Rock hydraulic-dynamic coupling catastrophe multi-domain index testing device

Abstract

The invention relates to the technical field of testing devices, in particular to a rock hydraulic-dynamic coupling catastrophe multi-domain index testing device. According to the technical scheme, the device comprises a dynamic loading system, a hydraulic loading system, a rock mass sample clamping and sealing system, a multi-domain parameter monitoring system and an intelligent data analysis and processing system. The dynamic loading system is matched with the hydraulic loading system to simulate a complex hydraulic-dynamic coupling process of the rock mass in a natural environment and in engineering activities, and diversified rock mass engineering research requirements are met; according to the invention, the multi-domain parameter indexes are synchronously monitored through the multi-domain parameter monitoring system, so that a comprehensive and accurate catastrophe model can be constructed; the intelligent data analysis and processing system can timely and accurately identify rock mass catastrophe features and invert high-precision model parameters, a reliable model basis is provided for deep research of a rock mass hydraulic-dynamic coupling catastrophe mechanism, and subsequent more accurate prediction of the catastrophe development trend is also facilitated.

Description

Technical Field

[0001] The invention relates to the technical field of testing devices, and in particular to a rock mass hydraulic-dynamic coupling disaster multi-domain index testing device. Background Art

[0002] In many rock engineering fields, such as dam foundations and underground caverns in water conservancy and hydropower projects, surrounding rock of tunnels in mining, and deep buried tunnels in transportation projects, rock masses are often under the interaction of complex stress, seepage and dynamic environment. As engineering construction expands to deep areas with complex geological conditions, the problem of hydraulic-dynamic coupling disasters in rock masses has become increasingly prominent.

[0003] In terms of simulating dynamic loading, most of the existing rock mass hydraulic-dynamic coupling disaster test devices have fixed dynamic loading modes and single frequencies, which makes it difficult to simulate the spectrum characteristics of real earthquakes or complex engineering vibrations. The hydraulic loading system is not accurate and flexible enough, making it difficult to accurately reproduce complex and changeable seepage boundary conditions, and unable to accurately control the dynamic changes of pore water pressure over time and space; moreover, the monitoring of various key parameters is often not synchronized, and the data acquisition frequency and time base of different types of sensors are inconsistent, making it impossible to effectively correlate and integrate the acquired multi-domain indicator data, making it difficult to construct a complete and accurate rock mass hydraulic-dynamic coupling disaster model, and unable to meet the urgent needs of current complex rock mass engineering for disaster prediction and prevention. Therefore, we propose a rock mass hydraulic-dynamic coupling disaster multi-domain indicator test device. Summary of the invention

[0004] The purpose of the present invention is to address the problems existing in the background technology and to propose a rock mass hydraulic-dynamic coupling disaster multi-domain index testing device.

[0005] The technical solution of the present invention is: a rock mass hydraulic-dynamic coupling disaster multi-domain index testing device, including a dynamic loading system, a hydraulic loading system, a rock mass sample clamping and sealing system, a multi-domain parameter monitoring system and an intelligent data analysis and processing system;

[0006] The dynamic loading system is used to apply dynamic load to the rock sample to simulate the dynamic load condition;

[0007] The hydraulic loading system is used to provide adjustable water pressure to simulate seepage conditions;

[0008] The rock sample clamping and sealing system is used to fix and seal the rock sample;

[0009] The multi-domain parameter monitoring system is used to perform real-time synchronous monitoring of the mechanics, seepage, and thermodynamics indicators of rock samples;

[0010] The intelligent data analysis and processing system includes a disaster feature extraction model based on deep learning, and the disaster feature extraction model analyzes and processes the monitoring data, extracts disaster features and inverts model parameters.

[0011] Optionally, the power loading system includes an electro-hydraulic servo actuator and an adaptive feedback control algorithm, the electro-hydraulic servo actuator is used to apply a dynamic load to the rock sample, and the adaptive feedback control algorithm is used to track the deviation between the output of the electro-hydraulic servo actuator and a preset loading curve in real time, and adjust the flow and pressure of the hydraulic oil through a PID controller.

[0012] Optionally, the adaptive feedback control algorithm formula is as follows:

[0013]

[0014] Wherein, u(t) is the control signal, e(t) is the deviation signal, and the deviation signal is the difference between the preset value and the actual value, K p , K i , K d They are proportional, integral and differential coefficients respectively;

[0015] The adaptive feedback control algorithm introduces a dynamic compensation algorithm to optimize the accuracy of the power loading of the electro-hydraulic servo actuator, and the dynamic model is constructed as follows:

[0016]

[0017] Where M is the mass of the actuator moving parts, C is the damping coefficient, K is the stiffness coefficient, x is the displacement, are acceleration and velocity respectively, F(t) is the loading force, and the loading force is dynamically compensated using the dynamic model by real-time monitoring of displacement, velocity and acceleration.

[0018] Optionally, the hydraulic loading system includes a constant pressure pump, a flow regulating valve, a pressure sensor, a precision diverter and a connecting pipe, the constant pressure pump is the power source for providing the water pressure, the output end of the constant pressure pump is connected to the flow regulating valve, the output end of the flow regulating valve is connected to the input end of the precision diverter, and the pressure sensor is fixedly installed on the connecting pipe between the constant pressure pump and the precision diverter.

[0019] Optionally, the hydraulic loading system further includes a fuzzy logic control algorithm, which combines Darcy's law with a non-Darcy seepage correction formula to adjust the water pressure and flow rate. Darcy's law is used at low flow rates. The Darcy's law formula is as follows:

[0020]

[0021] A correction formula is used at high flow rates, and the correction formula is as follows:

[0022]

[0023] Where v is the seepage velocity, k is the permeability coefficient, μ is the fluid dynamic viscosity, is the pressure gradient, ρ is the fluid density, β is the fluid compressibility, is the rate of change of pressure with time.

[0024] Optionally, the rock sample clamping and sealing system includes a hydraulic jack, an alloy steel clamping frame and a rubber sealing pad. The hydraulic jack is used to provide a clamping force for the rock sample, and the influence of the clamping force on the mechanical properties of the rock sample is quantified by the Hertz contact theory. The Hertz contact theory formula is expressed as follows:

[0025]

[0026] Where F is the clamping force, E * is the equivalent elastic modulus, R is the equivalent radius of curvature of the contact body, and δ is the contact deformation.

[0027] Optionally, the multi-domain parameter monitoring system includes a high-speed data acquisition card, the sampling frequency of the high-speed data acquisition card is not less than 10kHz, the high-speed data acquisition card is built with a synchronous clock signal generator for ensuring synchronous acquisition of data from each sensor, and the multi-domain parameter monitoring system includes a mechanical parameter monitoring component, a seepage parameter monitoring component and a thermodynamic parameter monitoring component;

[0028] The mechanical parameter monitoring component includes a strain gauge, a fiber Bragg grating sensor and a laser displacement meter. The strain gauge is attached to the surface and the inside of the rock sample. The measurement accuracy of the strain gauge is ±1με. The fiber Bragg grating sensor is used to monitor the stress and strain distribution of the rock sample. The laser displacement meter is used to detect the macro deformation of the rock sample. The displacement measurement accuracy of the laser displacement meter is ±0.01mm.

[0029] The seepage parameter monitoring component includes a thermal conductivity flowmeter and a pore water pressure sensor, wherein the thermal conductivity flowmeter is used to measure the seepage volume of each channel, and the pore water pressure sensor is used to detect the seepage posture of the rock sample, and the pressure measurement accuracy of the pore water pressure sensor is ±0.001MPa;

[0030] The thermodynamic parameter monitoring component includes a thermocouple, and the data reliability is verified by a data fusion algorithm based on the Kalman filter principle, and a state equation and an observation equation are established. The state equation is as follows:

[0031] x k =Axk-1 +w k-1

[0032] The observation equation is as follows:

[0033] z k =Hx k +v k

[0034] Among them, x k is the system state vector, A is the state transfer matrix, w k-1 is the process noise, z k is the observation vector, H is the observation matrix, v k is the observation noise.

[0035] Optionally, the disaster feature extraction model based on deep learning in the intelligent data analysis and processing system includes a convolutional neural network model, and the convolutional neural network model includes an input layer, a convolutional layer, a pooling layer and a fully connected layer. The input layer is used to receive the time series image of the multi-domain synchronous monitoring data monitored by the multi-domain parameter monitoring system, and the disaster features are extracted through the convolutional layer, the pooling layer and the fully connected layer. The model training adopts a generative adversarial network, and the optimized model loss function is:

[0036] L=L CNN -λL GAN

[0037] Among them, L CNN is the original loss of the CNN model, L GAN is the loss of the GAN model, and λ is the balance coefficient.

[0038] Optionally, the intelligent data analysis and processing system further includes a genetic algorithm, which combines finite element numerical simulation to invert rock mass coupling model parameters, encodes rock mass mechanical parameters and seepage parameters into genetic algorithm individual genes, and continuously optimizes individuals through selection, crossover, and mutation operations. The fitness function is defined as:

[0039]

[0040] in, are the simulated value and measured value of the i-th monitoring point respectively, n is the total number of monitoring points, and after multiple iterations, the model parameters are inverted.

[0041] Optionally, the genetic algorithm is combined with a simulated annealing algorithm to optimize the genetic algorithm search process. After the genetic algorithm iterates a certain number of rounds, an inferior solution is accepted with a certain probability. The simulated annealing acceptance criterion formula is expressed as:

[0042]

[0043] Among them, P is the probability of accepting an inferior solution, ΔE is the energy difference between the new solution and the current optimal solution, and T is the simulated annealing temperature, which controls the probability of accepting an inferior solution.

[0044] In summary, the present application includes at least one of the following beneficial technical effects:

[0045] 1. The present invention outputs loads of various waveforms, frequencies and amplitudes through a dynamic loading system, and cooperates with a hydraulic loading system that can simulate complex seepage paths to study the complex hydraulic-dynamic coupling process of rock mass in natural environments and engineering activities, meeting the diverse needs of rock mass engineering research;

[0046] 2. The present invention uses a multi-domain parameter monitoring system to synchronously monitor mechanical, seepage, and thermodynamic parameter indicators, ensuring that data from different types of sensors can be collected synchronously in real time, fully reflecting the instantaneous interaction between various physical quantities in the coupling process, and helping to build a comprehensive and accurate rock mass hydraulic-dynamic coupling disaster model;

[0047] 3. The disaster feature extraction model based on deep learning in the intelligent data analysis and processing system of the present invention can timely and accurately identify the key features of impending disasters in the rock mass, buy time for disaster prevention and control, and use genetic algorithms combined with finite element numerical simulation and simulated annealing algorithm optimized parameter inversion methods to invert high-precision model parameters based on monitoring data, providing a reliable model basis for in-depth research on the hydraulic-dynamic coupling disaster mechanism of the rock mass, and also helping to more accurately predict the development trend of the disaster in the future. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 A structural block diagram of a rock mass hydraulic-dynamic coupling disaster multi-domain index testing device of the present invention is given. DETAILED DESCRIPTION

[0049] The technical solution of the present invention will be clearly and completely described below in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all of the embodiments.

[0050] The components of the embodiments of the present invention generally described and shown in the drawings herein may be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present invention provided in the drawings is not intended to limit the scope of the claimed invention, but merely represents selected embodiments of the present invention.

[0051] Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in the field without making any creative work shall fall within the scope of protection of the present invention.

[0052] In the description of the present invention, it should be noted that the terms "center", "upper", "lower", "left", "right", "vertical", "horizontal", "inner", "outer", etc., indicating the orientation or positional relationship, are based on the orientation or positional relationship shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the present invention. In addition, the terms "first", "second", and "third" are used for descriptive purposes only, and cannot be understood as indicating or implying relative importance.

[0053] It should be noted that the terms "include", "comprises" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner.

[0054] In the description of the present invention, it should be noted that, unless otherwise clearly specified and limited, the terms "installed", "connected", and "connected" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0055] Example

[0056] The present invention proposes a rock mass hydraulic-dynamic coupling disaster multi-domain index testing device, such as Figure 1 As shown, it includes a power loading system, a hydraulic loading system, a rock sample clamping and sealing system, a multi-domain parameter monitoring system and an intelligent data analysis and processing system;

[0057] Among them, the dynamic loading system is used to apply dynamic loads to rock samples to simulate dynamic load conditions; the hydraulic loading system is used to provide adjustable water pressure to simulate seepage conditions; the rock sample clamping and sealing system is used to fix and seal the rock samples; the multi-domain parameter monitoring system is used to conduct real-time and synchronous monitoring of multi-domain indicators of rock samples, including mechanics, seepage, and thermodynamics; the intelligent data analysis and processing system includes a disaster feature extraction model based on deep learning; the disaster feature extraction model analyzes and processes the monitoring data, extracts disaster features, and inverts model parameters.

[0058] The dynamic loading system includes an electro-hydraulic servo actuator and an adaptive feedback control algorithm. The electro-hydraulic servo actuator is used to apply dynamic load to the rock sample. The adaptive feedback control algorithm is used to track the deviation between the output of the electro-hydraulic servo actuator and the preset loading curve in real time, and adjust the flow and pressure of the hydraulic oil through a PID controller.

[0059] The adaptive feedback control algorithm formula is as follows:

[0060]

[0061] Among them, u(t) is the control signal, e(t) is the deviation signal, the deviation signal is the difference between the preset value and the actual value, K p , K i , K d They are proportional, integral, and differential coefficients respectively. The adaptive feedback control algorithm introduces a dynamic compensation algorithm to optimize the accuracy of the power loading of the electro-hydraulic servo actuator. The dynamic model is constructed as follows:

[0062]

[0063] Where M is the mass of the actuator moving parts, C is the damping coefficient, K is the stiffness coefficient, x is the displacement, are acceleration and velocity respectively, and F(t) is the loading force. By real-time monitoring of displacement, velocity and acceleration, the loading force is dynamically compensated using the dynamic model.

[0064] The hydraulic loading system includes a hydraulic loading execution and monitoring component, which includes a constant pressure pump, a flow regulating valve, a pressure sensor, a precision diverter and a connecting pipe. The constant pressure pump is the power source for providing the water pressure. The output end of the constant pressure pump is connected to the flow regulating valve, and the output end of the flow regulating valve is connected to the input end of the precision diverter. The pressure sensor is fixedly installed on the connecting pipe between the constant pressure pump and the precision diverter. The hydraulic loading system also includes a fuzzy logic control algorithm. The fuzzy logic control algorithm combines Darcy's law and non-Darcy seepage correction formula to adjust the water pressure and flow. Darcy's law is used at low flow rates. The Darcy's law formula is as follows:

[0065]

[0066] At high flow rates, a correction formula is used, which is as follows:

[0067]

[0068] Where v is the seepage velocity, k is the permeability coefficient, μ is the fluid dynamic viscosity, is the pressure gradient, ρ is the fluid density, β is the fluid compressibility, is the rate of change of pressure with time.

[0069] The present invention outputs loads of various waveforms, frequencies and amplitudes through a dynamic loading system, and cooperates with a hydraulic loading system that can simulate complex seepage paths to study the complex hydraulic-dynamic coupling process of rock masses in natural environments and engineering activities, thus meeting the diverse needs of rock engineering research.

[0070] The rock sample clamping and sealing system includes a hydraulic jack, an alloy steel clamping frame and a rubber sealing pad. The hydraulic jack is used to provide clamping force for the rock sample, and the Hertz contact theory is used to quantify the influence of the clamping force on the mechanical properties of the rock sample. The Hertz contact theory formula is expressed as follows:

[0071]

[0072] Where F is the clamping force, E * is the equivalent elastic modulus, R is the equivalent radius of curvature of the contact body, and δ is the contact deformation.

[0073] The multi-domain parameter monitoring system includes a high-speed data acquisition card, the sampling frequency of the high-speed data acquisition card is not less than 10kHz, and the high-speed data acquisition card is equipped with a synchronous clock signal generator for ensuring synchronous acquisition of data from each sensor. The multi-domain parameter monitoring system includes a mechanical parameter monitoring component, a seepage parameter monitoring component, and a thermodynamic parameter monitoring component;

[0074] The mechanical parameter monitoring components include strain gauges, fiber Bragg grating sensors and laser displacement meters. The strain gauges are pasted on the surface and inside of the rock sample. The measurement accuracy of the strain gauges is ±1με. The fiber Bragg grating sensors are used to monitor the stress and strain distribution of the rock sample. The laser displacement meter is used to detect the macroscopic deformation of the rock sample. The displacement measurement accuracy of the laser displacement meter is ±0.01mm.

[0075] The seepage parameter monitoring components include a thermal conductivity flowmeter and a pore water pressure sensor. The thermal conductivity flowmeter is used to measure the seepage volume of each channel, and the pore water pressure sensor is used to detect the seepage posture of the rock sample. The pressure measurement accuracy of the pore water pressure sensor is ±0.001MPa.

[0076] The thermodynamic parameter monitoring component includes thermocouples, and the data fusion algorithm based on the Kalman filter principle verifies the data reliability, establishes the state equation and observation equation, and the state equation is as follows:

[0077] x k =Ax k-1 +w k-1

[0078] The observation equation is as follows:

[0079] z k =Hx k +v k

[0080] Among them, x k is the system state vector, A is the state transfer matrix, w k-1 is the process noise, z k is the observation vector, H is the observation matrix, v k is the observation noise.

[0081] The present invention uses a multi-domain parameter monitoring system to synchronously monitor mechanical, seepage, and thermodynamic parameter indicators, ensuring that data from different types of sensors can be collected synchronously in real time, fully reflecting the instantaneous interaction between various physical quantities in the coupling process, and helping to build a more comprehensive and accurate rock hydraulic-dynamic coupling disaster model.

[0082] The disaster feature extraction model based on deep learning in the intelligent data analysis and processing system includes a convolutional neural network model. The convolutional neural network model includes an input layer, a convolutional layer, a pooling layer, and a fully connected layer. The input layer is used to receive the time series image of the multi-domain synchronous monitoring data monitored by the multi-domain parameter monitoring system. The disaster features are extracted through the convolutional layer, the pooling layer, and the fully connected layer. The model training adopts the adversarial generation network. The optimized model loss function is:

[0083] L=L CNN -λL GAN

[0084] Among them, L CNN is the original loss of the CNN model, L GAN is the loss of the GAN model, and λ is the balance coefficient;

[0085] The intelligent data analysis and processing system also includes a genetic algorithm. The genetic algorithm combines finite element numerical simulation to invert the rock mass coupling model parameters, encodes the rock mass mechanical parameters and seepage parameters into genetic algorithm individual genes, and continuously optimizes individuals through selection, crossover, and mutation operations. The fitness function is defined as:

[0086]

[0087] in, are the simulated value and measured value of the i-th monitoring point, respectively, and n is the total number of monitoring points. After multiple iterations, the model parameters are inverted. The genetic algorithm is combined with the simulated annealing algorithm to optimize the genetic algorithm search process. After a certain number of iterations of the genetic algorithm, the inferior solution is accepted with a certain probability. The simulated annealing acceptance criterion formula is expressed as:

[0088]

[0089] Among them, P is the probability of accepting an inferior solution, ΔE is the energy difference between the new solution and the current optimal solution, and T is the simulated annealing temperature, which controls the probability of accepting an inferior solution.

[0090] The disaster feature extraction model based on deep learning in the intelligent data analysis and processing system of the present invention can autonomously learn the data feature patterns of different disaster stages, and can timely and accurately identify the key features of impending disasters in the rock mass, thereby buying more time for disaster prevention and control. It also uses a genetic algorithm combined with finite element numerical simulation and a parameter inversion method optimized by a simulated annealing algorithm to invert high-precision model parameters based on monitoring data, providing a reliable model basis for in-depth research on the hydraulic-dynamic coupling disaster mechanism of the rock mass, and also helping to more accurately predict the development trend of the disaster in the future.

[0091] The testing process using the above rock mass hydraulic-dynamic coupling disaster multi-domain index testing device includes the following steps:

[0092] 1. Device Construction

[0093] Power loading system: electro-hydraulic servo actuator is selected, the loading frequency is set to 30-80Hz, the loading force peak is 600kN, and the coefficients in the adaptive feedback control algorithm are determined through repeated debugging. Set K p =0.8, K i =0.05, K d =0.02, the electro-hydraulic servo actuator outputs random wave load;

[0094] Hydraulic loading system: The constant pressure pump provides 0-10MPa water pressure in stable operation. The flow control valve adjusts the opening according to the fuzzy logic algorithm. When simulating slow seepage, Darcy's law is used, combined with the feedback of the pressure sensor, to finely control the flow rate monitored by the thermal conductivity flowmeter to within ±0.05mL / min of the preset value. In case of strong water inrush simulation, the non-Darcy seepage correction formula is used to ensure that the water pressure and flow rate are accurately matched to the complex seepage. The precision diverter is designed with 4 diversion channels according to the crack direction of the tunnel surrounding rock to simulate the real seepage path;

[0095] Rock sample clamping and sealing system: The selected alloy steel frame has a side length of 400mm, which is suitable for standard rock samples. Four hydraulic jacks are symmetrically distributed around the alloy steel frame, each with a clamping force of 300kN, ensuring that the clamping force applied to the sample is uniform and stable. The leakage rate of the rubber sealing gasket is less than 0.01L / h. The Hertz contact theory is used for quantitative analysis, and micro strain gauges are arranged in the contact area between the rock sample and the clamping surface to monitor the contact deformation in real time. According to the Hertz formula, the clamping force distribution is reversed, and the loading strategy of the hydraulic jack is optimized and adjusted accordingly to further ensure that the sample is stable and sealed during the entire test process;

[0096] Multi-domain parameter monitoring system: high-speed acquisition card synchronously collects data at 10kHz, 15 strain gauges, 8 fiber Bragg grating sensors and laser displacement meters with an accuracy of ±0.01mm are pasted on the surface and key internal parts of the rock sample, the accuracy of the thermal conductivity flowmeter is ±0.05mL / min, and the thermal conductivity flowmeter is installed at the key nodes of each branch pipeline of the precision diverter. The accuracy of the pore water pressure sensor is ±0.001MPa. The pore water pressure sensor is distributed at different positions around the rock sample at specific intervals and angles according to the requirements of seepage field simulation. The two work together to control the rock seepage parameters in real time and accurately. The accuracy of the thermocouple is ±0.5℃. According to the temperature field monitoring point layout plan optimized and designed by numerical simulation in advance, a total of 12 are arranged inside and on the surface of the sample to accurately measure the temperature change of the sample during the dynamic-hydraulic coupling process, providing data support for exploring the heat generation and transfer laws during the disaster process;

[0097] 2. Testing Process

[0098] Preparation stage: Select rock samples with properties highly similar to those of the tunnel surrounding rock mass. After transporting the samples to the laboratory, use CNC processing equipment to process them into standard sizes, ensuring that the side length error is controlled within ±1mm and the surface flatness error is less than ±0.1mm;

[0099] Install the processed rock samples in the clamping and sealing system, and complete the connection and tightening of each component in turn according to the pre-established installation process. After the installation is completed, check each system of the device, including line connection, sensor calibration and algorithm parameter setting. After confirmation, start the device and preheat the key components. The preheating time is set to 30-60 minutes according to the characteristics of the components to ensure that each component reaches the best working condition.

[0100] Steady-state simulation stage: turn on the hydraulic loading system to provide water pressure and flow loading for the specimen. At this time, the dynamic loading system remains static and does not apply any dynamic load;

[0101] The multi-domain parameter monitoring system is started simultaneously to collect mechanical, seepage, and thermodynamic parameter data in real time at a high-frequency sampling rate of 10kHz, and the data is preliminarily processed and stored in the database in real time as benchmark data for subsequent comparative analysis. At this stage, the monitoring duration is set to 2-3 hours to ensure that sufficient steady-state data is collected to fully reflect the physical properties of the sample under normal groundwater seepage environment.

[0102] Disaster simulation stage: Start the dynamic loading system to simulate earthquakes of different intensities, and carry out loading tests in sequence according to the equivalent loads of Richter 5.0, 6.0, and 7.0 earthquakes. During the loading process, the dynamic loading system outputs dynamic loads of corresponding waveforms, frequencies, and amplitudes according to the preset loading curves;

[0103] At the same time, the hydraulic loading system cooperates with the dynamic loading, and according to the groundwater level fluctuation caused by the earthquake and the prediction model of the seepage velocity change, the fuzzy logic control algorithm is combined with the corresponding seepage formula to dynamically adjust the water pressure and flow rate;

[0104] During the entire coupled disaster process, the multi-domain parameter monitoring system continuously and synchronously collects data, and the intelligent data analysis and processing system analyzes the data in real time online. The disaster feature extraction model based on deep learning is used to capture the disaster features, which include changes in rock mass cracks, sudden changes in pore water pressure, and stress and strain concentration. Once the disaster features are detected to exceed the preset warning conditions, the system will issue a warning signal and record the warning time, the corresponding parameter status, and the data change trend, providing information for subsequent in-depth analysis of the disaster process.

[0105] Ending stage: After completing all the scheduled test items, stop the power loading system and the hydraulic loading system in turn. According to the operating procedures, slowly shut down the power supply, valves and other equipment of each system to ensure that the device smoothly exits the working state. Classify and archive the data collected in this test, use data mining and analysis software to analyze the data, and use the model parameters obtained by inversion to further optimize the numerical model. By comparing and verifying with the field measured data, improve the accuracy of the model, and lay a solid technical foundation for the design, construction and operation decisions of the tunnel project.

[0106] 3. Effect Verification Details

[0107] Comparison between simulation and measured data: At the tunnel construction site, in-situ monitoring equipment, including high-precision strain gauges, pore water pressure sensors, seismic accelerometers, etc., were arranged in advance along the key parts of the planned route to continuously collect stress, seepage, and dynamic response parameter data of the rock mass in a natural environment. The rock mass parameters obtained by the device test were compared and analyzed with the measured data on site, and the degree of consistency between the two was calculated using statistical software.

[0108] Early warning effect evaluation: In multiple simulation tests, the early warning effect of the intelligent data analysis and processing system is evaluated to analyze the accuracy of the early warning;

[0109] Parameter inversion verification: Use the model parameters obtained by inversion to perform numerical simulation and make detailed fit calculations with the test results.

[0110] The above specific embodiments are only several optional embodiments of the present invention. Based on the technical solutions of the present invention and the relevant inspirations of the above embodiments, those skilled in the art can make various alternative improvements and combinations to the above specific embodiments.

Claims

1. A rock mass hydraulic-dynamic coupling disaster multi-domain index testing device, characterized in that: include: A dynamic loading system, which is used to apply a dynamic load to the rock sample to simulate a dynamic load condition; A hydraulic loading system, the hydraulic loading system is used to provide adjustable water pressure to simulate seepage conditions; A rock sample clamping and sealing system, wherein the rock sample clamping and sealing system is used to fix and seal the rock sample; A multi-domain parameter monitoring system, which is used to perform real-time synchronous monitoring of the mechanics, seepage, and thermodynamics indicators of rock samples; An intelligent data analysis and processing system, wherein the intelligent data analysis and processing system comprises a disaster feature extraction model based on deep learning, wherein the disaster feature extraction model analyzes and processes the monitoring data, extracts disaster features and inverts model parameters.

2. A rock mass hydraulic-dynamic coupling disaster multi-domain index testing device according to claim 1, characterized in that: The power loading system includes an electro-hydraulic servo actuator and an adaptive feedback control algorithm. The electro-hydraulic servo actuator is used to apply a dynamic load to the rock sample. The adaptive feedback control algorithm is used to track the deviation between the output of the electro-hydraulic servo actuator and a preset loading curve in real time, and adjust the flow and pressure of the hydraulic oil through a PID controller.

3. A rock mass hydraulic-dynamic coupling disaster multi-domain index testing device according to claim 2, characterized in that: The adaptive feedback control algorithm formula is as follows: Wherein, u(t) is the control signal, e(t) is the deviation signal, and the deviation signal is the difference between the preset value and the actual value, K p , K i , K d They are proportional, integral and differential coefficients respectively; The adaptive feedback control algorithm introduces a dynamic compensation algorithm to optimize the accuracy of the power loading of the electro-hydraulic servo actuator, and the dynamic model is constructed as follows: Where M is the mass of the actuator moving parts, C is the damping coefficient, K is the stiffness coefficient, x is the displacement, are acceleration and velocity respectively, F(t) is the loading force, and the loading force is dynamically compensated using the dynamic model by real-time monitoring of displacement, velocity and acceleration.

4. The rock mass hydraulic-dynamic coupling disaster multi-domain index testing device according to claim 1 is characterized in that: The hydraulic loading system includes a hydraulic loading execution and monitoring component, which includes a constant pressure pump, a flow regulating valve, a pressure sensor, a precision diverter and a connecting pipe. The constant pressure pump is the power source for providing the water pressure. The output end of the constant pressure pump is connected to the flow regulating valve, and the output end of the flow regulating valve is connected to the input end of the precision diverter. The pressure sensor is fixedly installed on the connecting pipe between the constant pressure pump and the precision diverter.

5. The rock mass hydraulic-dynamic coupling disaster multi-domain index testing device according to claim 4 is characterized in that: The hydraulic loading system also includes a fuzzy logic control algorithm, which combines Darcy's law with a non-Darcy seepage correction formula to adjust the water pressure and flow rate. Darcy's law is used at low flow rates. The Darcy's law formula is as follows: A correction formula is used at high flow rates, and the correction formula is as follows: Where v is the seepage velocity, k is the permeability coefficient, μ is the fluid dynamic viscosity, is the pressure gradient, ρ is the fluid density, β is the fluid compressibility, is the rate of change of pressure with time.

6. The rock mass hydraulic-dynamic coupling disaster multi-domain index testing device according to claim 1 is characterized in that: The rock sample clamping and sealing system includes a hydraulic jack, an alloy steel clamping frame and a rubber sealing pad. The hydraulic jack is used to provide a clamping force for the rock sample, and the Hertz contact theory is used to quantify the influence of the clamping force on the mechanical properties of the rock sample. The Hertz contact theory formula is expressed as follows: Where F is the clamping force, E * is the equivalent elastic modulus, R is the equivalent radius of curvature of the contact body, and δ is the contact deformation.

7. The rock mass hydraulic-dynamic coupling disaster multi-domain index testing device according to claim 1 is characterized in that: The multi-domain parameter monitoring system includes a high-speed data acquisition card, the sampling frequency of the high-speed data acquisition card is not less than 10kHz, the high-speed data acquisition card is built with a synchronous clock signal generator for ensuring synchronous acquisition of data from each sensor, and the multi-domain parameter monitoring system includes a mechanical parameter monitoring component, a seepage parameter monitoring component and a thermodynamic parameter monitoring component; The mechanical parameter monitoring component includes a strain gauge, a fiber Bragg grating sensor and a laser displacement meter. The strain gauge is attached to the surface and the inside of the rock sample. The measurement accuracy of the strain gauge is ±1με. The fiber Bragg grating sensor is used to monitor the stress and strain distribution of the rock sample. The laser displacement meter is used to detect the macro deformation of the rock sample. The displacement measurement accuracy of the laser displacement meter is ±0.01mm. The seepage parameter monitoring component includes a thermal conductivity flowmeter and a pore water pressure sensor, wherein the thermal conductivity flowmeter is used to measure the seepage volume of each channel, and the pore water pressure sensor is used to detect the seepage posture of the rock sample, and the pressure measurement accuracy of the pore water pressure sensor is ±0.001MPa; The thermodynamic parameter monitoring component includes a thermocouple, and the data reliability is verified by a data fusion algorithm based on the Kalman filter principle, and a state equation and an observation equation are established. The state equation is as follows: x k =Ax k-1 +w k-1 The observation equation is as follows: z k =Hx k +v k Among them, x k is the system state vector, A is the state transfer matrix, w k-1 is the process noise, z k is the observation vector, H is the observation matrix, v k is the observation noise.

8. The rock mass hydraulic-dynamic coupling disaster multi-domain index testing device according to claim 1 is characterized in that: The disaster feature extraction model based on deep learning in the intelligent data analysis and processing system includes a convolutional neural network model, which includes an input layer, a convolutional layer, a pooling layer and a fully connected layer. The input layer is used to receive the time series image of the multi-domain synchronous monitoring data monitored by the multi-domain parameter monitoring system, and extract the disaster features through the convolutional layer, the pooling layer and the fully connected layer. The model training adopts the adversarial generation network, and the optimized model loss function is: L=L CNN -λL GAN Among them, L CNN is the original loss of the CNN model, L GAN is the loss of the GAN model, and λ is the balance coefficient.

9. The rock mass hydraulic-dynamic coupling disaster multi-domain index testing device according to claim 8, characterized in that: The intelligent data analysis and processing system also includes a genetic algorithm, which combines finite element numerical simulation to invert rock mass coupling model parameters, encodes rock mass mechanical parameters and seepage parameters into genetic algorithm individual genes, and continuously optimizes individuals through selection, crossover, and mutation operations. The fitness function is defined as: in, are the simulated value and measured value of the i-th monitoring point respectively, n is the total number of monitoring points, and after multiple iterations, the model parameters are inverted.

10. A rock mass hydraulic-dynamic coupling disaster multi-domain index testing device according to claim 9, characterized in that: The genetic algorithm is combined with the simulated annealing algorithm to optimize the genetic algorithm search process. After the genetic algorithm iterates a certain number of rounds, the inferior solution is accepted with a certain probability. The simulated annealing acceptance criterion formula is expressed as: Among them, P is the probability of accepting an inferior solution, ΔE is the energy difference between the new solution and the current optimal solution, and T is the simulated annealing temperature, which controls the probability of accepting an inferior solution.

Images