A geological risk early warning and active support monitoring system for a pumped storage power station

Abstract

This invention relates to the field of underground engineering safety monitoring technology, and discloses a geological risk early warning and active support monitoring system and method for pumped storage power stations. The system consists of multimodal intelligent anchoring units, a data acquisition and control host, and a data processing and early warning analysis platform. The multimodal intelligent anchoring units are deployed in the rock mass, integrating multiple sensing functions and forming a three-dimensional array. The data acquisition and control host receives instructions, drives the anchoring units to detect and collect raw data, and the data processing and early warning analysis platform processes the data, runs algorithms, and issues early warnings. The method includes obtaining an initial physical property benchmark model of the surrounding rock, calibrating the in-situ pressure sensitivity coefficient, periodically acquiring a time-series variation model of physical properties, and then inverting the stress field and fusing multi-dimensional risk characteristics for early warning. This invention achieves accurate early warning and active support monitoring of geological risks in pumped storage power stations through multimodal detection, data acquisition and processing, and multi-feature fusion analysis, thereby improving the safety of underground engineering.

Description

Technical Field

[0001] This invention relates to the field of underground engineering safety monitoring technology, specifically a geological risk early warning and active support monitoring system for pumped storage power stations. Background Technology

[0002] Pumped storage power stations are characterized by their large scale, deep burial, and complex geological conditions. The long-term stability of the surrounding rock is crucial to the safety of the entire project. Therefore, effectively monitoring the condition of the surrounding rock and providing timely warnings of geological risks is a core task during the construction and operation of the project.

[0003] Existing methods for monitoring the stability of surrounding rock mainly include deformation monitoring and stress monitoring. Deformation monitoring typically uses equipment such as multi-point displacement gauges and convergence gauges to measure the displacement or convergence deformation of the surrounding rock. This method can reflect the macroscopic trend of surrounding rock deformation, but it measures the result caused by stress changes, which has a certain lag. Moreover, it is usually a point or line measurement, making it difficult to provide complete spatial deformation information within the surrounding rock. Stress monitoring directly measures the stress state at specific points by deploying stress gauges or strain gauges in the rock mass. However, due to the high degree of non-uniformity of the stress field in the rock mass, a limited number of point measurements are insufficient to accurately characterize the stress distribution and evolution of the entire area.

[0004] In recent years, geophysical exploration methods such as resistivity tomography and elastic wave velocity tomography have also been applied to detect the internal structure and state of surrounding rocks. These methods can provide spatial distribution images of the physical properties inside the surrounding rock, thereby qualitatively inferring the location of water-rich areas, fractured areas, or stress concentration areas. However, a major limitation of these methods is that the quantitative relationship between the physical properties of the rock mass and the stress state is usually unclear, and it largely depends on laboratory rock sample tests. Parameters obtained at the laboratory scale are difficult to directly apply to complex, fractured, and heterogeneous rock masses in the field. Therefore, it is difficult to achieve quantitative inversion of the stress field inside the surrounding rock based solely on geophysical images.

[0005] Furthermore, for structures using anchor bolts for active support, conventional monitoring often focuses on measuring the axial force at the bolt head. This only reflects the overall stress on the bolt and cannot reveal the load distribution along the bolt or the collaborative working state between the bolt and the surrounding rock mass. Simultaneously, most existing early warning systems are based on threshold values ​​for single parameters such as displacement and stress, failing to comprehensively consider the coupled effects of multiple factors such as stress concentration, rock mass damage, and groundwater activity. This makes it difficult to comprehensively and accurately identify precursory information regarding rock mass instability and failure. Summary of the Invention

[0006] To address the shortcomings of existing technologies, this invention provides a geological risk early warning and active support monitoring system for pumped storage power stations. This system solves the problem that existing methods for monitoring the stability of surrounding rock in underground engineering projects such as pumped storage power stations mostly rely on the measurement of apparent deformation quantities such as displacement and strain, or qualitative detection based on geophysical methods. These methods are difficult to directly and quantitatively invert the stress field inside the surrounding rock, resulting in a lag in the identification of geological risks and an inability to accurately assess the collaborative working state between the active support system and the surrounding rock.

[0007] To achieve the above objectives, the present invention provides the following technical solution: A geological risk early warning and active support monitoring system for pumped storage power stations includes: Multiple multimodal intelligent anchoring units 100 are deployed inside the rock mass. While providing support functions, the multimodal intelligent anchoring units 100 integrate sensors for electrical and acoustic detection as well as sensors for measuring their own strain and temperature. The multiple multimodal intelligent anchoring units 100 form a three-dimensional detection array in the rock mass. The data acquisition and control host 200 connected to the multimodal intelligent anchoring unit 100 drives the multimodal intelligent anchoring unit 100 array to perform specific detection tasks and collect raw data returned by all sensors according to preset instructions or instructions from the data processing and early warning analysis platform 300. The data processing and early warning analysis platform 300 is communicatively connected to the data acquisition and control host 200. The data processing and early warning analysis platform 300 receives and stores data from the data acquisition and control host 200, runs core algorithms, and generates analysis results and early warning information.

[0008] In the aforementioned pumped storage power station geological risk early warning and active support monitoring system, the base of the multimodal intelligent anchoring unit 100 is the anchor rod or anchor cable commonly used in the underground cavern support project of pumped storage power stations. It is made of high-strength steel and is used to bear the design support load.

[0009] In the aforementioned pumped storage power station geological risk early warning and active support monitoring system, the multimodal intelligent anchoring unit 100 integrates sensing elements for electrical detection. The rod base serves as the first electrode. The anchoring unit also includes at least one accompanying electrode made of flexible conductive material. The accompanying electrode is arranged parallel to the axial direction of the rod base and is electrically insulated from the rod base by its own insulating sheath or a preset insulating isolator. After the anchor borehole is filled and cured by grouting material, the rod base and the accompanying electrode form an electrical path with the surrounding rock mass, constituting two or more electrodes in a four-electrode measurement device for resistivity tomography.

[0010] In the aforementioned pumped storage power station geological risk early warning and active support monitoring system, the multimodal intelligent anchoring unit 100 integrates a sensing element for acoustic detection. A piezoelectric transducer is encapsulated at the end of the anchoring unit or at a predetermined position along the length of the rod. The piezoelectric transducer is encapsulated and fixed by a waterproof and sealed shell that matches the mechanical properties of the surrounding rock. An acoustic impedance matching layer is filled between the transducer and the encapsulation shell. The signal lead of the piezoelectric transducer is laid along the length of the rod base and led out together with other cables to the orifice end of the anchoring unit.

[0011] In the aforementioned pumped storage power station geological risk early warning and active support monitoring system, the multimodal intelligent anchoring unit 100 integrates a sensing element for distributed self-sensing. A distributed optical fiber sensor is arranged along the axial direction of the pole base. The optical fiber sensor is fixed in a pre-made groove on the surface of the pole or directly integrated into its interior during the pole manufacturing process. The optical fiber sensor has an enhanced protective layer and works based on the Brillouin scattering or Rayleigh scattering principle. It is used to continuously measure the axial strain and temperature distribution along the entire length of the pole base.

[0012] In the aforementioned pumped storage power station geological risk early warning and active support monitoring system, the data acquisition and control host 200 includes: The central processing and clock synchronization module 201 includes an embedded industrial computer or a programmable logic controller, which is responsible for executing a preset sequence of control instructions, managing data storage, handling network communication with the data processing and early warning analysis platform 300, and integrating a high-precision clock source to provide a unified time reference for all data acquisition processes. The multi-channel resistivity meter module 202 consists of a high-power constant current source, a high-precision voltmeter, and a large-scale relay switch matrix. The relay switch matrix is ​​connected to the electrical electrodes of all the multi-modal intelligent anchoring units 100. The central processing and clock synchronization module 201 controls the on / off state of the relay switch matrix to dynamically connect the constant current source and the voltmeter to four designated electrodes, thereby realizing automatic scanning and measurement of different four-electrode device combinations. The acoustic transceiver control module 203 consists of a high-voltage pulse generator and a multi-channel high-speed data acquisition card. The high-voltage pulse generator is used to generate the instantaneous high-voltage electrical signal required to drive the piezoelectric transducer to emit elastic waves. The multi-channel high-speed data acquisition card is responsible for synchronously amplifying, filtering and converting the weak analog voltage signal output by the piezoelectric transducer, which acts as a receiver. The module is controlled by the central processing and clock synchronization module 201 to realize the excitation of a specific transducer and the synchronous acquisition of data from all channels. The fiber optic demodulator interface module 204 provides a standard data and control interface for commercial or customized Brillouin / Rayleigh fiber optic demodulators. The data acquisition and control host 200 sends acquisition commands to the fiber optic demodulator through this interface and receives strain and temperature data distributed along the fiber optic cable calculated by the demodulator. An external trigger and synchronization interface is used to receive synchronization signals from external devices. When a preset trigger signal is received, the central processing and clock synchronization module 201 immediately executes a predefined fast data acquisition sequence.

[0013] In the aforementioned pumped storage power station geological risk early warning and active support monitoring system, the data acquisition and control host 200 operates in a normalized periodic scanning mode, including the following steps: When performing resistivity tomography scanning, the central processing and clock synchronization module 201 controls the relay switch matrix of the multi-channel resistivity meter module 202 according to the predefined measurement sequence file, sequentially configures the power supply electrode pair and the measurement electrode pair, performs a power-on-measurement operation for each configuration, and caches the recorded current value, potential difference value and corresponding electrode combination information together, and traverses all electrode combinations in the sequence file. After performing elastic wave tomography scanning and completing the resistivity scan, the central processing and clock synchronization module 201 switches to the acoustic wave transceiver control module 203. The module selects one piezoelectric transceiver in the array as the vibration source in turn according to the preset excitation sequence and drives the high-voltage pulse generator to excite it. At the same time, the high-speed data acquisition card synchronously records the waveform data of other piezoelectric transceivers as receivers within the preset time window, traversing all preset vibration sources. The central processing and clock synchronization module 201 collects distributed fiber optic sensing data. Through the fiber optic demodulator interface module 204, it sends a collection command to the fiber optic demodulator and receives the returned strain and temperature data distributed along all connected distributed fiber optic sensors. After completing the above acquisition steps, the central processing and clock synchronization module 201 associates all resistivity data, elastic wave waveform data, and distributed optical fiber data obtained within the monitoring period with the unified timestamp provided by the high-precision clock source, packages them into a data file, and uploads it to the data processing and early warning analysis platform 300 through the network communication interface.

[0014] In the aforementioned pumped storage power station geological risk early warning and active support monitoring system, the data acquisition and control host 200 operates in tension synchronous acquisition mode, including the following steps: The central processing and clock synchronization module 201 continuously monitors the status of the external trigger and synchronization interface by listening to external trigger signals. Triggering a high-speed acquisition sequence: When a predefined trigger signal is detected on the interface, the host pauses the regular scanning task and enters the tension synchronous acquisition mode to continuously and rapidly perform resistivity and elastic wave tomography scans at a frequency higher than that of regular scanning. Simultaneously recording time-varying data, while performing high-speed tomographic scanning, the host computer acquires distributed optical fiber sensing data on the multimodal intelligent anchoring unit 100 being tensioned at high frequency through the optical fiber demodulator interface module 204. A synchronous event dataset is generated. When the tensioning end signal is detected on the external trigger and synchronization interface, the host stops high-speed acquisition, packages all high-frequency tomographic imaging data and corresponding distributed strain data acquired during the entire tensioning process into a synchronous event dataset with start and end timestamps, and uploads it to the data processing and early warning analysis platform 300, and then returns to the normalized periodic scanning mode.

[0015] A method for geological risk early warning and active support monitoring of pumped storage power stations, applied to the aforementioned system, includes the following steps: A baseline model of initial physical properties is obtained. Multiple multimodal intelligent anchoring units 100 are deployed in the surrounding rock to be monitored. A baseline scan is performed to collect the original data set of resistivity and elastic wave at the initial moment. Based on the original data set, a three-dimensional baseline model characterizing the initial physical properties of the surrounding rock is constructed. The in-situ pressure sensitivity coefficient is calibrated by simultaneously collecting strain data of the unit and geophysical response data measured by the surrounding anchor unit array during the application of tension load to at least one multimodal smart anchor unit 100. Based on the stress disturbance and physical property changes caused by the tension load, the in-situ pressure sensitivity coefficient field covering the monitoring area is solved and constructed. The physical properties time-series variation model is obtained, resistivity and elastic wave tomography scans are periodically performed, and the data collected in each period are inverted to generate a series of three-dimensional time-series models characterizing the evolution of the internal physical properties of the rock mass over time. The stress field is inverted and integrated for early warning. The in-situ pressure sensitivity coefficient is used to interpret the time-series variation model of the physical properties to invert the spatiotemporal evolution of the stress field inside the surrounding rock. Furthermore, multi-dimensional risk characteristics such as stress, damage, and seepage are integrated, and early warning information is generated when typical disaster precursor combination characteristics are identified.

[0016] In the aforementioned method for geological risk early warning and active support monitoring of pumped storage power stations, the risk feature fusion and early warning module in the data processing and early warning analysis platform 300 performs the following steps: Extracting single-dimensional risk features: From the output results of the stress field dynamic inversion module and the tomographic imaging inversion module, identify and extract single risk feature bodies related to surrounding rock instability, including identifying stress concentration areas from the four-dimensional absolute stress field model, identifying damage evolution areas from the elastic wave velocity time series variation model, identifying seepage enrichment areas from the resistivity time series variation model, and identifying microseismic active areas from microseismic event data acquired in the acoustic detection mode passive mode; Risk feature fusion and co-evolution analysis are performed to assess the temporal and spatial correlation of different risk features and construct a comprehensive risk index. The comprehensive risk index is a weighted sum of the normalized indicators of each individual risk feature. When multiple features simultaneously exhibit high index values ​​in the same region and their temporal evolution trends are synchronous, the comprehensive risk index increases significantly. A tiered early warning system is generated by comparing the comprehensive risk index with a set of preset thresholds. When the comprehensive risk index exceeds different thresholds, level one, level two, and level three early warnings are generated respectively. The early warning information includes the early warning level, the three-dimensional spatial coordinates of the risk area, the risk index, and the main risk characteristics that trigger the early warning. The system enables the visualization of early warning information, presenting the analysis and early warning results to users in a graphical manner. The extracted individual risk features are rendered as three-dimensional isosurfaces or volumes of different colors and superimposed on the three-dimensional geological model of the cavern group. At the same time, the comprehensive risk index is rendered in the form of a three-dimensional heat map. Users can interactively query the risk index and the evolution curve of each sub-indicator over time at any location.

[0017] The present invention has the following beneficial effects: 1. This invention establishes a quantitative relationship between stress changes and changes in physical properties such as resistivity and elastic wave velocity directly at the engineering site through an in-situ pressure-sensitive calibration module and the controllable mechanical loading process of anchor tension. This overcomes the limitation of traditional geophysical methods that can only make qualitative inferences, and enables the monitoring results to directly reflect the stress values ​​and distribution of the surrounding rock, providing a more accurate data basis for risk assessment.

[0018] 2. This invention uses a risk feature fusion and early warning module to simultaneously identify and analyze stress concentration zones, damage evolution zones, seepage enrichment zones, and microseismic active zones. By evaluating the coupling evolution relationship of these four types of risk features in time and space, this invention can jointly determine instability risks from multiple physical dimensions, effectively avoiding misjudgments or omissions that may be caused by monitoring a single indicator, making early warning decisions more comprehensive and robust.

[0019] 3. This invention integrates distributed fiber optic sensors into the multimodal intelligent anchoring unit, which can not only acquire geophysical information of the surrounding rock mass, but also monitor the axial stress distribution of the anchor rod itself in real time. It can directly evaluate the bearing capacity and load transfer efficiency of the anchor rod, and at the same time use the anchor rod itself as a mechanical benchmark for detecting changes in the stress of the surrounding rock, thus realizing the organic unity of support monitoring and geological environment monitoring. Attached Figure Description

[0020] Figure 1 This is a system architecture diagram of the present invention; Figure 2 This is a flowchart of the method of the present invention; Figure 3 This is a hardware architecture block diagram of the data acquisition and control host of the present invention.

[0021] Among them, 100 is the multimodal intelligent anchoring unit; 200 is the data acquisition and control host; 201 is the central processing and clock synchronization module; 202 is the multi-channel resistivity meter module; 203 is the acoustic transceiver control module; 204 is the fiber optic demodulator interface module; and 300 is the data processing and early warning analysis platform. Detailed Implementation

[0022] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0023] Please see the appendix Figure 1 - Appendix Figure 3 This invention provides a geological risk early warning and active support monitoring system for pumped storage power stations, the system comprising: Multiple multimodal intelligent anchoring units 100 deployed inside the rock mass, a data acquisition and control host 200 connected to the multimodal intelligent anchoring units 100, and a data processing and early warning analysis platform 300 communicatively connected to the data acquisition and control host 200.

[0024] The multimodal intelligent anchoring unit 100, as the field sensing and execution component of the system, provides support functions while integrating sensors for electrical and acoustic detection as well as sensors for measuring its own strain and temperature. Multiple multimodal intelligent anchoring units 100 form a three-dimensional detection array in the rock mass.

[0025] The data acquisition and control host 200, as the execution control and data acquisition center on site, drives the multimodal intelligent anchoring unit 100 array to perform specific detection tasks according to preset instructions or instructions from the data processing and early warning analysis platform 300, and is responsible for collecting the raw data returned by all sensors.

[0026] The data processing and early warning analysis platform 300, as the central processing unit of the system, is responsible for receiving and storing data from the data acquisition and control host 200, running the core algorithm of the method of this invention, and generating analysis results and early warning information.

[0027] Please see the appendix Figure 2 This invention provides a method for geological risk early warning and active support monitoring of pumped storage power stations, which may include the following steps: S101. Obtain the initial physical property benchmark model. Specifically, the steps are as follows: deploy multiple multimodal intelligent anchoring units in the surrounding rock to be monitored, perform a baseline scan to collect the original resistivity and elastic wave datasets at the initial moment, and construct a three-dimensional benchmark model characterizing the initial physical properties of the surrounding rock based on the original datasets.

[0028] S102. Calibrate the in-situ pressure sensitivity coefficient; the specific steps are as follows: during the application of tension load to at least one multimodal intelligent anchoring unit, simultaneously collect the strain data of the unit and the geophysical response data measured by the surrounding anchoring unit array, and solve and construct the in-situ pressure sensitivity coefficient field covering the monitoring area based on the stress disturbance and physical property changes caused by the tension load. S103. Obtain a time-series model of physical properties. Specifically, the steps are to periodically perform resistivity and elastic wave tomography scans and perform inversion calculations on the data collected in each period to generate a series of three-dimensional time-series models characterizing the evolution of the internal physical properties of the rock mass over time.

[0029] S104. Invert the stress field and perform fusion early warning; the specific steps are as follows: use the in-situ pressure sensitivity coefficient to interpret the time-series variation model of physical properties in order to invert the spatiotemporal evolution of the stress field inside the surrounding rock. Furthermore, it integrates multi-dimensional risk characteristics such as stress, damage, and seepage, and generates early warning information when typical disaster precursor combinations are identified.

[0030] The multimodal intelligent anchoring unit 100 in this embodiment of the invention is a composite functional body formed by functional integration on the basis of standard geotechnical support components. The base of the unit is a load-bearing structure, which can be the anchor rod or anchor cable commonly used in the underground cavern support project of pumped storage power station. The load-bearing structure, i.e. the rod base, is usually made of high-strength steel, and its primary function is to bear the design support load.

[0031] In one embodiment of the invention, a sensing element for electrical detection is integrated into the anchoring unit.

[0032] Specifically, the rod base itself serves as the first electrode for injecting current into the rock mass or measuring potential. In addition, the anchoring unit includes at least one accompanying electrode made of a flexible conductive material, such as graphene-based composite wire, carbon fiber bundle, or metal wire. During installation, this accompanying electrode is arranged parallel to the axial direction of the rod base. It maintains electrical insulation from the rod base through its own insulating sheath or through a pre-set insulating isolator. After the anchor borehole is filled and cured with grouting material, both the rod base and the accompanying electrode form an electrical path with the surrounding rock mass, thus collectively constituting two or more electrodes in a four-electrode measurement device that can be used for resistivity tomography. The arrangement of multiple mutually insulated accompanying electrodes supports more complex electrical measurement arrangements.

[0033] The multimodal intelligent anchoring unit 100 in this embodiment of the invention also integrates a sensing element for acoustic detection. Specifically, a piezoelectric transducer is encapsulated at the end of the anchoring unit or at a predetermined position along the length of the rod. The piezoelectric transducer is encapsulated and fixed by a waterproof and sealed shell with mechanical properties matching the surrounding rock to ensure its long-term reliable operation in underground high-pressure water environments. To ensure efficient transmission and reception of acoustic energy, an acoustic impedance matching layer may be filled between the transducer and the encapsulation shell. The signal lead of the piezoelectric transducer is laid along the length of the rod base and led out together with other cables to the orifice end of the anchoring unit.

[0034] In another embodiment of the invention, a sensing element for distributed self-sensing is also integrated into the anchoring unit. Specifically, a distributed optical fiber sensor is arranged along the axial direction of the pole base. The optical fiber sensor can be fixed in a pre-fabricated groove on the pole surface or directly integrated into the pole during manufacturing. To resist mechanical damage during installation and grouting, the optical fiber sensor may have a reinforcing protective layer. This distributed optical fiber sensor operates based on the principles of Brillouin scattering or Rayleigh scattering and is used to continuously measure the axial strain and temperature distribution along the entire length of the pole base.

[0035] In practice, the aforementioned accompanying electrodes, signal leads of the piezoelectric transducer, and distributed fiber optic sensors are integrated and fixed to the rod base before leaving the factory, forming an integral unit. After the unit is installed into the borehole, it is finally fixed by injecting grouting material. After the grouting material has cured, it not only transfers the load of the anchoring unit to the surrounding rock, but also acts as a physical field coupling medium, tightly coupling the entire multimodal intelligent anchoring unit 100 with the surrounding rock mass, forming a unified whole of mechanical and physical field response. This is the basis for all subsequent geophysical exploration and state interpretation.

[0036] The multimodal intelligent anchoring unit 100 in this embodiment of the invention realizes three different modes of operation: electrical detection, acoustic detection, and distributed self-sensing through its integrated different sensing elements.

[0037] The electrical detection mode of this unit is used to obtain information on the resistivity distribution of the rock mass. During resistivity tomography data acquisition, the data acquisition and control host 200 selects the electrodes of any two multimodal intelligent anchoring units in the array as the power supply electrode pair (A, B), and selects the electrodes of the other two units as the measurement electrode pair (M, N). The electrodes here can be the rod matrix or accompanying electrodes.

[0038] The host computer injects a stable current of known magnitude into the rock mass through the power supply electrodes. The current establishes a stable electric field in the rock mass, and simultaneously, the main unit measures the potential difference generated by this electric field on the measuring electrode pair. By systematically switching different electrode combinations as power supply and measurement electrode pairs, a dense potential difference dataset covering the entire detection area can be obtained. The resistivity of the rock mass is closely related to parameters such as water content, pore structure, and ion concentration; therefore, this dataset is fundamental for assessing the integrity and seepage state of the rock mass.

[0039] The acoustic detection mode of this unit is used to obtain information on the distribution of elastic wave velocity in rock mass or to monitor the micro-fracture activity of rock mass. This mode can operate in active mode or passive mode.

[0040] In active mode, the data acquisition and control host 200 applies a high-voltage electrical pulse to the piezoelectric transducer of a selected multimodal smart anchoring unit. The transducer converts electrical energy into mechanical vibration based on the inverse piezoelectric effect, thereby emitting a beam of elastic wave into the surrounding rock mass. The piezoelectric transducers of all other multimodal smart anchoring units in the array operate in receiving mode, converting the received mechanical vibration after propagation through the rock mass into electrical signals based on the direct piezoelectric effect.

[0041] The data acquisition and control host 200 accurately records the initial arrival travel time of elastic waves from transmission to reception. By sequentially exciting each transducer in the array as a source and receiving the data from all other transducers, a complete cross-hole travel time dataset can be constructed. The elastic wave velocity of the rock mass is mainly affected by its elastic modulus, density, and degree of fracture development. Therefore, this dataset is the basis for evaluating the mechanical properties and structural damage of the rock mass.

[0042] In passive mode, all piezoelectric transducers in the multimodal intelligent anchoring units of the array are in a continuous signal listening state. When micro-fractures occur inside the rock mass due to stress adjustment, transient elastic waves, i.e., micro-vibrations or acoustic emission signals, are released. After these signals are received by multiple transducers, the data acquisition and control host 200 can calculate the location of the micro-fracture event based on the time difference of the signal arriving at different receiving points using a positioning algorithm. Spatiotemporal statistics of the micro-fracture events can reveal the damage accumulation and stress release process inside the rock mass.

[0043] The distributed self-sensing mode of this unit is used to acquire the strain and temperature distribution of the anchoring unit itself. The data acquisition and control host 200, through a connected fiber optic demodulator, emits a laser pulse of a specific frequency to the distributed fiber optic sensors integrated in the unit. As the laser propagates in the fiber, Brillouin scattering or Rayleigh scattering occurs. By analyzing the frequency shift and intensity of the backscattered light, the fiber optic demodulator can calculate the strain and temperature values ​​at each point on the fiber.

[0044] Because the fiber optic sensor is tightly coupled to the rod base, the strain distribution it measures reflects the axial force distribution state borne by the rod base. This allows for a precise understanding of the transmission law of the support load from the anchor head to the deep rock mass, providing direct quantitative basis for assessing the health status of the anchoring system and the interaction between the rock mass and the support structure.

[0045] The data acquisition and control host 200 serves as the hub connecting the field sensing devices and the back-end analysis platform. It is designed with a modular hardware architecture to adapt to the needs of different detection modes and ensure the overall reliability of the system.

[0046] The core of the data acquisition and control host 200 is the central processing and clock synchronization module 201. This module includes an embedded industrial computer or programmable logic controller, which is responsible for executing preset control command sequences, managing data storage, and handling network communication with the data processing and early warning analysis platform 300.

[0047] The central processing and clock synchronization module 201 also integrates a high-precision clock source, such as a GPS synchronization clock or a temperature-controlled crystal oscillator, to provide a unified time reference for all data acquisition processes. This is crucial for subsequent synchronous analysis of multi-channel data and microseismic positioning.

[0048] To achieve electrical detection modes, the data acquisition and control host 200 includes a multi-channel resistivity meter module 202, which consists of a high-power constant current source, a high-precision voltmeter, and a large-scale relay switch matrix. The relay switch matrix is ​​connected to the electrical electrodes of all multi-modal intelligent anchoring units 100 deployed in the field. The central processing and clock synchronization module 201 controls the on / off state of the relay switch matrix to dynamically connect the constant current source and voltmeter to any four specified electrodes, thereby realizing automatic scanning and measurement of different four-electrode combinations.

[0049] To realize the acoustic detection mode, the data acquisition and control host 200 includes an acoustic transceiver control module 203. This module consists of a high-voltage pulse generator and a multi-channel high-speed data acquisition card. The high-voltage pulse generator generates the instantaneous high-voltage electrical signal required to drive the piezoelectric transducer to emit elastic waves. The multi-channel high-speed data acquisition card is responsible for synchronously amplifying, filtering, and converting the weak analog voltage signals output by all piezoelectric transducers acting as receivers. This module is also controlled by the central processing and clock synchronization module 201 to achieve the excitation of specific transducers and the synchronous acquisition of data from all channels.

[0050] To achieve distributed self-sensing modes, the data acquisition and control host 200 is equipped with a fiber optic demodulator interface module 204. This module provides standard data and control interfaces, such as Ethernet interfaces or serial ports, for commercial or custom Brillouin / Rayleigh fiber optic demodulators. The data acquisition and control host 200 sends acquisition commands to the fiber optic demodulator through this interface and receives strain and temperature data distributed along the fiber optic cable calculated by the demodulator.

[0051] In addition, the data acquisition and control host 200 is also equipped with an external trigger and synchronization interface 205. This interface is used to receive synchronization signals from external devices, such as analog signals or digital switch signals from pressure sensors on anchor tensioning equipment. When this interface receives a preset trigger signal, the central processing and clock synchronization module 201 immediately executes a predefined fast data acquisition sequence. This interface is the hardware foundation for implementing the in-situ pressure-sensitive self-calibration method of this invention.

[0052] The control logic and working mode of the data acquisition and control host 200 are the basis for realizing the monitoring and early warning method of the present invention. The host mainly operates in two working modes: normal periodic scanning mode and tension synchronous acquisition mode.

[0053] In the normalized periodic scanning mode, the data acquisition and control host 200 automatically executes routine monitoring tasks according to instructions issued by the data processing and early warning analysis platform 300 or an internally preset schedule. A complete routine monitoring cycle may include the following steps: During resistivity tomography scanning, the central processing and clock synchronization module 201, based on a predefined measurement sequence file, sequentially configures the power supply electrode pairs and measurement electrode pairs by controlling the relay switch matrix of the multi-channel resistivity meter module 202. For each configuration, the module performs a power-on-measurement operation and caches the recorded current and potential difference values ​​along with the corresponding electrode combination information. This process iterates through all electrode combinations in the sequence file.

[0054] Elastic wave tomography scanning is performed. After the resistivity scan is completed, the central processing and clock synchronization module 201 switches to the acoustic transceiver control module 203. This module, according to a preset excitation sequence, selects one piezoelectric transceiver in the array as the vibration source in turn and drives a high-voltage pulse generator to excite it. Simultaneously, the high-speed data acquisition card records the waveform data of all other piezoelectric transceivers acting as receivers within a preset time window. This process traverses all preset vibration sources.

[0055] Collect distributed fiber optic sensing data. The central processing and clock synchronization module 201 sends acquisition commands to the fiber optic demodulator through the fiber optic demodulator interface module 204, and receives the returned strain and temperature data distributed along all connected distributed fiber optic sensors.

[0056] After completing all the above acquisition steps, the central processing and clock synchronization module 201 associates all resistivity data, elastic wave waveform data and distributed optical fiber data obtained within the monitoring period with a unified timestamp provided by a high-precision clock source, packages them into a data file, and uploads it to the data processing and early warning analysis platform 300 through the network communication interface.

[0057] The tensioning synchronous acquisition mode is a special event-triggered operating mode used to achieve in-situ pressure sensitivity self-calibration. The control logic for this mode is as follows: Listening for external trigger signals. The central processing and clock synchronization module 201 of the data acquisition and control host 200 continuously monitors the status of the external trigger and synchronization interface module 205. This interface is connected to the pressure sensor or operating switch of the anchor tensioning equipment.

[0058] When a high-speed acquisition sequence is triggered, and a predefined trigger signal is detected on the interface, such as when the pressure value exceeds the start threshold or the tensioning operation switch is closed, the host immediately suspends the ongoing regular scanning task and enters the tensioning synchronous acquisition mode.

[0059] In this mode, the host performs resistivity and elastic wave tomography scans continuously and rapidly at a frequency much higher than that of conventional scans to capture transient changes in the physical properties of the rock mass during tension loading.

[0060] Simultaneously recording time-varying data, while performing high-speed tomographic scanning, the host computer acquires distributed fiber optic sensing data on the multimodal intelligent anchoring unit 100 being tensioned at high frequency through the fiber optic demodulator interface module 204, so as to obtain the complete dynamic process of tension force transmission along the rod.

[0061] A synchronization event dataset is generated. When the tensioning end signal is detected on the external trigger and synchronization interface module 205, the host stops high-speed data acquisition.

[0062] Subsequently, it packages all high-frequency tomographic imaging data and corresponding distributed strain data collected throughout the tensioning process into an independent synchronous event dataset with start and end timestamps, and uploads it to the data processing and early warning analysis platform 300 for subsequent pressure sensitivity coefficient calibration calculation. After completing this operation, the host returns to the normalized periodic scanning mode.

[0063] The data preprocessing and tomographic imaging inversion module in the data processing and early warning analysis platform 300 is responsible for converting the raw dataset received from the data acquisition and control host 200 into a three-dimensional physical property model that can characterize the internal structure of the surrounding rock.

[0064] The module first preprocesses the received raw resistivity dataset. Preprocessing steps include data quality checks, such as removing data points with abnormal measurements due to poor electrode grounding or on-site electromagnetic interference, and performing reciprocity checks, i.e., swapping the power supply electrode pair with the measurement electrode pair and performing repeated measurements to assess and filter data reliability. After quality checks are completed, the module calculates the apparent resistivity for each measurement combination according to the following formula. : ; in, This is a device coefficient, the value of which depends only on the three-dimensional spatial coordinates of the power supply electrode (A,B) and the measuring electrode (M,N); To measure the potential difference recorded on the electrode pair; This refers to the current injected into the power supply electrode pair.

[0065] For the raw elastic wave dataset, the preprocessing steps of this module include digital filtering of the acquired raw waveform data, such as bandpass filtering, to suppress noise and enhance the signal-to-noise ratio of the first arrival signal. Then, the module employs an automatic picking algorithm, such as one based on the energy ratio of short and long time windows, to identify and extract the first arrival travel times of the P-wave or S-wave from the filtered waveform. Perform a consistency check on all picked-up travel time data and remove outliers that significantly deviate from the normal propagation time.

[0066] After preprocessing, this module performs tomographic inversion calculations. Mathematically, the inversion process is structured as solving a regularized optimization problem, with the goal of finding a model parameter vector that best fits the observed data and satisfies geophysical prior constraints. The general objective function of this optimization problem. It can be represented as: ; in, For data fitting terms; For model regularization terms; This is the vector of observation data obtained after preprocessing; Forward operands; and These are the data weighting matrix and the model constraint matrix, respectively. This is the regularization factor.

[0067] In resistivity tomography inversion, the model parameter vector To monitor the logarithmic resistivity of each discrete unit in the region, the observation data vector The logarithmic apparent resistivity for all measurement combinations. Forward modeling operator. The potential distribution under a given resistivity model is calculated by solving the three-dimensional Poisson equation. This optimization problem can be solved using iterative methods such as the Gauss-Newton method. In each iteration, the model update amount This is obtained by solving the following system of linear equations: ; in, In the current model The Jacobian matrix is ​​calculated at the given location. The solution is iterated until the model converges, ultimately yielding the three-dimensional resistivity distribution model.

[0068] In elastic wave velocity tomography inversion, the model parameter vector The slowness of each discrete unit is represented by the observation data vector. This represents the time taken. The relationship between time taken and slowness can be approximated by a system of linear equations. ,in It is a geometric matrix whose elements Indicates the first The first ray The length traversed within each cell is such that, since the ray path itself depends on an unknown velocity field, this problem employs an iterative solution strategy: In each iteration, ray tracing is first performed based on the current velocity model to update the matrix. Then, the regularized optimization problem described above is solved to update the slowness model. The solution to this system of linear equations can be achieved using methods such as simultaneous iterative reconstruction. Iteration continues until the model converges, ultimately yielding the three-dimensional slowness distribution, which is then converted into a wave velocity distribution model.

[0069] The in-situ pressure sensitivity calibration module in the data processing and early warning analysis platform 300 processes synchronous event datasets collected during anchor tensioning to establish a quantitative relationship between changes in rock mass physical properties and stress changes. This quantitative relationship is characterized by a spatially varying pressure sensitivity coefficient field, which forms the physical basis for subsequent quantitative inversion of the surrounding rock stress field. The execution process of this module may include the following steps: Calculate the stress disturbance field. The goal of this step is to calculate the incremental stress field generated by the anchor bolt in the surrounding rock mass based on the anchor bolt tension load. The module first processes the distributed fiber optic sensing data synchronously acquired on this stressed anchoring unit to obtain the stress increment field along the length of the bolt base. Distributed axial strain Based on this strain distribution, the shear stress distributed along the contact surface between the rod and the grout can be calculated. : ; in, Let be the elastic modulus of the rod matrix. The cross-sectional area of ​​the rod base is... Given the borehole diameter, this shear stress distribution represents the load distribution applied by the anchor bolt to the surrounding rock. The module then discretizes this load distribution into a series of concentrated forces acting on the anchor bolt axis and, based on elasticity theory, calculates these forces at any point in the rock mass. The resulting stress increment. This calculation can be performed using numerical methods or analytical / semi-analytical methods.

[0070] In one specific embodiment, this calculation is achieved by integrating the Mindlin solution to obtain a three-dimensional stress increment tensor field distributed throughout the monitoring area caused by a single tensioning event. For rock materials, changes in their physical properties are usually related to changes in effective stress; therefore, the change in effective stress can be extracted from the stress increment tensor. For example, the change in average principal stress or the change in maximum principal stress.

[0071] The goal of this step is to calculate the physical property change field based on the tomographic imaging data collected before and after tensioning, and to invert the physical property change field caused by stress disturbance. The module selects one of the industry-known time series inversion or differential imaging methods to perform this calculation.

[0072] In one embodiment, the module inverts the data collected before tensioning to obtain a baseline physical property model, and inverts the data collected during tensioning to obtain a physical property model under loading conditions. Subsequently, the physical property change field is obtained by subtracting the models. and .

[0073] In another embodiment, to obtain more stable results, the module can employ a difference inversion method to directly solve for the solutions that can explain the differences between the two periods of data. Model change This process involves solving... Implementation, in which This is the sensitivity matrix calculated based on the benchmark model.

[0074] Solving for the pressure sensitivity coefficient field yields the stress increment field. and the corresponding physical property change field and Then, based on the approximate linear relationship between stress and physical properties in rock physics, the module solves for the pressure sensitivity coefficient point-by-point or element-by-element in space. Wave velocity pressure sensitivity coefficient. and resistivity varistor coefficient The calculation formula is as follows: ; ; in, It is a spatial position vector; and These are the elastic wave velocity and resistivity at that point before tensioning, respectively; and These represent the changes in wave velocity and resistivity at that point caused by tensioning, respectively. This represents the effective stress change experienced at that point. By tensioning and calibrating multiple multimodal intelligent anchoring units 100 at different locations, and then spatially interpolating or fusing the calculation results, a spatially non-uniform pressure-sensitive coefficient field covering the entire three-dimensional monitoring area can be constructed. and .

[0075] The stress field dynamic inversion module in the data processing and early warning analysis platform 300 functions by interpreting the time-series model of physical properties obtained during routine monitoring using the quantitative relationships established by the in-situ pressure-sensitive calibration module, thereby achieving quantitative inversion of the dynamic changes in the stress field inside the surrounding rock. The execution process of this module may include the following steps: Calculates the relative change field of physical properties. This module receives a series of three-dimensional physical property models arranged in chronological order from the data preprocessing and tomographic inversion modules. ,in Represents elastic wave velocity or resistivity , It is a spatial position vector. For the first At each monitoring point, this module selects the physical property model of the system under its initial steady-state condition. As a baseline, calculate any subsequent time point Relative variation field of physical properties relative to this reference : ; Through this calculation, the relative change field of elastic wave velocity and the relative change field of resistivity can be obtained respectively.

[0076] The module inverts the incremental stress field using a calibrated, spatially varying wave velocity pressure sensitivity coefficient field. and resistivity piezoresistive coefficient field The calculated relative change field of physical properties is then converted into an effective stress increment field caused by geological activity or engineering disturbance. In the case of a single physical field, this inversion can be directly calculated using the following formula: ; in, To correspond to the pressure sensitivity coefficient of the physical property, in a specific embodiment, to improve the stability and reliability of the inversion results, the module adopts a joint inversion strategy, utilizing the variation information of both wave velocity and resistivity. This joint inversion is achieved by solving an optimization problem, namely, finding a stress increment field. This allows it to optimally explain the observed changes in both physical properties simultaneously. The objective function of this optimization problem can be set as: ; in, To balance the weighting factors contributed by the two data sources, an optimization problem is solved to obtain a more reliable three-dimensional stress increment field that integrates multiphysics information. .

[0077] The module updates the absolute stress field by comparing the inverted stress increment field with the known initial geostress field. By superimposing the data, any monitoring time can be obtained. absolute stress field of surrounding rock Initial geostress field It can be obtained through well-known techniques in the field, such as in-situ in-situ stress testing or numerical simulation based on geological models. The updated formula for the absolute stress field is as follows: ; The module ultimately outputs a four-dimensional absolute stress field model of the surrounding rock, enabling dynamic and quantitative monitoring of the stress state of the surrounding rock.

[0078] The risk feature fusion and early warning module in the Data Processing and Early Warning Analysis Platform 300 performs in-depth analysis of multi-source data and inversion models to identify precursors of geological disasters and generate early warning information. This module achieves its function by extracting multi-dimensional risk features and analyzing their spatiotemporal coupling relationships, specifically including the following steps: This module extracts single-dimensional risk features from the outputs of the stress field dynamic inversion module and the tomographic imaging inversion module, identifying and extracting single risk features related to surrounding rock instability. Specifically: From the four-dimensional absolute stress field model In this process, by comparing with the rock mass mechanical strength criteria, areas where the stress level is close to or exceeds the rock mass strength are identified and marked as stress concentration areas.

[0079] From the elastic wave velocity time series variation model Within the rock mass, a region exhibiting a sustained and significant decrease in wave velocity was identified. This decrease in wave velocity is directly related to the expansion and penetration of microfractures within the rock mass; therefore, this region was designated as the damage evolution zone.

[0080] From the resistivity time-series variation model In the study, regions with a sustained and significant decrease in resistivity were identified. In underground engineering environments, a decrease in resistivity usually indicates the enrichment of pore water or the formation of dominant seepage channels, and therefore these regions were identified as seepage enrichment zones.

[0081] From the microseismic event data acquired in the passive mode of acoustic detection, spatiotemporal clustering analysis of the microseismic events was performed to identify areas where microseismic events were highly concentrated in time and space, and these areas were identified as microseismic active zones.

[0082] The risk feature fusion and co-evolution analysis is performed. The goal of this step is to assess the temporal and spatial correlation of different risk features. The module determines the degree of risk coupling by analyzing the degree of overlap and temporal synchronous evolution trend of the extracted stress concentration area, damage evolution area, seepage enrichment area and microseismic active area in three-dimensional space.

[0083] In one specific embodiment, the module constructs a comprehensive risk index. The calculation method involves a weighted summation of the normalized indices for each individual risk characteristic: ; in, These are normalization functions representing the intensity of stress, damage, seepage, and microseismic activity, respectively, with values ​​between 0 and 1; These are the corresponding weighting coefficients, whose values ​​can be determined based on engineering geological conditions, expert experience, or machine learning training based on historical data. When multiple features simultaneously exhibit high index values ​​in the same area, and their temporal evolution trends are synchronous, the comprehensive risk index... It will increase significantly.

[0084] The module generates tiered early warning information and calculates a comprehensive risk index. It compares the data with a set of preset thresholds to achieve tiered alerts; for example, three thresholds can be set. : when Exceed At that time, the system generates a Level 1 warning, indicating that potential risks have emerged in the area.

[0085] when Exceed At that time, the system generates a level-two warning, indicating that the risk characteristics are showing obvious co-evolution and the possibility of instability is increasing.

[0086] when Exceed At that time, the system generates a three-level warning, indicating that a combination of typical disaster precursor characteristics has been formed and the risk of instability is extremely high. The warning information includes the warning level, the three-dimensional spatial coordinates of the risk area, the risk index, and the main risk characteristics that triggered the warning.

[0087] This module enables the visualization of early warning information. It presents the analysis and early warning results to users in a graphical manner. The visualization interface can overlay the extracted individual risk features as three-dimensional isosurfaces or volumes of different colors onto the three-dimensional geological model of the cavern complex. At the same time, it displays the comprehensive risk index. High-risk areas are visually marked using a 3D heatmap. Users can interactively query the risk index and the evolution curves of each sub-indicator over time at any location to trace the evolution of risk.

[0088] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A geological risk early warning and active support monitoring system for pumped storage power stations, characterized in that, include: Multiple multimodal intelligent anchoring units (100) deployed inside the rock mass provide support functions while integrating sensors for electrical and acoustic detection as well as sensors for measuring their own strain and temperature. The multiple multimodal intelligent anchoring units (100) form a three-dimensional detection array in the rock mass. The data acquisition and control host (200) connected to the multimodal intelligent anchoring unit (100) drives the multimodal intelligent anchoring unit (100) array to perform specific detection tasks and collect raw data returned by all sensors according to preset instructions or instructions from the data processing and early warning analysis platform (300); The data processing and early warning analysis platform (300) is communicatively connected to the data acquisition and control host (200). The data processing and early warning analysis platform (300) receives and stores data from the data acquisition and control host (200), runs the core algorithm, and generates analysis results and early warning information.

2. The pumped storage power station geological risk early warning and active support monitoring system according to claim 1, characterized in that, The base of the multimodal intelligent anchoring unit (100) is an anchor rod or anchor cable commonly used in the underground cavern support project of pumped storage power station. It is made of high-strength steel and is used to bear the design support load.

3. The pumped storage power station geological risk early warning and active support monitoring system according to claim 2, characterized in that, The multimodal intelligent anchoring unit (100) integrates sensing elements for electrical detection. The rod base serves as the first electrode. The anchoring unit also includes at least one accompanying electrode made of flexible conductive material. The accompanying electrode is arranged parallel to the axial direction of the rod base and is electrically insulated from the rod base by its own insulating sheath or a preset insulating isolator. After the anchor borehole is filled and cured by grouting material, the rod base and the accompanying electrode form an electrical path with the surrounding rock mass, constituting two or more electrodes in a four-electrode measurement device for resistivity tomography.

4. The pumped storage power station geological risk early warning and active support monitoring system according to claim 2, characterized in that, The multimodal intelligent anchoring unit (100) integrates a sensing element for acoustic detection. A piezoelectric transducer is encapsulated at the end of the anchoring unit or at a predetermined position along the length of the rod. The piezoelectric transducer is encapsulated and fixed by a waterproof and sealed shell that matches the mechanical properties of the surrounding rock. An acoustic impedance matching layer is filled between the transducer and the encapsulation shell. The signal lead of the piezoelectric transducer is laid along the length of the rod base and led out together with other cables to the orifice end of the anchoring unit.

5. The pumped storage power station geological risk early warning and active support monitoring system according to claim 2, characterized in that, The multimodal intelligent anchoring unit (100) integrates a sensing element for distributed self-sensing. A distributed optical fiber sensor is arranged along the axial direction of the pole base. The optical fiber sensor is fixed in a pre-made groove on the surface of the pole or directly integrated into it during the pole manufacturing process. The optical fiber sensor has an enhanced protective layer and works based on the Brillouin scattering or Rayleigh scattering principle. It is used to continuously measure the axial strain and temperature distribution over the entire length of the pole base.

6. The pumped storage power station geological risk early warning and active support monitoring system according to claim 1, characterized in that, The data acquisition and control host (200) includes: The central processing and clock synchronization module (201) includes an embedded industrial computer or a programmable logic controller, which is responsible for executing a preset control instruction sequence, managing data storage, processing network communication with the data processing and early warning analysis platform (300), and integrating a high-precision clock source to provide a unified time reference for all data acquisition processes; The multi-channel resistivity meter module (202) consists of a high-power constant current source, a high-precision voltmeter, and a large-scale relay switch matrix. The relay switch matrix is ​​connected to the electrical electrodes of all the multi-modal intelligent anchoring units (100). The central processing and clock synchronization module (201) controls the on / off state of the relay switch matrix to dynamically connect the constant current source and the voltmeter to four designated electrodes, thereby realizing automatic scanning and measurement of different four-electrode device combinations. The acoustic transceiver control module (203) consists of a high-voltage pulse generator and a multi-channel high-speed data acquisition card. The high-voltage pulse generator is used to generate the instantaneous high-voltage electrical signal required to drive the piezoelectric transducer to emit elastic waves. The multi-channel high-speed data acquisition card is responsible for synchronously amplifying, filtering and converting the weak analog voltage signal output by the piezoelectric transducer, which acts as the receiver. The module is controlled by the central processing and clock synchronization module (201) to realize the excitation of a specific transducer and the synchronous acquisition of data from all channels. The fiber optic demodulator interface module (204) provides a standard data and control interface for commercial or customized Brillouin / Rayleigh fiber optic demodulators. The data acquisition and control host (200) sends acquisition commands to the fiber optic demodulator through this interface and receives strain and temperature data distributed along the fiber optic cable calculated by the demodulator. An external trigger and synchronization interface is used to receive synchronization signals from external devices. When a preset trigger signal is received, the central processing and clock synchronization module (201) immediately executes a predefined fast data acquisition sequence.

7. The pumped storage power station geological risk early warning and active support monitoring system according to claim 1, characterized in that, The data acquisition and control host (200) operates in a normalized periodic scanning mode, including the following steps: When performing resistivity tomography scanning, the central processing and clock synchronization module (201) controls the relay switch matrix of the multi-channel resistivity meter module (202) according to the predefined measurement sequence file, sequentially configures the power supply electrode pair and the measurement electrode pair, performs a power-on-measurement operation for each configuration, and caches the recorded current value, potential difference value and corresponding electrode combination information together, and traverses all electrode combinations in the sequence file. After performing elastic wave tomography scanning and completing the resistivity scan, the central processing and clock synchronization module (201) switches to the acoustic wave transceiver control module (203). The module selects one piezoelectric transceiver in the array as the source of vibration according to the preset excitation sequence and drives the high-voltage pulse generator to excite it. At the same time, the high-speed data acquisition card records the waveform data of other piezoelectric transceivers as receivers within the preset time window and traverses all preset sources of vibration. The central processing and clock synchronization module (201) collects distributed fiber optic sensing data. Through the fiber optic demodulator interface module (204), it sends a collection command to the fiber optic demodulator and receives the returned strain and temperature data distributed along all connected distributed fiber optic sensors. After completing the above acquisition steps, the central processing and clock synchronization module (201) associates all resistivity data, elastic wave waveform data and distributed optical fiber data obtained within the monitoring period with the unified timestamp provided by the high-precision clock source, packages them into a data file, and uploads it to the data processing and early warning analysis platform (300) through the network communication interface.

8. The pumped storage power station geological risk early warning and active support monitoring system according to claim 1, characterized in that, The data acquisition and control host (200) operates in tension synchronous acquisition mode, including the following steps: Listening to external trigger signals, the central processing and clock synchronization module (201) continuously monitors the status of the external trigger and synchronization interface; Triggering a high-speed acquisition sequence: When a predefined trigger signal is detected on the interface, the host pauses the regular scanning task and enters the tension synchronous acquisition mode to continuously and rapidly perform resistivity and elastic wave tomography scans at a frequency higher than that of regular scanning. Simultaneously recording time-varying data, while performing high-speed tomographic scanning, the host computer acquires distributed optical fiber sensing data on the multimodal intelligent anchoring unit (100) being tensioned at high frequency through the optical fiber demodulator interface module (204). A synchronous event dataset is generated. When the tensioning end signal is detected on the external trigger and synchronization interface, the host stops high-speed acquisition and packages all high-frequency tomographic imaging data and corresponding distributed strain data acquired during the entire tensioning process into a synchronous event dataset with start and end timestamps. The dataset is then uploaded to the data processing and early warning analysis platform (300) and then returns to the normalized periodic scanning mode.

9. A method for geological risk early warning and active support monitoring of pumped storage power stations, applied to the system described in any one of claims 1-8, characterized in that, Includes the following steps: A baseline model of initial physical properties is obtained, multiple multimodal intelligent anchoring units (100) are deployed in the surrounding rock to be monitored, a baseline scan is performed to collect the original data set of resistivity and elastic wave at the initial moment, and a three-dimensional baseline model characterizing the initial physical properties of the surrounding rock is constructed based on the original data set. In calibrating the in-situ pressure sensitivity coefficient, during the process of applying tension load to at least one multimodal smart anchor unit (100), the strain data of the unit and the geophysical response data measured by the surrounding anchor unit array are collected simultaneously, and the in-situ pressure sensitivity coefficient field covering the monitoring area is solved and constructed based on the stress disturbance and physical property changes caused by the tension load. The physical properties time-series variation model is obtained, resistivity and elastic wave tomography scans are periodically performed, and the data collected in each period are inverted to generate a series of three-dimensional time-series models characterizing the evolution of the internal physical properties of the rock mass over time. The stress field is inverted and integrated for early warning. The in-situ pressure sensitivity coefficient is used to interpret the time-series variation model of the physical properties to invert the spatiotemporal evolution of the stress field inside the surrounding rock. Furthermore, multi-dimensional risk characteristics such as stress, damage, and seepage are integrated, and early warning information is generated when typical disaster precursor combination characteristics are identified.

10. The method for geological risk early warning and active support monitoring of pumped storage power stations according to claim 9, characterized in that, The risk feature fusion and early warning module in the data processing and early warning analysis platform (300) performs the following steps: Extracting single-dimensional risk features: From the output results of the stress field dynamic inversion module and the tomographic imaging inversion module, identify and extract single risk feature bodies related to surrounding rock instability, including identifying stress concentration areas from the four-dimensional absolute stress field model, identifying damage evolution areas from the elastic wave velocity time series variation model, identifying seepage enrichment areas from the resistivity time series variation model, and identifying microseismic active areas from microseismic event data acquired in the acoustic detection mode passive mode; Risk feature fusion and co-evolution analysis are performed to assess the temporal and spatial correlation of different risk features and construct a comprehensive risk index. The comprehensive risk index is a weighted sum of the normalized indicators of each individual risk feature. When multiple features simultaneously exhibit high index values ​​in the same region and their temporal evolution trends are synchronous, the comprehensive risk index increases significantly. A tiered early warning system is generated by comparing the comprehensive risk index with a set of preset thresholds. When the comprehensive risk index exceeds different thresholds, level one, level two, and level three early warnings are generated respectively. The early warning information includes the early warning level, the three-dimensional spatial coordinates of the risk area, the risk index, and the main risk characteristics that trigger the early warning. The system enables the visualization of early warning information, presenting the analysis and early warning results to users in a graphical manner. The extracted individual risk features are rendered as three-dimensional isosurfaces or volumes of different colors and superimposed on the three-dimensional geological model of the cavern group. At the same time, the comprehensive risk index is rendered in the form of a three-dimensional heat map. Users can interactively query the risk index and the evolution curve of each sub-indicator over time at any location.

Images