i+1th information vector is selected as follows

[0097]

[0098] The above formula means: select a sample index j in the current candidate set J that can maximize the reduction of the posterior distribution information entropy, and use the j sample as the i+1th information vector.

[0099] Generally, for a probability distribution, information entropy H represents its degree of uncertainty, and statistical learning naturally hopes to obtain a certain (predicted) posterior distribution as much as possible. The above process is repeated until d information vectors are selected (I = Id). At this point, we can get:

[0100]

[0101] In the IVM binary classification model, the optimal solution of the covariance function hyperparameter θ is adaptively obtained by maximizing the marginal likelihood p(yI|XI,:,θ). Specifically, by taking the negative logarithm -log(p(yI|XI,:,θ)), the maximization problem is transformed into a minimization problem, and then the conjugate gradient descent method is used to achieve the optimal hyperparameter Adaptive acquisition.

[0102] The above learning process realizes the replacement of the learning of the original sample set with the learning of the samples (information vectors) specified by the effective set I. The subsequent prediction process is consistent with the Bayesian classification (when the prior and noise satisfy the Gaussian distribution, it is also called information vector machine classification). Substituting the corresponding variables, the IVM prediction posterior distribution is obtained:

[0103]

[0104] For the binary classification problem involved in this embodiment, after obtaining the above-mentioned IVM prediction posterior distribution, the prediction of the category to which the unknown sample belongs can be obtained by simply performing the following transformation.

[0105]

[0106] Where Φ(x) represents the cumulative probability density function of the standard normal distribution, which can of course be replaced by other response functions, such as the sigmoid logistic function.

[0107] In addition, in the IVM binary classification model, the covariance matrix k, also known as the kernel matrix, has the same idea as the kernel function: if the basis function set φ(x) is used to map the original sample data to the high-order feature space, the dot product of the original sample input vector is generalized to the dot product of the basis function through the kernel, that is:

[0108] K i,j ≡k(x i ,x j )=cov(f(x i ),f(x j ))=φ(x i ) T ∑φ (7)

[0109] It is worth noting that the above mapping relationship is nonlinear, and the dimension of the feature space may be extremely high or even infinite. There may be infinite φ(x) in the basis function. Therefore, the introduction of the covariance function (kernel method) enables IVM to have powerful nonlinear processing capabilities.

[0110] Step S2-4, performance verification of the information vector machine model for identifying the rupture type of the rockfill dam body.

[0111] This embodiment performs a feasibility test on the results of the test samples output by the optimal information vector machine classification model. Specifically, the test index is the prediction accuracy of the verification sample, that is, the actual possibility of rockfill body rupture of the verification sample is checked with the predicted possibility of rockfill block rupture. If the prediction accuracy is above 95%, it is considered that the performance of the established optimal information vector machine classification model meets the requirements. Otherwise, the information vector machine model is retrained after adding training samples.

[0112] Embodiment 3

[0113] This embodiment takes the rockfill dam of a large pumped storage power station as an example to specifically illustrate the potential seepage channel identification method of the rockfill dam of the upper reservoir of the pumped storage power station based on microseismic monitoring. Figure 5 .

[0114] In this embodiment, the 200m×200m range of the western dam surface of the rockfill dam is used as the monitoring area, and sensors are installed on the dam surface of the monitoring area, providing 6 sensors, and the sensor numbers are S1, S2, S3, S4, S5, and S6. The microseismic monitoring system mainly includes a microseismic signal sensor, a digital signal acquisition system (i.e., an acquisition instrument), and a digital signal processing system. Each microseismic signal sensor is connected to the digital signal acquisition system through a cable, and the digital signal acquisition system is connected to the digital signal processing system through a network cable. The sensitivity of the sensor is 30V / g, the frequency response range is 50Hz~5kHz, the sampling frequency of the digital signal acquisition system is 20kHz, the sensor converts the received stress wave into an electrical signal, and stores it in the digital signal processing system after converting it into a digital signal through the digital signal acquisition system. In this embodiment, the starting time of the elastic wave collected by the sensor is the starting time of the P wave.

[0115] Step S3-1: Arrange and fix the microseismic sensor B1 (see Figure 6 ). The microseismic sensor B1 is paired with the fixing plate A3 through screws A5 and installed, with at least six sets; the top output end of the microseismic sensor B1 is connected to the signal transmission protection tube A6, the end of the signal transmission line B8 passes through the upper end of the signal transmission protection tube A6, and the signal transmission protection tube A6 is rotated clockwise to fix the torsion structure with the outside of the fixing plate A3;

[0116] Step S3-2: Drill fixed holes on the upstream panel and downstream horseway of the dam body. The diameter of the fixed hole should be slightly smaller than the fixing rod A1. The elevations of the fixed holes are different and form a spatial mesh structure distribution. The arrangement of the fixed holes avoids any three being located on the same straight line and any four being located on the same plane. The spatial structure where the fixing rod is located covers the upstream concrete panel and its side wall of the dam as much as possible. The length of the fixed hole should be greater than 80mm. The specific drilling depth is determined according to the specifications and types of the dam concrete.

[0117] Step S3-3: Fix the microseismic sensor B1 at the center of the fixing hole. Insert the expansion screw A2, the fixing plate A3 and the device in step S3-1, and twist the fixing screw A5 to fix it as a whole, so that the microseismic sensor B1 is tightly connected to the fixing plate A3, and the bottom end of the fixing rod A1 is in close contact with the concrete surface of the dam body.

[0118] Specifically, multiple taps are performed around each sensor, and the sensor records the starting time of the elastic wave generated by each tap. The tapping time of the jth sensor is recorded as t j , the starting time of the elastic wave generated by the knocking of the i-th sensor after the j-th sensor is knocked is recorded as t ji During the test, pedestrians and construction on the dam surface are stopped to interfere with the sensor's collection of the elastic wave signal generated by the knocking.

[0119] According to the distance between the jth sensor and each sensor, and the relationship between speed and time, corresponding to each knocking hole, equation (8) is listed according to the two-point distance formula:

[0120] (x j -x 1 ) 2 +(y j -y 1 ) 2 +(z j -z 1 ) 2 =v j 2 (t j1 -t j ) 2 (8)

[0121] Substitute the coordinates of the first and second striking points, the striking moments of the corresponding striking points, and the values ​​of the starting moments of the elastic waves generated by the striking received by the i-th sensor after the striking points are struck into equation (8) to solve the equivalent wave velocity v of the dam body. 1 =4078m / s,v 2 =4105m / s, and then calculate the average equivalent wave velocity of the dam body v=4091.5m / s.

[0122] During the operation of the reservoir dam, the monitoring area is monitored through the microseismic monitoring system to determine the source location and time of microseismic events generated in the monitoring area. The method for determining the source location and time of microseismic events generated in the monitoring area is as follows:

[0123] Assume that the coordinates of the epicenter of the microseismic event are (x k ,yk,z k), the time when the microseism occurs is t k , define t ki is the starting time of the elastic wave generated by the microseismic event received by the i-th sensor. According to the distance between the source of the microseismic event and each sensor, and the relationship between speed and time, the following six equations are listed according to the two-point distance formula:

[0124] (x k -x 1 ) 2 +(y k -y 1 ) 2 +(z k -z 1 ) 2 =v k 2 (t k1 -t k ) 2 (9)

[0125] (x k -x 2 ) 2 +(y k -y 2 ) 2 +(z k -z 2 ) 2 =v k 2 (t k2 -t k ) 2 (10)

[0126] (x k -x 3 ) 2 +(y k -y 3 ) 2 +(z k -z 3 ) 2 =v k 2 (t k3 -t k ) 2 (11)

[0127] (x k -x 4 ) 2 +(y k -y 4 ) 2 +(z k -z 4 ) 2 =v k2 (t k4 -t k ) 2 (12)

[0128] (x k -x 5 ) 2 +(y k -y 5 ) 2 +(z k -z 5 ) 2 =v k 2 (t k5 -t k ) 2 (13)

[0129] (x k -x 6 ) 2 +(y k -y 6 ) 2 +(z k -z 6 ) 2 =v k 2 (t k6 -t k ) 2 (14)

[0130] By combining the above six equations and substituting the average equivalent wave velocity v of the rock mass, the coordinates of each sensor, and the value of the starting time when each sensor receives the elastic wave generated by the microseismic event, the coordinates of the epicenter of the microseismic event (x k ,y k , z k ) and the time t at which the microseism occurs k .

[0131] Step S3-4: The microseismic sensor B1 is connected to the high-speed AD converter B3 via the signal transmission line B8, the storage B4, the clock B5 and the FGPA microcontroller B6 are connected via internal cables, and the communication component B7 sends a signal to the cloud data analysis system C.

[0132] Step S3-5: Monitor the potential seepage channels of the dam body. Due to the large fluctuation of reservoir water level and frequent stress changes inside the rockfill dam body, internal cracks are generated, and then seepage channels are generated; the incubation or expansion of cracks causes a certain amplitude of vibration in the dam body, and the dam body transmits vibration signals through the rockfill medium. Multiple microseismic sensors B1 collect microseismic signals caused by seepage cracking inside the dam body and transmit them to the cloud data analysis system C through the signal transmission line B8.

[0133] Specifically, in the cloud analysis (see Figure 7 ) period, the real-time collected soundprint of the seepage microseismic signal (see Figure 8(a)-Figure 8(d) ) and the historical real-time statistics of the hypocenter positions of microseismic events occurring in the monitoring area, mark the hypocenter positions in real time in a three-dimensional rectangular coordinate system, and obtain a spatial distribution map of the hypocenter positions. Combined with the distribution of the hypocenter positions in the spatial distribution map of the hypocenter positions, a judgment is made. When the hypocenter positions of the seepage microseismic events are distributed and aggregated in a strip-like or surface-like manner in one or some local areas of the monitoring area, there may be potential seepage channels in the corresponding local areas; if the hypocenter positions of the microseismic events are discretely distributed in one or some local areas of the monitoring area without aggregation, it means that there are no structural surfaces such as proven controlling faults in the corresponding local areas, that is, there are no potential seepage channels in the corresponding local areas.

[0134] Specifically, the voiceprint features of the acoustic emission signal are briefly described.

[0135] The voiceprint of the acoustic emission signal is a joint expression of the acoustic emission signal in the time domain and frequency domain. The expression of the acoustic emission signal in the time domain is a wave file, which can show the characteristics of the amplitude of the acoustic emission signal changing with time. When the waveform file of the acoustic emission signal in the entire time domain is FFT-transformed, the acoustic emission signal can be transferred from the time domain to the frequency domain, but the above-mentioned spatial transformation can only be reflected in a two-dimensional plane state. By framing and windowing the speech signal for STFFT transformation, the changes in the acoustic emission signal are reflected in the three-dimensional space, which is the time domain-frequency domain-amplitude signal distribution diagram of the acoustic emission signal. The stripes on the voiceprint are the amplitude representation of the acoustic emission signal, and the depth of the color of the stripes is the energy of the acoustic signal. The amplitude of the color spectrum from top to bottom indicates that the energy is getting bigger and bigger. The density of the stripes reflects the number of acoustic emission signals at this moment, and the depth of the color reflects the high and low degree of the signal amplitude. From the above analysis, we can see that for any acoustic emission signal, the texture features of the voiceprint pattern are unique, which provides a classification basis for the recognition of voiceprint patterns.

[0136] Step S3-6: Disassemble the device. Separate the signal acquisition device B from the fixing device A, the microseismic signal sensor B1 and the signal transmission line B8 successively; loosen the fixing nut A4 counterclockwise to restore to the separated state, remove the fixing plate A4 and the spring A3; separate the signal transmission line and the high-speed AD converter B3, and disassemble multiple fixing devices.

[0137] During the monitoring process of Example 3, after 60 days of monitoring, a total of 221 dam microseismic seepage rupture microseismic events occurred, as shown in Table 1.

[0138] Table 1 IVM model identification results of seepage rupture microseismic signals

[0139]

[0140] The spatial distribution of microseismic events is shown in Figure 5 As shown in Figure 2, the epicenter of the microseismic event occurred in a local area near the upstream dam surface at the dam pile number S0+100 ( Figure 5 The local area (the local area framed by the dotted line) is concentrated and distributed in a strip shape, indicating that there is a seepage channel in this local area. It is suggested that during the operation of the dam body, protective measures should be taken in this local area, such as concrete grouting, to ensure the safe operation of the dam body.

[0141] Exemplarily, through the steps S3-1 to S3-6, multiple groups of seepage path prediction values ​​of the dam body to be tested can be obtained, which can be used to form a seepage channel surface to form a final processing result.

[0142] This embodiment overcomes the shortcomings of existing monitoring methods and can simply and quickly monitor whether there is a seepage channel inside the rockfill dam of the upper reservoir of a pumped-storage power station. Moreover, the method uses a pre-trained IVM learning model to identify the characteristics of the dam body to be tested in real time, such as the rupture type, and then predict whether there is a seepage channel in the dam section. It can realize accurate detection of the seepage channel of the dam body, and is particularly suitable for dam body seepage survey during the process of reservoir dam removal and reinforcement.

[0143] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A system for monitoring the seepage channel of a reservoir rockfill dam using microseismic signals, characterized in that: include: Fixed devices, signal acquisition devices, and cloud data analysis systems; The fixing device comprises a fixing plate, a fixing rod and screws, the fixing plate is connected to the signal acquisition device through the screws, through holes are provided at both ends of the fixing plate, and the fixing rod passes through the through holes, so that the signal acquisition device is installed on the dam body; The fixing device also includes an expansion screw, and the fixing rod is a multi-diameter hollow rod. A hole smaller than the diameter of the fixing rod is pre-drilled on the surface concrete panel, and the fixing rod is passed through the fixing plate starting from the top direction, and the expansion screw is inserted into the hole. The fixing nut is screwed into the outer plate surface of the fixing plate at the top of the fixing rod, so that the expansion screw inside the hole begins to expand and forms a firm whole with the dam body around the hole wall, so that the signal acquisition device is installed on the dam body; then the signal acquisition device is preliminarily fixed at the longitudinal axis hole of the fixing plate, and the screw is screwed in from the bottom end of the fixing plate to stably install the signal acquisition device, thereby constructing a vibration system composed of the rockfill dam panel, the fixing device and the signal acquisition device; The signal acquisition device includes a microseismic signal sensor for acquiring microseismic signals of rockfill material fracture and sending the microseismic signals to a cloud data analysis system; The fixing device further comprises a signal transmission protection tube, and the microseismic signal sensor is arranged inside the signal transmission protection tube; The signal transmission protection tube is prefabricated with an inner knob thread structure, and the fixing plate is prefabricated with a knob tooth structure, and the fixing device and the signal collection device are integrally fixed on the outer side or top of the fixing plate for re-fixing operation; The cloud data analysis system is used to determine whether the rockfill dam body is broken based on the received microseismic signals and the information vector machine model for identifying the type of rockfill dam body rupture; and is used to analyze the distribution of the earthquake source position in the earthquake source position spatial distribution map based on the received microseismic signals and the earthquake source position spatial distribution map to determine whether there is a potential seepage channel in the local area; The cloud data analysis system is used to determine whether the rockfill dam body is broken based on the received microseismic signal and using an information vector machine model for identifying the type of rockfill dam body rupture, and includes: Obtain microseismic signals of rockfill material rupture; Decomposing and reconstructing the microseismic signal by wavelet transform to obtain a feature vector; Based on the feature vector, using the trained information vector machine model for identifying the type of rockfill dam rupture, it is determined whether the rockfill dam is ruptured; The process of training the information vector machine model for identifying the rupture type of the rockfill dam body comprises: obtaining microseismic signals of the rupture of the rockfill body material under the action of non-seepage and seepage; using wavelet transform to decompose and reconstruct the microseismic signals of the rupture of the rockfill body material under the action of non-seepage and seepage, and constructing a training set; using the training set to train the information vector machine model for identifying the rupture type of the rockfill dam body until the accuracy of the prediction result reaches a set threshold, and obtaining a trained information vector machine model for identifying the rupture type of the rockfill dam body; The expression of the information vector machine model is: Where Φ(x) represents the cumulative probability density function of the standard normal distribution; θ is the covariance function hyperparameter; p represents the probability distribution; X and y represent the input feature vector set and output target vector of the training sample respectively; x * and * Respectively represent the input feature vector and output target of the sample to be predicted; μ * and σ * Represent the predicted mean and predicted variance respectively; The information vector machine model for identifying the type of rockfill dam rupture includes an IVM binary classification model; the classification of seepage rupture and non-seepage rupture of the rockfill dam rupture is achieved through the binary classification of the IVM model; The analyzing the distribution of the earthquake source positions in the earthquake source position spatial distribution map to determine whether there is a potential seepage channel in the local area specifically includes: When the hypocenter locations of seepage microseismic events are distributed and aggregated in a strip-like or surface-like manner in one or some local areas of the monitoring area, there are potential seepage channels in the corresponding local areas; if the hypocenter locations of microseismic events are discretely distributed in one or some local areas of the monitoring area without aggregation, there are no potential seepage channels in the corresponding local areas.

2. The system for monitoring the seepage channel of a reservoir rockfill dam using microseismic signals according to claim 1 is characterized in that: The signal acquisition device also includes an FGPA microcontroller and a power supply. The FGPA microcontroller is connected to the microseismic signal sensor through a signal transmission line. The FGPA microcontroller is connected to the power supply, high-speed AD converter, storage, clock and communication components. The power supply supplies power to the microseismic signal sensor, FGPA microcontroller, high-speed AD converter, storage, clock and communication components.

3. The system for monitoring the seepage channel of a reservoir rockfill dam using microseismic signals according to claim 1 is characterized in that: The cloud data analysis system includes a central processing unit, which is connected to a storage, a multimedia component, an I / O interface and a communication component.

4. A method for monitoring the seepage channel of a reservoir rockfill dam using microseismic signals, characterized in that: The system for monitoring the seepage channel of a reservoir rockfill dam using microseismic signals as claimed in any one of claims 1 to 3 comprises: Based on the received microseismic signals, the information vector machine model for identifying the type of rockfill dam rupture is used; the seepage rupture and non-seepage rupture of the rockfill dam rupture are classified through the binary classification of the IVM model; Obtain microseismic signals and spatial distribution maps of earthquake source locations; Based on the microseismic signals and the spatial distribution map of the earthquake source positions, the distribution of the earthquake source positions in the spatial distribution map of the earthquake source positions is analyzed to determine whether there are potential seepage channels in the local area; The analyzing the distribution of the earthquake source positions in the earthquake source position spatial distribution map to determine whether there is a potential seepage channel in the local area specifically includes: When the hypocenter locations of seepage microseismic events are distributed and aggregated in a strip-like or surface-like manner in one or some local areas of the monitoring area, there are potential seepage channels in the corresponding local areas; if the hypocenter locations of microseismic events are discretely distributed in one or some local areas of the monitoring area without aggregation, there are no potential seepage channels in the corresponding local areas.

5. The method for monitoring the seepage channel of a reservoir rockfill dam using microseismic signals according to claim 4 is characterized in that: The acquisition of the spatial distribution map of the earthquake source position includes: Assume that the coordinates of the epicenter of the microseismic event are (x k ,y k , z k ), the time when the microseism occurs is t k , define t ki is the starting time of the elastic wave generated by the microseismic event received by the i-th sensor. According to the distance between the source of the microseismic event and each sensor, and the relationship between speed and time, the following six equations are listed according to the two-point distance formula: (x k -x1) 2 +(y k -y1) 2 +(z k -z1) 2 =v k 2 (t k1 -t k ) 2 (x k -x2) 2 +(y k -y2) 2 +(z k −z2) 2 =v k 2 (t k2 -t k ) 2 (x k -x3) 2 +(y k -y3) 2 +(z k -z3) 2 =v k 2 (t k3 -t k ) 2 (x k -x4) 2 +(y k -y4) 2 +(z k -z4) 2 =v k 2 (t k4 -t k ) 2 (x k -x5) 2 +(y k -y5) 2 +(z k -z5) 2 =v k 2 (t k5 -t k ) 2 (x k -x6) 2 +(y k -y6) 2 +(z k -z6) 2 =v k 2 (t k6 -t k ) 2 By combining the above six equations, substituting the average equivalent wave velocity v of the rock mass, the coordinates of each sensor, and the value of the starting time of the elastic wave generated by the microseismic event received by each sensor, the coordinates of the epicenter of the microseismic event (x k ,y k , z k ) and the time t at which the microseism occurs k ; According to the coordinates of the epicenter of the microseismic event (x k ,y k , z k ) and the time t at which the microseism occurs k , draw a spatial distribution map of the earthquake source location.

Images