A device and method for detecting the water content of the surface layer of surrounding rock of a pumped storage underground powerhouse

Abstract

The present application belongs to the technical field of microseismic signal detection, and provides a device and method for detecting the water content of the surface layer of surrounding rock of a pumped storage underground powerhouse. The device comprises a signal acquisition module, a signal excitation module and a signal analysis and visualization module, wherein the signal acquisition module and the signal excitation module are connected with the signal analysis and visualization module; the signal excitation module is used for continuously applying equal-amplitude vibration to the surrounding rock area to be measured to excite multiple standard vibration waves; the signal acquisition module is used for collecting microseismic signals generated on the surface of the surrounding rock to be measured due to excitation; and the signal analysis and visualization module is used for generating a Mel time-frequency voiceprint according to the microseismic signals, and identifying the water content of the surrounding rock according to the Mel time-frequency voiceprint. The present application fully utilizes the high representation ability of the internal properties of rocks possessed by microseismic signals, and realizes the detection of the water content of the surrounding rock of an underground powerhouse.

Description

Technical Field

[0001] This invention belongs to the field of microseismic signal detection technology, specifically relating to a device and method for detecting the surface water content of the surrounding rock of a pumped storage underground powerhouse. Background Technology

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

[0003] Currently, methods for detecting rock moisture content are mainly divided into indoor and outdoor methods. Indoor methods primarily refer to the gravimetric method, which involves a series of procedures such as sampling, drying, and weighing to obtain the rock moisture content result. This method has advantages such as ease of operation, low cost, and minimal tools, and has long been the primary method for detecting rock moisture content. However, the gravimetric method requires drilling holes in the rock mass to be tested, and the sampled rock blocks must be transported to indoor conditions for drying and testing, resulting in a time-consuming limitation. Outdoor methods mainly involve installing sensors (a standalone moisture sensor or a combination of moisture anchor and sensor) into the rock mass. By monitoring the dynamically changing dielectric constant within the rock mass and performing corresponding calculations, the moisture content of the surrounding rock is obtained. Similarly, this method requires installing the sensor in pre-drilled holes within the tunnel surrounding rock, making installation and disassembly complex and consuming significant manpower and resources. In summary, while the above methods can meet the requirements for detecting the surface moisture content of surrounding rock, the implementation process is complex and cumbersome, making it difficult to achieve rapid and convenient detection of surrounding rock moisture content.

[0004] Microseismic signals are low-frequency stress waves released during the initiation, development, propagation, and penetration of cracks in rocks, ultimately leading to instability. Their frequency range is between 10 and 20 Hz. -1 ~10 -3The frequency range of Hz, belonging to the low-frequency acoustic signal band, has garnered increasing attention from academic and engineering communities in rock mechanics and engineering due to its high correlation with rock fracture. It has been successfully applied to safety monitoring in major infrastructure projects such as highways, water conservancy projects, and mining tunnels, earning it the nickname "rock fracture stethoscope." Furthermore, the microseismic signals generated by rock fractures with different water contents exhibit significant differences. For example, higher water content rocks produce more high-amplitude microseismic signals and exhibit greater fluctuations in fractal dimension and b-value. Simultaneously, the microseismic acoustic signature, which considers both time and frequency domain characteristics, can better characterize the differences in fractures with different rock water contents. Therefore, quantitatively detecting the surface water content of the surrounding rock of pumped-storage underground powerhouses using microseismic signals is highly feasible. However, in reality, current microseismic signal detection devices are mostly used for long-term monitoring of tunnel surrounding rock stability, and their application for detecting the surface water content of the surrounding rock of pumped-storage underground powerhouses is relatively rare. Meanwhile, existing microseismic signal detection devices are often used as a passive detection method. However, in monitoring the surface water content of surrounding rock, it is necessary to actively generate microseismic signals on the surface of the surrounding rock, and then collect, process, and analyze these signals in real time, requiring them to be used as an active detection method. Therefore, existing microseismic signal detection devices are insufficient to meet the needs of water content detection in the surrounding rock of underground powerhouses. This paper proposes a device and method suitable for detecting the surface water content of surrounding rock in pumped storage underground powerhouses, which is of great significance for the safe operation of pumped storage power stations. Summary of the Invention

[0005] To address the aforementioned problems, this invention proposes a device and method for detecting the surface water content of the surrounding rock of pumped storage underground powerhouses. This invention fully utilizes the high characterization capability of microseismic signals of the internal properties of rocks, enabling efficient and convenient acquisition of large amounts of microseismic signals. Furthermore, it automatically extracts Mel acoustic signatures from the microseismic signals and identifies them using a pre-trained PASST model, thereby achieving the detection of the water content of the surrounding rock of the underground powerhouse. This overcomes the limitations of traditional methods for detecting the water content of surrounding rock, which are cumbersome, time-consuming, and damage to the original rock structure (through drilling or sensor installation).

[0006] According to some embodiments, the present invention adopts the following technical solution:

[0007] In the first aspect, the present invention provides a device for detecting the surface water content of the surrounding rock of a pumped storage underground powerhouse.

[0008] A device for detecting the surface water content of the surrounding rock of a pumped storage underground powerhouse includes: a signal acquisition module, a signal excitation module, and a signal analysis and visualization module, wherein the signal acquisition module and the signal excitation module are both connected to the signal analysis and visualization module;

[0009] The signal excitation module is used to continuously apply equal-amplitude vibrations to the area of ​​the surrounding rock to be tested, thereby generating multiple standard vibration waves.

[0010] The signal acquisition module is used to acquire the micro-vibration signal generated on the surface of the surrounding rock under test due to excitation.

[0011] The signal analysis and visualization module is used to generate a Mel time-frequency acoustic pattern based on the microseismic signal; and to identify the water content of the surrounding rock based on the Mel time-frequency acoustic pattern.

[0012] Furthermore, the signal excitation module is located at the front end of the signal acquisition module and is used to fit against the rock wall of the surrounding rock to be tested.

[0013] Furthermore, the signal acquisition module includes several micro-vibration signal sensors, a signal transmission cable, a fixed frame, and a sturdy handle;

[0014] The sturdy grip is fixedly connected to the fixed frame, and the fixed frame is equipped with several micro-vibration signal sensors.

[0015] The signal acquisition module and the signal excitation module are connected to the signal analysis and visualization module via a signal transmission cable.

[0016] Furthermore, the fixed frame includes a sensor carrier plate, an annular guide rail, and a connecting rod. The sturdy handle is connected to the annular guide rail via the connecting rod. Several sensor carrier plates are provided on the annular guide rail, and the sensor carrier plates are used to fix the micro-vibration signal sensor.

[0017] Furthermore, the sensor carrier plate includes a retractable reverse buckle, a fixed forward slot, and an embedded rail;

[0018] The retractable reverse buckle and the fixed forward slot form a pair of equally sized inlaid structures, and the micro-vibration signal sensor is located between the retractable reverse buckle and the fixed forward slot;

[0019] The embedded rail integrates the sensor carrier plate, which carries the micro-vibration signal sensor, onto the annular guide rail.

[0020] Furthermore, the signal excitation module includes, from the inside out, a guide shaft, a linear mover, a coil support, an energized coil, a guide seat, and a vibration excitation block;

[0021] The energized coil is fixed on the coil support. The analysis module and visualization module supply power to generate current inside the energized coil, forming a constant magnetic field to drive the linear mover and the linear mover of the hollow cylinder of the vibration excitation block to perform low-frequency periodic reciprocating motion along the guide axis.

[0022] Furthermore, the step of identifying the water content of the surrounding rock based on the Mel time-frequency acoustic pattern specifically includes: using the PASST deep learning model to identify the water content of the surrounding rock to be detected based on the Mel time-frequency acoustic pattern.

[0023] Secondly, the present invention provides a method for detecting the surface water content of the surrounding rock of a pumped storage underground powerhouse.

[0024] A method for detecting the surface water content of the surrounding rock of a pumped-storage underground powerhouse, employing the apparatus described in the first aspect for detecting the surface water content of the surrounding rock of a pumped-storage underground powerhouse, comprising:

[0025] Acquire several standard microseismic signals;

[0026] The microseismic signal is converted into a nonlinear frequency scale-sensitive Mel time-frequency acoustic pattern using a short-time Fourier time-frequency analysis algorithm based on Mel filter.

[0027] The PASST deep learning model was used to identify the features of the microseismic Mel-frequency acoustic pattern and to identify the water content of the surrounding rock relative to multiple standard microseismic signals.

[0028] Calculate the average water content of the measuring points in the surrounding rock to be tested.

[0029] Furthermore, the acquisition of several standard microseismic signals specifically includes: using a signal excitation module to continuously and automatically excite standardized tapping on the surface of the surrounding rock to be tested, and simultaneously acquiring multiple microseismic signal sensors to obtain a unified standard microseismic signal waveform.

[0030] Furthermore, the specific steps for identifying the features of the microseismic Mel-frequency acoustic pattern using the PASST deep learning model include: extracting overlapping patches from the Mel-frequency acoustic pattern, linearly projecting the overlapping patches onto a vector to form the input sequence of the Transformer; adding classification and knowledge distillation labels using Patchout based on the input sequence of the Transformer; transforming the sequences of classification and distillation labels using TransformerEncoder, and processing the average values ​​of the transformed classification and distillation knowledge labels using MLPClassifier (Multilayer Perceptron Classifier) ​​to obtain the surrounding rock water content relative to multiple standard microseismic signals.

[0031] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0032] (1) The present invention provides a device for detecting the surface water content of the surrounding rock of a pumped storage underground powerhouse. This device fully considers the high correlation between microseismic signals and the internal properties of the rock mass. It has a unique structural design that integrates active excitation of vibration signals, passive acquisition of microseismic signals and automatic signal analysis. It overcomes the limitations of the current rock water content detection method in terms of process - drilling (installation / sampling), and significantly reduces the time consumption and economic cost of rock water content detection. In particular, the detection process of this device belongs to the category of non-destructive testing, which is especially suitable for detecting the water content of the surrounding rock during the operation period of a pumped storage underground powerhouse.

[0033] (2) The device described in this invention adopts a structure design that uses multiple micro-seismic sensors to collect and magnetic field-driven vibration signals to automatically excite the structure. It can acquire a large number of standard micro-seismic signals in a short time at the site of the surrounding rock to be tested, which greatly enriches the data sample of the water content prediction of the surrounding rock to be tested. This effectively solves the problem of large uncertainty in the detection results caused by the discreteness of the rock mass, and can realize accurate detection of the water content of the surrounding rock.

[0034] (3) The PASST deep feature algorithm used in this invention is a novel Transformer model for acoustic signatures. This model incorporates a novel Patchout structure that optimizes and standardizes the Transformer model, which significantly reduces the computational complexity of training the Transformer model in the signal domain while also improving its generalization ability. Therefore, this algorithm can maintain a high prediction accuracy even with low datasets, providing strong technical support for achieving efficient and accurate detection of water content in surrounding rocks.

[0035] (4) The present invention provides a method for detecting the water content of the surrounding rock in the surface layer of a pumped storage underground powerhouse. Based on the Mel time-frequency acoustic signature features that can characterize both the time and frequency domain information of the surrounding rock microseismic signals, the method uses a PASST deep learning model to identify the differences in the patch structure of the Mel time-frequency acoustic signature features of the microseismic signals, thereby achieving prediction of the surrounding rock water content. This method overcomes the contradiction between the accuracy of existing rock water content detection methods and the difficulty in balancing complex procedures (requiring in-situ drilling), and can achieve efficient and accurate detection of rock water content.

[0036] (5) The microseismic monitoring technology used in this invention is a common means of monitoring the stability and safety of rock mass. This invention extends the technology from the field of passive safety monitoring to the field of active detection by proposing a device and method for detecting the water content of the surrounding rock of pumped storage underground powerhouse, thus expanding the application scope of microseismic technology. At the same time, this invention not only significantly reduces the time and cost of rock water content detection, but also significantly improves the efficiency and accuracy of rock water content detection. Attached Figure Description

[0037] The accompanying drawings, which form part of this invention, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an improper limitation of the invention.

[0038] Figure 1 This is a schematic diagram of a device for detecting the surface water content of the surrounding rock of a pumped storage underground powerhouse, provided in Embodiment 1 of the present invention.

[0039] Figure 2(a) is a schematic diagram of the rear structure of a device for detecting the surface water content of the surrounding rock of a pumped storage underground powerhouse provided in Embodiment 1 of the present invention;

[0040] Figure 2(b) is a schematic diagram of the front end structure of a device for detecting the surface water content of the surrounding rock of a pumped storage underground powerhouse provided in Embodiment 1 of the present invention;

[0041] Figure 3(a) is a schematic diagram of the installation of the micro-vibration signal sensor and the sensor carrier plate provided in Embodiment 1 of the present invention;

[0042] Figure 3(b) is a side view of the micro-vibration signal sensor provided in Embodiment 1 of the present invention after installation;

[0043] Figure 3(c) is a front view of the micro-vibration signal sensor provided in Embodiment 1 of the present invention after installation;

[0044] Figure 4(a) is a front view of the signal excitation module structure provided in Embodiment 1 of the present invention;

[0045] Figure 4(b) is a side view of the signal excitation module structure provided in Embodiment 1 of the present invention;

[0046] Figure 5 This is a flowchart of a method for detecting the surface water content of the surrounding rock of a pumped storage underground powerhouse, provided in Embodiment 2 of the present invention.

[0047] Figure 6 This is a schematic diagram of the device arrangement array provided in Embodiment 2 of the present invention;

[0048] Figure 7(a) is an example of the microseismic Mel-ray acoustic waveform of the surrounding rock with a water content of 0.8% provided in Example 2 of the present invention;

[0049] Figure 7(b) is an example of a microseismic Mel-ray acoustic waveform of a surrounding rock with a water content of 1.6% provided in Example 2 of the present invention;

[0050] Figure 7(c) is an example of the microseismic Mel-ray acoustic waveform of the surrounding rock with a water content of 2.1% provided in Example 2 of the present invention;

[0051] Figure 8 This is a schematic diagram of the PASST depth algorithm architecture provided in Embodiment 2 of the present invention;

[0052] In the diagram, A—signal acquisition module; A1—micro-vibration signal sensor; A2—signal transmission cable; A3—fixed frame; A31—sensor carrier plate; A311—retractable reverse buckle; A312—fixed forward slot; A313—embedded rail; A32—circular guide rail; A33—connecting rod; A4—stable handle; B—signal excitation module; B1—guide seat; B2—guide shaft; B3—energized coil; B4—coil bracket; B5—linear mover; B6—vibration excitation block; C—signal analysis and visualization module; C1—signal processing device; C2—data storage device; C3—multimedia device; C4—power supply; C5—input device; C6—output device. Detailed Implementation

[0053] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0054] It should be noted that the following detailed descriptions are exemplary and intended to provide further illustration of the invention. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0055] It should be noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the scope of exemplary embodiments according to the invention. As used herein, the singular form is intended to include the plural form as well, unless the context clearly indicates otherwise. Furthermore, it should be understood that when the terms "comprising" and / or "including" are used in this specification, they indicate the presence of features, steps, operations, devices, components, and / or combinations thereof.

[0056] In this invention, terms such as "upper," "front," "rear," and "bottom" indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are merely relational terms determined for the convenience of describing the structural relationship of the various components or elements of this invention, and do not specifically refer to any component or element in this invention, and should not be construed as limiting this invention.

[0057] In this invention, terms such as "fixed" and "connected" should be interpreted broadly, indicating a fixed connection, an integral connection, or a detachable connection; a direct connection or an indirect connection through an intermediate medium. Those skilled in the art can determine the specific meaning of these terms in this invention based on the specific circumstances, and they should not be construed as limitations on the invention.

[0058] Example 1

[0059] like Figure 1As shown in Figures 2(a), 2(b), 3(a), 3(b), 3(c), 4(a), and 4(b), this embodiment provides a device for detecting the surface water content of the surrounding rock of a pumped storage underground powerhouse. The device includes a signal acquisition module A, a signal excitation module B, and a signal analysis and visualization module C.

[0060] The signal acquisition module A is used to acquire microseismic signals generated by the surface of the surrounding rock under test. It includes multiple microseismic signal sensors A1, a signal transmission cable A2, a fixed frame A3, and a sturdy handle A4. The fixed frame A3 is used to fix the microseismic signal sensors A1 and connect the sturdy handle A4, and includes a sensor carrier plate A31, a circular guide rail A32, and a connecting rod A33. The sensor carrier plate A31 includes a retractable reverse buckle A311, a fixed forward slot A312, and an embedded rail A313. The retractable reverse buckle A311 and the fixed forward slot A312 are a pair of equally sized inlaid structures. By adjusting the depth of the upper screw of the retractable reverse buckle A311 within the sensor carrier plate A31, the reverse circular support at the bottom of its structure is gradually placed closer to the microseismic signal sensor A1 on the fixed forward slot A312. When the reverse circular support on the telescopic reverse buckle A311 is in complete and tight contact with the fixed forward slot A312, the micro-vibration signal sensor A1 can be installed and fixed on the sensor carrier plate A31. In addition, the sensor carrier plate A31 carrying the micro-vibration signal sensor A1 is integrally fitted onto the annular guide rail A32 through the embedded rail A313, and the signal transmission cable A2 is led out from the circular through hole at the end of the sensor carrier plate A31 to the signal analysis and visualization module C for connection.

[0061] The connecting rod A33 consists of a triangular support structure formed by three inclined auxiliary rods. The front end of each auxiliary rod is fitted and connected to the rear end of the annular guide rail A32 at a 180° angle. Furthermore, the rear end of each auxiliary rod is fixedly connected to the stabilizing handle A4, thereby connecting the multiple microseismic signal sensors A1, signal transmission cable A2, fixed frame A3, and stabilizing handle A4 in the signal acquisition module A into a single unit. Based on this, by pressing the stabilizing handle A4 laterally with one hand to constrain the lateral displacement of the microseismic signal sensors A1, good contact between the multiple microseismic signal sensors A1 and the surface of the surrounding rock to be measured can be achieved.

[0062] The signal excitation module B is used to continuously apply constant-amplitude vibrations to the area of ​​the surrounding rock under test, thereby generating multiple standard vibration waves. It includes a guide seat B1, a guide shaft B2, an energized coil B3, a coil support B4, a linear actuator B5, and a vibration excitation block B6. The energized coil B3 is fixed to the coil support B4 and is powered by the analysis module and visualization module C to generate current within it, thereby forming a constant magnetic field to drive the linear actuator B5 (a magnetically insulating material) and the hollow cylinder of the vibration excitation block B6. The linear actuator B5, a hollow cylinder, performs low-frequency periodic reciprocating motion along the guide shaft B2. Through the action of the inner axis (based on the linear actuator) and the outer axis (based on the vibration excitation block), the internal electrical energy can be converted into low-frequency vibrational kinetic energy with separation characteristics applied to the surface of the surrounding rock under test.

[0063] The signal analysis and visualization module C is used to generate Mel time-frequency acoustic waveforms from the acquired microseismic signals and identify the water content of the surrounding rock to be detected through a pre-trained PASST deep learning model, thereby visualizing the results (detection acoustic waveform / water content). It includes a signal processing device C1, a data storage device C2, a multimedia device C3, a power supply C4, an input device C5, and an output device C6. The signal analysis device or other desired functions are realized through the joint action of each component.

[0064] Example 2

[0065] like Figure 5 and Figure 6 As shown in the figure, this embodiment provides a method for detecting the surface water content of the surrounding rock of a pumped storage underground powerhouse, including:

[0066] Step 1: Use the device of this invention to acquire multiple standard microseismic signals from the surrounding rock measuring points.

[0067] Specifically, the standard microseismic signal data is obtained through a device for detecting the water content of the surrounding rock of a pumped storage underground powerhouse using microseismic signals. The device continuously and automatically (magnetically driven) excites standardized tapping on the surface of the surrounding rock to be tested through the signal excitation module in the device, and relies on the simultaneous acquisition mechanism of multiple microseismic signal sensors installed on the acquisition module, so that this embodiment can acquire a large number of uniform standard microseismic signal waveforms in a short time.

[0068] Optionally, the tapping method employs a multi-point, multiple-tapping strategy. The multi-point strategy refers to dividing the area to be tested in the surrounding rock into multiple measurement points. The area of ​​the measurement area must be larger than the area of ​​the signal excitation module. Loose and easily detached measurement points need to be pre-processed before the test. If the area of ​​the measurement area is too large, it can be divided into multiple smaller sub-measurement areas. Multiple tapping refers to performing multiple standardized taps at the same measurement point to reduce the dispersion of the microseismic signal.

[0069] Optionally, the microseismic signal dataset obtained by the multi-point, multi-impact strategy constitutes a set of original feature vectors with an indeterminate number. By establishing a multi-to-one relationship between microseismic signal data and rock water content, the discreteness bias problem caused by the detection method can be reduced, thereby significantly improving the accuracy of rock water content detection.

[0070] Step 2: Convert the microseismic signal into a nonlinear frequency scale-sensitive (more suitable for macroscopic rupture with high sensitivity at low frequencies) Mel time-frequency acoustic pattern using a short-time Fourier time-frequency analysis algorithm based on Mel filter. This pattern has dual representation capabilities in both the time and frequency domains.

[0071] Specifically, the Mel-scale voiceprint mapping utilizes a filter bank with a Mel frequency scale to process the corresponding frequency domain signal, transforming the linear spectrum into a non-linear Mel-scale voiceprint. Since the human ear is more sensitive to low-frequency signals and less sensitive to high-frequency signals, the Mel scale used in the Mel-scale voiceprint mapping exhibits a linear relationship with normal frequencies in the low-frequency range, while in the high-frequency range, due to weakened auditory perception, it shows a logarithmic relationship with normal frequencies. This processing highlights the low-frequency components of the signal, achieving both dimensionality reduction of the data and preservation of the basic information needed for the human ear to understand the original signal, offering advantages when processing large amounts of data.

[0072] Specifically, the Mel acoustic waveforms corresponding to different surface water contents of the surrounding rock are as follows: Figures 7(a)-7(c) As shown in Figure 7(a), the Mel acoustic waveform of the surrounding rock surface with a water content of 0.8% is obtained. It can be seen that the main frequency band of the waveform is mainly in the low-to-mid frequency range (271Hz), and the maximum energy amplitude is at a relatively low level. Figure 7(b) shows the Mel time-frequency acoustic waveform of the surrounding rock surface with a water content of 1.6%. It can be seen that the main frequency band of the waveform is still in the low-to-mid frequency range (271Hz), the maximum energy amplitude has increased, and higher energy bands have begun to appear in the low-to-high frequency range. Figure 7(c) shows the Mel acoustic waveform of the surrounding rock surface with a water content of 2.1%. It can be seen that bands with certain energy amplitudes appear in the low, mid, and high frequency ranges, and the maximum energy amplitude remains roughly unchanged. Therefore, the microseismic Mel acoustic waveform can effectively display the differences in water content of different surrounding rock surfaces in terms of frequency band distribution and energy, and can serve as a good characteristic indicator for detecting the water content of the surrounding rock surface of underground powerhouses.

[0073] Step 3: Use a pre-trained PASST deep learning model to identify the features of the microseismic Mel-time frequency acoustic pattern and identify the water content of the surrounding rock corresponding to multiple standard microseismic signals.

[0074] Preferably, the PASST deep feature algorithm used in the PASST model is a novel Transformer model for speaker maps. This model incorporates a novel Patchout structure that optimizes and standardizes the Transformer model, which significantly reduces the computational complexity of training the Transformer model in the signal domain, while also improving its generalization ability.

[0075] The specific process is as follows: Figure 8 As shown, overlapping patches are extracted from the InputSpectrogram; the overlapping patches are linearly projected onto the vector to form the input sequence of the Transformer; Patchout is applied, and classification labels and knowledge distillation labels are added; the average value of the transformed classification labels and distilled knowledge labels is then processed by the TransformerEncoder and the MLPClassifier classifier; the trained model is then output.

[0076] Specifically, the PASST model used in the algorithm is the PaSST-S model pre-trained on the AudioSet dataset. Through multiple pre-training sessions, the optimal parameters of this model are obtained. After training, the model with the highest accuracy on the test set is saved as the classification model.

[0077] Step 4: Calculate the average water content of the surrounding rock at the measuring point.

[0078] Preferably, the water content of the surrounding rock detected at multiple measuring points in steps 1 to 3 can be calculated by taking multiple low-frequency separable vibration wave signals from each measuring point, which means that multiple water content values ​​of the surrounding rock can be obtained from each measuring point.

[0079] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A device for detecting the surface water content of the surrounding rock of a pumped-storage underground powerhouse, characterized in that, include: The system includes a signal acquisition module, a signal excitation module, and a signal analysis and visualization module, wherein the signal acquisition module and the signal excitation module are both connected to the signal analysis and visualization module. The signal excitation module is used to continuously apply constant amplitude vibration to the area of ​​the surrounding rock to be tested, thereby generating multiple standard vibration waves. The signal excitation module includes, from the inside out, a guide shaft, a linear mover, a coil support, an energized coil, a guide seat, and a vibration excitation block. The energized coil is fixed on the coil support and is powered by the analysis module and visualization module to generate current inside the energized coil, forming a constant magnetic field to drive the linear mover and the hollow cylinder of the vibration excitation block to perform low-frequency periodic reciprocating motion along the guide shaft. The signal acquisition module is used to acquire micro-vibration signals generated by excitation on the surface of the surrounding rock to be tested. The signal acquisition module includes several micro-vibration signal sensors, a signal transmission cable, a fixed frame, and a stable handle. Several micro-vibration signal sensors are mounted on the fixed frame. The fixed frame includes a sensor carrier plate, an annular guide rail, and a connecting rod. The sensor carrier plate includes a retractable reverse buckle, a fixed forward slot, and an embedded rail. The retractable reverse buckle and the fixed forward slot form a pair of equal-diameter inlay structures, and the micro-vibration signal sensors are located between the retractable reverse buckle and the fixed forward slot. The embedded rail embeds the sensor carrier plate carrying the micro-vibration signal sensors into the annular guide rail. The connecting rod consists of a triangular support structure composed of three inclined secondary rods. The front end of the secondary rod is engaged with the 180° end of the rear end of the annular guide rail, and the rear end of the secondary rod is fixedly connected to the stable handle. The signal analysis and visualization module is used to generate a Mel time-frequency acoustic pattern based on the microseismic signal; and to identify the water content of the surrounding rock based on the Mel time-frequency acoustic pattern.

2. The device for detecting the surface water content of the surrounding rock of a pumped-storage underground powerhouse according to claim 1, characterized in that, The signal excitation module is located at the front end of the signal acquisition module and is used to fit against the rock wall of the surrounding rock to be tested.

3. The device for detecting the surface water content of the surrounding rock of a pumped storage underground powerhouse according to claim 1, characterized in that, The sturdy grip is fixedly connected to the fixed frame; The signal acquisition module and the signal excitation module are connected to the signal analysis and visualization module via a signal transmission cable.

4. The device for detecting the surface water content of the surrounding rock of a pumped storage underground powerhouse according to claim 3, characterized in that, The sturdy grip is connected to an annular guide rail via a connecting rod. Several sensor carrier plates are provided on the annular guide rail, and the sensor carrier plates are used to fix the micro-vibration signal sensor.

5. The device for detecting the surface water content of the surrounding rock of a pumped-storage underground powerhouse according to claim 1, characterized in that, The specific steps for identifying the water content of the surrounding rock based on the Mel time-frequency acoustic pattern include: using the PASST deep learning model to identify the water content of the surrounding rock to be detected based on the Mel time-frequency acoustic pattern.

6. A method for detecting the surface water content of the surrounding rock of a pumped-storage underground powerhouse, characterized in that, The apparatus for detecting the surface water content of the surrounding rock of a pumped storage underground powerhouse as described in any one of claims 1-5 comprises: Acquire several standard microseismic signals; The microseismic signal is converted into a nonlinear frequency scale-sensitive Mel time-frequency acoustic pattern using a short-time Fourier time-frequency analysis algorithm based on Mel filter. The PASST deep learning model was used to identify the features of the microseismic Mel-frequency acoustic pattern and to identify the water content of the surrounding rock relative to multiple standard microseismic signals. Calculate the average water content of the measuring points in the surrounding rock to be tested.

7. The method for detecting the surface water content of the surrounding rock of a pumped storage underground powerhouse according to claim 6, characterized in that, The acquisition of several standard microseismic signals specifically includes: using a signal excitation module to continuously and automatically excite standardized tapping on the surface of the surrounding rock to be tested, and using multiple microseismic signal sensors installed simultaneously to acquire uniform standard microseismic signal waveforms.

8. The method for detecting the surface water content of the surrounding rock of a pumped storage underground powerhouse according to claim 6, characterized in that, The specific steps for identifying the features of the microseismic Mel-frequency acoustic pattern using the PASST deep learning model include: extracting overlapping patches from the Mel-frequency acoustic pattern, linearly projecting the overlapping patches onto a vector to form the input sequence of the Transformer; adding classification and knowledge distillation labels using Patchout based on the input sequence of the Transformer; transforming the sequences of classification and distillation labels using TransformerEncoder, and processing the average values ​​of the transformed classification and distillation knowledge labels using the MLP Classifier to obtain the water content of the surrounding rock relative to multiple standard microseismic signals.

Images