Roadway surrounding rock anisotropic aging deformation monitoring and predicting model under mining disturbance

Abstract

The invention discloses a roadway surrounding rock anisotropic aging deformation monitoring and predicting model under mining disturbance. A method comprises the following steps: step 1, preparing an in-situ sample; step 2, constructing a sample mounting cavity; step 3, sample installation and monitoring system arrangement; 4, in-situ test environment construction and in-situ stress recovery are carried out; step 5, data acquisition; step 6, constructing an aging deformation model; and step 7, constructing an aging deformation prediction model. According to the method, the anisotropic sample is installed in the deep underground roadway on site, an automatic monitoring system is combined, roadway surrounding rock stress-strain characteristics under mining-induced stress changes are collected in real time, the aging deformation prediction model is established, and a scientific basis can be provided for safe mining of an underground mine.

Description

Technical Field

[0001] The present invention relates to the fields of deep underground rock mechanics, deep geological engineering and intelligent monitoring and prediction technology, and in particular to the time-dependent deformation monitoring and prediction of deep anisotropic tunnel surrounding rocks under mining disturbance conditions, and constructs a dynamic early warning model by using multi-sensor data fusion and deep reinforcement learning. Background Art

[0002] The mining of deep coal resources faces complex geological conditions and strong mining disturbances. In particular, the surrounding rock of deep tunnels has significant anisotropy characteristics, and its mechanical behavior shows obvious differences in different directions. During the mining process, the dynamic changes in mining stress will cause the tunnel surrounding rock to undergo time-dependent deformation, that is, deformation that changes with time. This deformation may cause disasters such as tunnel instability and rock burst, which seriously threatens mine safety production.

[0003] At present, most of the research on the time-dependent deformation of tunnel surrounding rock relies on indoor tests or numerical simulation methods. However, it is often difficult for indoor tests to truly reproduce the dynamic stress path under mining disturbance, and the sample size effect may lead to distortion of anisotropic characteristics; in terms of numerical simulation, due to the use of idealized constitutive models, it is also difficult to fully reflect the complex behavior of tunnel surrounding rock under real mining conditions, such as damage, seepage and stress. In addition, traditional monitoring technologies such as single-point displacement meters and anchor stress meters can only obtain local and unidirectional data, making it difficult to achieve spatial continuity of anisotropic time-dependent deformation of tunnel surrounding rock.

[0004] Therefore, it is urgent to develop a field monitoring and prediction method that can synchronously obtain the time-dependent deformation evolution data of anisotropic tunnel surrounding rock under real mining disturbance conditions, and integrate the micro-crack characteristics and macro-mechanical response, so as to provide a scientific basis for in-depth revelation of the damage mechanism of tunnel surrounding rock, realize real-time dynamic monitoring, and accurately predict future deformation trends, thus providing strong technical support for safe mining in mines. Summary of the invention

[0005] The present invention aims to provide a monitoring and prediction model for the anisotropic time-dependent deformation of tunnel surrounding rocks under mining disturbance. By installing anisotropic specimens on site in a coal mine and combining them with an automated monitoring system, the stress-strain characteristics of tunnel surrounding rocks under mining stress changes are collected in real time, and a time-dependent deformation prediction model is established to provide a scientific basis for safe mining in the mine.

[0006] To achieve the above purpose, the technical solution adopted by the present invention is as follows: The beneficial effects of the present invention are as follows: The anisotropic time-dependent deformation monitoring and prediction model of tunnel surrounding rock under mining disturbance includes the following steps: Step 1, in-situ sample preparation, select the tunnel surrounding rock that is not disturbed by mining in the advanced support pressure area of ​​the coal mining face, obtain coal samples by coring, cut along the horizontal bedding 0°, vertical bedding 90°, and inclined bedding 45° of the coring coal samples, and obtain three kinds of samples. The three samples have the same size and are rectangular; Step 2: constructing a sample installation cavity, constructing an installation cavity in a surrounding rock area of ​​the mining tunnel that is not disturbed by mining in front of the coal mining face, wherein the size of the installation cavity is larger than the sample; Step 3: Sample installation and monitoring system arrangement: Place the sample in the installation cavity, arrange a hydraulic loading device at the open end of the installation cavity, support the sample and the five side walls of the installation cavity through the hydraulic loading device, and the hydraulic loading device is supported on the end surface of the sample, finally forming a three-axis support; The loading end of the hydraulic loading device is provided with a stress sensing device, a fiber Bragg grating stress sensor is embedded in the stress sensing device, a strain gauge is arranged on the telescopic arm of the hydraulic loading device, a temperature sensor and a humidity sensor are arranged in the installation cavity, the stress sensor, the strain gauge, the temperature sensor and the humidity sensor are all connected to a data acquisition device, and the data acquisition device is connected to a ground control center; The hydraulic loading device is controlled by a ground control center; Step 4: In-situ test environment construction and in-situ stress recovery: According to the field exploration data, the target value of the in-situ ground stress of the sample is determined, which is spatially represented as three directions: x, y, and z. A preload is applied to the sample through a hydraulic loading device to simulate the in-situ stress state; Step 5: Data collection: Under the disturbance of mining, real-time data from stress sensors, strain gauges, temperature sensors, and humidity sensors are collected to correct thermal stress effects and other environmental factors during the accelerated test; Step 6: construct a time-dependent deformation model based on the data obtained in step 5; Step 7: Construct a time-dependent deformation prediction model. Based on the data obtained in step 5, construct a deep reinforcement learning prediction model to predict the strain increment in the future period.

[0007] Preferably, after step 7, the following steps are further included: Step 8: Apply the time-dependent deformation prediction model to mine tunnel support design, coal mining face advancement speed optimization, and rock burst prevention and control.

[0008] Preferably, in step 1: The sampling location was selected at a depth of 100m inside the advanced support pressure zone of the coal mining face, and the coal core was drilled using a drill rod with a diameter of 200mm. The dimensions of the sample are 50mm × 50mm × 100mm in width × length × height, with the width being in the x direction, the length being in the y direction, and the height being in the z direction. The dimensional error of the sample in each direction is no more than ±0.5mm.

[0009] Preferably, in step 2: There are at least 9 installation cavities in the same mining tunnel, and at least 3 of each type of specimen are installed; The size of the installation cavity is based on the size of the specimen with a safety margin added, and the depth, width, and height of the installation cavity satisfy the following formulas: in, L chamber , W chamber , H chamber They are the depth, width and height of the installation cavity respectively; L sample , W sample , H sample are the length, width and height of the specimen respectively, d For installation allowance, L max After determining the width and height, the depth direction is determined by the achievable safe excavation depth. L max Not less than 1 m; The specimen installation cavity closest to the coal mining face is at least 400 m away from the coal mining face.

[0010] Preferably, step 3 includes constructing a preload force feedback system, implementing regulation by real-time monitoring of the force sensors on each hydraulic loading system, and timely adjusting the parameters of the hydraulic device to control the loading error within 5% to ensure balanced preload in each axis direction. The regulation method is as follows; Set the sampling frequency of the stress sensor to T s , at each sampling moment k Collect the current record actual load force F actual [ k ], the target preload force is recorded as F target , then the error can be calculated e [ k ]: (2) Adopt closed-loop control algorithm based on PID control to generate control instructions u [ k] is used to adjust the hydraulic device parameters in time to compensate for errors: (3) in, K p is the proportional gain, which controls the amplification of the instantaneous error; K i is the integral gain, which accumulates past errors to eliminate steady-state deviations; K d The differential gain is used to predict the future error trend to suppress oscillation. To cope with the nonlinearity and uncertainty of the on-site environment, the preload force feedback system introduces an adaptive correction algorithm. The system compares the real-time collected loading force data with the target preload force, calculates the error, and generates a control signal through the PID controller. u [ k ], this signal drives the hydraulic actuator to adjust the pressure parameters so that the loading force gradually approaches the target value: (4)

[0011] Preferably, step 4 includes fine-tuning control, and the fine-tuning control is as follows: When the error is small, a higher sampling frequency is used to capture the subtle changes in stress error and calculate: (5) According to the online error and change trend, the PID parameters are updated using the adaptive algorithm: (6) in s target,x , s target,y and s target,z is the target stress value of the specimen in the x, y, and z directions, s x , s y、 s z is the actual stress value of the specimen in the x, y, and z directions, α p , α i and α d is the adaptive factor, f [ k ] is the system dynamic prediction function.

[0012] Preferably, in step 5, the data collection method and the signal processing of each sensor are as follows: Assume the real signal of the sensor is Sj ( t ), the measurement noise is N j ( t ), assuming the total number of sensors is N , then the continuous output signal of the sensor can be described as: (7) The sampling period of all sensors is unified as T s , k Used to represent the sampling point index during discrete sampling, that is, k sampling points, then the discrete sampling time t k The definition can be defined as: (8) At each sampling moment, data is collected from each sensor to obtain discrete sampling data: (9) In order to ensure the synchronization of multi-channel acquisition, the timestamp of each channel is required to meet the following requirements: (10) x The permissible deviation is extremely small.

[0013] After that, the data collected by each channel is t k Composition of data vector: (11) To eliminate high-frequency noise, firstly, the data vector X[ k ] to perform low-pass filtering and set the filter impulse response to h [ n ], through the convolution operation, the high-frequency noise is suppressed, so that the signal data after filtering is obtained as follows: (12) in, L is the filter length. Discrete wavelet transform is performed on the data after low-pass filtering. Discrete wavelet transform decomposes the signal into different scales and time positions. W i is the wavelet coefficient, which reflects the frequency component of the signal in the local area. l The wavelet coefficients are soft-thresholded and the criterion is: (13) parameter l Depending on the noise level and number of samples n Determine as follows: (14) By inverse wavelet transform Reconstruct the signal and get the denoised signal as , perform inverse wavelet transform on the processed wavelet coefficients, and get the denoised signal as follows: (15) In order to further improve data smoothness and estimation accuracy, the Kalman filter is used to estimate the state of the denoised signal, where A is the state transfer matrix: (16) Error covariance prediction: (17) in, Q is the process noise covariance.

[0014] set up H is the observation matrix, R To measure the noise covariance, the Kalman gain is calculated as: (18) Then the updated state is estimated to be: (19) The updated error covariance is: (20) In order to reduce the transmission bandwidth requirement, the denoised signal is compressed after preprocessing, and the transform coding method is used for data compression. Then the data is converted into a bit stream through adaptive quantization and entropy coding. The compression function is recorded as Q , then the final preprocessing output vector data is: (twenty one) In summary, the overall operation model of the preprocessing module is: (twenty two) The preprocessed data vector Y[k] is stored in the data acquisition unit and then transmitted in the form of a data packet. The data packet is defined as: (twenty three) The entire data stream is represented as a continuous data set D: (twenty four) For multi-channel configurations, each unit m Generate a data set D m, including their respective timestamps and preprocessed data, set the m The data set of units is recorded as: (25) The data integration system performs time alignment and fusion of data from each unit to form an overall data set, providing a unified data source for global monitoring and subsequent analysis: (26) When building data transmission, choose a low-latency transmission network to cooperate with the construction of the data transmission model, and the data return delay is less than 50 ms: (27) Construct the final return data at the ground control center, Indicates that the ground side is at time t The data received at the time is in the form of: (28).

[0015] Preferably, step 6, constructing a time-dependent deformation model comprises: Construct the cumulative strain as follows: (29) in, e ij 0 is the initial strain, is the creep strain rate, A ij To reflect the material constants of the basic creep characteristics in different directions, s ij ( t ) is the corresponding component of the stress tensor, expressed as s x ( t ), s y ( t ) s z ( t ); n ij is the stress sensitivity index; O ij is the activation energy, J is the gas constant, T C ( t ) is temperature; g ( RH ( t )) is the humidity modulation function, D ij ( t ) is the anisotropic damage variable,m ij is the damage inhibition parameter; The sample is ij The creep strain rate in the direction is as follows: (30) in, B ij is the damage rate constant; p , q , r A sensitivity index that describes the effects of stress and strain on damage accumulation; O D,ij is the damage activation energy; h ( RH ( t )) is the humidity modulation function.

[0016] Preferably, in step 7, the data set is synthesized based on the downhole transmission data , build a deep reinforcement learning prediction model, which consists of the following modules: First of all Perform convolution operations to extract local spatial features: (31) Then, LSTM is used to process the time series features output by the convolutional layer to capture the dynamic evolution of time-dependent deformation: (32) Map the LSTM output to the prediction target and output the strain increment Δ in the future period e pre ( t k + Δt ) is predicted as follows: (33) in, i CNN Represents the weight parameters of the convolutional neural network layer; i LSTM Represents the weight parameters of the long short-term memory network layer; i FC Represents the weight parameters of the fully connected layer (output layer); The deep reinforcement learning prediction model uses the actual strain increment Δ acquired in real time e true ( t k + Δt ) and the difference between the predicted results to construct the reward function: (34) This reward function constitutes a feedback signal, which is used to update the model parameters online, achieve adaptive adjustment, and finally form a closed loop.

[0017] Preferably, the stress sensing devices corresponding to the top and bottom of the sample completely cover the top and bottom surfaces of the sample.

[0018] The present invention can truly reflect the impact of mining disturbance: during the implementation of deep tunnel surrounding rock engineering, the present invention installs samples in undisturbed areas to monitor in real time the time-dependent deformation characteristics of the tunnel surrounding rock under the changes in mining disturbance stress, thereby truly reflecting the impact of mining disturbance on rock behavior.

[0019] The present invention can fully capture the anisotropic characteristics and use multi-directional specimens and corresponding monitoring equipment to fully reflect the anisotropic mechanical behavior of the tunnel surrounding rock. The multi-sensor technology is used to collect key parameters such as stress, strain, displacement, and environmental temperature and humidity inside the rock in real time and at a high sampling rate, and ensure that the data of each sensor is collected synchronously under a unified time reference, so as to obtain comprehensive and accurate rock mass status information.

[0020] The present invention adopts automated monitoring and data analysis: the system realizes automated data collection and remote transmission, which greatly improves monitoring efficiency and data reliability, and provides solid data support for on-site real-time decision-making.

[0021] The prediction model of the present invention is highly practical: the time-dependent deformation prediction model constructed based on field monitoring data can accurately predict the deformation trend of the rock mass in the future, and provide a scientific basis and early warning information for safe support of surrounding rock in deep tunnels, and has high practical value.

[0022] The invention has high engineering application value: through real-time data collection, online prediction and intelligent early warning, the system can detect potential risks in advance and realize refined management. It provides a scientific basis for tunnel surrounding rock control and safety support and tunnel surrounding rock health monitoring, which helps to optimize the tunnel surrounding rock filling support plan and improve the overall tunnel surrounding rock strength. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 It is a flow chart of the present invention.

[0024] Figure 2 Schematic diagram of the sample structure.

[0025] Figure 3 Schematic diagram of sample installation.

[0026] Figure 4 Schematic diagram of stress sensing device and strain acquisition device.

[0027] Figure 5 It is a structural diagram of the hydraulic loading device and data acquisition system.

[0028] In the figure, 1-coal seam, 2-upper hydraulic loading device, 3-sample, 4-x-axis hydraulic loading device, 5-lower hydraulic loading device, 6-hydraulic device, 7-data integration system, 8-installation cavity, 9-y-axis hydraulic loading device, 10-ground control center, 11-stress sensing device, 12-deformation monitoring meter, 13-telescopic arm, 14-mining tunnel. DETAILED DESCRIPTION

[0029] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings.

[0030] The present invention discloses a monitoring and prediction model for anisotropic time-dependent deformation of surrounding rock in tunnels under mining disturbance. Figure 1 As shown, the specific steps include: Step 1: In-situ sample preparation: The sampling location is selected at a depth of 100m inside the advanced support pressure zone of the coal mining face to ensure that the surrounding rock of the mined tunnel is not disturbed by mining and truly reflect the original state of the surrounding rock of the deep tunnel. A drill rod with a diameter of 200mm is used for drilling sampling. This diameter not only ensures sufficient cross-sectional size of the coal core, but also maintains the integrity of the anisotropic structure inside the sample, providing sufficient materials for subsequent multi-directional specimen preparation. The drilled samples are pretreated in time, and the temperature, humidity and other environmental parameters are recorded to ensure that the samples are kept in situ as much as possible during storage. Use suitable wrapping materials (such as anti-static film or thermal insulation materials) to seal the coal core to prevent additional damage due to changes in the external environment during transportation and storage. According to the natural bedding structure of the tunnel surrounding rock, the coal core is pretreated and divided, such as Figure 2 As shown, high-precision CNC cutting equipment is used to cut and sample along three directions: horizontal bedding 0°, vertical bedding 90°, and inclined bedding 45°, to form a rectangular standard sample of uniform size, 50mm × 50mm × 100mm, and the error of each size of the sample is no more than ±0.5mm. The number of samples in each direction must meet statistical requirements to ensure that the data is significant and representative.

[0031] Step 2, sample installation cavity construction: In order to restore the stress state of the tunnel surrounding rock under real mining disturbance, this step constructs a sample installation cavity in front of the mining working face to accommodate the standard sample prepared above and simulate the on-site in-situ stress environment.

[0032] Specific as Figure 3As shown, a mounting cavity 8 is constructed in the coal seam 1 of the mining tunnel 14 in the spatial area not disturbed by mining, ensuring that the location of the mounting cavity can truly reflect the on-site stress field and crack distribution characteristics in the later stage. In order to ensure that the sample can fully restore the original stress state in the mounting cavity, the size of the mounting cavity 8 should be added with a safety margin based on the sample size. In theory, the depth, width, and height of the mounting cavity meet the following requirements: (1) in, L chamber , W chamber , H chamber They are the depth, width and height of the installation cavity respectively; L sample , W sample , H sample are the length, width and height of the sample respectively (the size of the sample in the present invention is 50mm×50mm×100mm); d To install margin, adjust according to monitoring equipment; L max After determining the width and height, the safe excavation depth that can be achieved in the depth direction is determined according to the actual situation on site, generally not less than 1 m. In practical applications, in order to ensure structural stability and construction process requirements, the size of the installation cavity is usually slightly larger than the above theoretical value. During the construction process, the roughness of the slot wall is strictly controlled, and the surface of the cavity is machined or manually trimmed to ensure flatness and rounded corners at the end to reduce local stress concentration. In order to obtain the anisotropic characteristics of the tunnel surrounding rock, at least 9 sample installation cavities should be arranged in the same tunnel to carry out field monitoring experiments on the anisotropic time-dependent deformation of the deep tunnel surrounding rock under real mining disturbance. The sample installation cavity closest to the working face should be more than 400m away from the working face to ensure the stability of the excavation of the installation cavity and the installation and stress recovery of the sample.

[0033] Step 3, sample installation and monitoring system layout: This step aims to use six hydraulic loading devices with telescopic arms to accurately install and preload standard samples, and realize stress recovery and real-time monitoring in all directions (x, y, and z axes). To this end, a step-by-step installation and adjustment scheme is adopted to fully preload the samples in the in-situ environment and ensure uniform and coaxial loading. Among them, facing the spatial cavity, the y-axis is defined as the depth direction of the spatial cavity; the z-axis is defined as the up and down direction of the sample, where the bottom end of the sample corresponds to the lower end of the z-axis, and the upper end of the sample corresponds to the upper end of the z-axis; the x-axis is defined as the left and right direction of the sample. A total of six hydraulic loading devices are configured, and each device is equipped with a telescopic arm to facilitate flexible adjustment of the sample position and loading direction during installation. The upper and lower z-axis hydraulic loading devices are respectively installed at the lower and upper ends of the installation cavity, and the telescopic adjustment is synchronously applied to balance the load. The x-axis hydraulic loading devices 4 on both sides: They are installed on the left and right sides of the sample, respectively, and the balanced load is synchronously applied through telescopic adjustment. Y-axis hydraulic loading device 9: one is fixed in the center of the y-axis in the installation cavity, and the other is installed outside the cavity and equipped with a support device for final pre-tightening adjustment in the y-axis direction. The six hydraulic loading devices are used together to accurately adjust the position during the specimen installation process, achieve multi-axial coaxial pre-tightening, and greatly improve the installation accuracy and loading uniformity. Figure 4 As shown, there is a stress sensing device 11 at the end of the hydraulic loading device, in which a fiber Bragg grating stress sensor is embedded to measure the three-dimensional stress σ x ,σ y ,σ z The stress sensing device is a cube. The loading surface area of ​​the cube in the z-axis direction is not less than the bottom area of ​​the sample (50mm×50mm), and the loading surface area of ​​the cube in the y-axis and x-axis directions is not less than the bottom area of ​​the sample (50mm×100mm). A deformation monitoring meter 12 is installed on the telescopic arm 13 of the hydraulic loading device to monitor its displacement and deformation along the axial direction.

[0034] like Figure 5As shown, the specific process of sample installation is as follows: First, on the y-axis inside the sample installation cavity 8, fix the y-axis hydraulic loading device 9 in the center of the cavity wall to ensure that it can accurately collect deep stress data. Install the lower hydraulic loading device 5 at the lower end of the sample installation cavity 8. Place the bottom end of the sample 3 (corresponding to the lower end of the z-axis) on the lower hydraulic loading device 5 so that the bottom of the sample is coaxial with the loading surface of the device. Put the z-axis upper hydraulic loading device 2 with the bottom end of the sample fixed into the installation cavity together with the sample 3. Adjust the telescopic arm of the lower hydraulic loading device 5 so that the sample 3 and the aforementioned y-axis hydraulic loading device 9 are precisely coaxial in the y-axis direction. Install two x-axis hydraulic loading devices 4 on the left and right sides of the sample respectively. At the same time, adjust the telescopic arms on both sides so that they move synchronously to achieve balanced pre-tightening of the sample in the x-axis direction, ensure consistent loading on the left and right sides, and prevent lateral deviation caused by unbalanced loading. Install the hydraulic loading device at the upper end of the sample, and also use the telescopic arm to dock the upper surface of the sample. Adjust the upper hydraulic loading device to ensure that it works together with the lower hydraulic loading device to uniformly preload the specimen in the z-axis direction. Install the hydraulic device 6 at the open end of the specimen mounting cavity 8, supplemented by an external support device to ensure that the device remains stable during the adjustment process. Adjust the device to achieve the final preload with the specimen in the y-axis direction. Finally, apply the initial preload force through the telescopic arms of all six hydraulic devices to make the preload forces in the six directions consistent, but the preload force should not be too large.

[0035] In addition, micro temperature and humidity sensors are added inside the cavity to monitor the local temperature in real time. A data integration system 7 is added underground, and each sensor is connected to the data acquisition unit (DAQ) through a dedicated interface. The unit supports high sampling rate and multi-channel synchronous acquisition to ensure the uniformity of data timestamps, and finally summarizes the data in the data integration system. The data integration system is connected to the underground communication sensor network, and finally the data is transmitted back to the ground control center 10 to achieve safe and effective monitoring.

[0036] In this process, a preload force feedback system is constructed to implement regulation by real-time monitoring of the force sensors on each hydraulic loading system, and timely adjust the hydraulic device parameters to control the loading error within 5% to ensure balanced preload in each axis direction. Specifically, the sampling frequency of the stress sensor is set to T s , at each sampling moment k Collect the current record actual load force F actual [ k ], the target preload force is recorded as F target , then the error can be calculated e [ k ]: (2) Adopt closed-loop control algorithm based on PID control to generate control instructions u [ k ] is used to timely adjust the hydraulic device parameters (such as oil pressure, flow, etc.) to compensate for errors: (3) in, K p is the proportional gain, which controls the amplification of the instantaneous error; K i is the integral gain, which accumulates past errors to eliminate steady-state deviations; K d The differential gain is used to predict the future error trend to suppress oscillation. To cope with the nonlinearity and uncertainty of the on-site environment, the preload force feedback system introduces an adaptive correction algorithm. The system compares the real-time collected loading force data with the target preload force, calculates the error, and generates a control signal through the PID controller. u [ k ], this signal drives the hydraulic actuator to adjust the pressure parameters so that the loading force gradually approaches the target value: (4) Step 4: In-situ test environment construction and in-situ stress recovery: According to the field exploration data, determine the target stress values ​​of the sample in the x, y, and z directions, which are recorded as s target,x , s target,y and s target,z The actual stress values ​​are s x , s y、 s z . An independent and synchronous loading scheme is adopted in each direction. Through a unified hydraulic loading system, it is ensured that the loading devices of each axis apply preload at the same time and realize independent regulation, which truly simulates the in-situ stress state. In order to ensure loading stability, a fine-tuning control theory is constructed: When the error is small, a higher sampling frequency is used to capture the subtle changes in stress error and calculate: (5) According to the online error and change trend, the PID parameters are updated using the adaptive algorithm: (6) in α p , α i and αd is the adaptive factor, f [ k ] is the system dynamic prediction function. This fine-tuning strategy ensures that when the system approaches the target stress state, the loading can be fine-tuned in a smooth and high-precision manner, further reducing the error.

[0037] Step 5: Data acquisition: Through the multi-sensor data acquisition platform, the multi-dimensional parameter data of the sample during the loading process is acquired in real time and for a long time. Including the stress sensor to collect the stress data of the sample in the x, y, and z directions in real time s x ( t ), s y ( t ) 、 s z ( t ), strain data recorded by the strain gauge installed on the telescopic arm of the hydraulic loading device e ij ( t ) and displacement data is d i ( t ), collect local temperature T C ,humidity RH Environmental data such as temperature and humidity can be collected to correct for thermal stress effects and other environmental factors during the accelerated test.

[0038] The present invention specially constructs a data storage and transmission model for storing and transmitting field monitoring data. Including stress data s x ( t ), s y ( t ) 、 s z ( t ), strain data e ij ( t ) and displacement data d i ( t ),temperature T C ,humidity RH. Through the acquisition unit (DAQ), high sampling rate and multi-channel synchronous acquisition are achieved to ensure the uniformity of data timestamps, and finally the data is summarized in the data integration system and transmitted back to the ground control center to achieve safe and effective monitoring. The model is as follows: For each sensor (such as stress, strain, displacement, temperature sensor, humidity sensor), let its true signal be S j (t ), the measurement noise is N j ( t ), assuming the total number of sensors is N , then the continuous output signal of the sensor can be described as: (7) The sampling period of all sensors is unified as T s , k Used to represent the sampling point index during discrete sampling, that is, k sampling points, then the discrete sampling time t k The definition can be defined as: (8) At each sampling moment, data is collected from each sensor to obtain discrete sampling data: (9) In order to ensure the synchronization of multi-channel acquisition, the timestamp of each channel is required to meet the following requirements: (10) x The permissible deviation is extremely small.

[0039] After that, the data collected by each channel is t k Composition of data vector: (11) In order to improve data quality and facilitate subsequent transmission, this system has built-in multi-level pre-processing modules in the data acquisition unit. To eliminate high-frequency noise, the data vector X[ k ] to perform low-pass filtering. Set the filter impulse response to h [ n ], through the convolution operation, the high-frequency noise is suppressed, so that the signal data after filtering is obtained as follows: (12) in, L is the filter length. This operation can be designed with FIR filter to make the system have a smooth frequency domain response. Discrete wavelet transform is performed on the data after low-pass filtering. Wavelet transform decomposes the signal into different scales and time positions. W i Wavelet coefficients reflect the frequency components of the signal in the local area. l The wavelet coefficients are soft-thresholded, and the specific criteria are: (13) parameter l It is usually based on the noise level (estimated by the median absolute deviation) and the number of samples. n Determine as follows: (14) By inverse wavelet transform Reconstruct the signal and get the denoised signal as W ( [ k ]). Perform inverse wavelet transform on the processed wavelet coefficients to obtain the denoised signal: (15) In order to further improve data smoothness and estimation accuracy, the Kalman filter is used to estimate the state of the denoised signal, where A is the state transfer matrix: (16) Error covariance prediction: (17) in, Q is the process noise covariance.

[0040] set up H is the observation matrix, R To measure the noise covariance, the Kalman gain is calculated as: (18) Then the updated state is estimated to be: (19) The updated error covariance is: (20) This step adaptively smoothes the collected data, providing a signal with a higher signal-to-noise ratio for subsequent data compression and transmission. In order to reduce the transmission bandwidth requirement, the denoised signal is compressed after preprocessing. The transform coding method is used for data compression, and then the data is converted into a bit stream through adaptive quantization and entropy coding. The compression function is denoised as Q , then the final preprocessing output vector data is: (twenty one) In summary, the overall operation model of the preprocessing module can be expressed as: (twenty two) After multi-level pre-processing, not only the signal quality is significantly improved, but also the amount of data is greatly reduced, facilitating subsequent low-latency transmission.

[0041] The preprocessed data vector Y[k] is stored in the data acquisition unit and then transmitted in the form of a data packet. The data packet is defined as: (twenty three) Therefore, the entire data stream can be represented as a continuous data set D: (twenty four) For multi-channel configurations, each unit m Generate a data set D m , including their respective timestamps and preprocessed data, set the m The data set of units is recorded as: (25) The data integration system performs time alignment and fusion of data from each unit to form an overall data set, providing a unified data source for global monitoring and subsequent analysis: (26) This process involves timestamp correction and multi-sensor data fusion algorithms to ensure that each data packet has high consistency in time and content.

[0042] After the downhole system integration stores the data, it is transmitted through the downhole communication network. In order to achieve real-time monitoring, when building data transmission, a low-latency transmission network should be selected to cooperate with the construction of the data transmission model, and the data return delay should be less than 50 ms: (27) Therefore, the final data is sent back to the ground control center. Indicates that the ground side is at time t The data received at the time is in the form of: (28) By correcting for time offset, it is ensured that the data has the correct time reference when analyzing.

[0043] Step 6: Construct a time-dependent deformation model based on the comprehensive data set transmitted to the ground control center in step 5 , which contains stress data s x ( t ), s y ( t ) s z ( t ), strain data e ij (t), displacement data d i (t ), and environmental parameters temperature T C ( t ) and humidity RH ( t ), a prediction model of anisotropic time-dependent deformation of surrounding rock in deep tunnels under real mining disturbance is constructed. Firstly, combined with the real environment measurement parameters underground, the cumulative strain is constructed by combining damage mechanics and creep theory: (29) in, e ij 0 is the initial strain, is the creep strain rate; A ij is a material constant, reflecting the basic creep characteristics in different directions; s ij ( t ) represents the corresponding component of the stress tensor, which can be expressed as s x ( t ), s y ( t ) s z ( t ); n ij is the stress sensitivity index; O ij is the activation energy, J is the gas constant, T C ( t ) is temperature; g ( RH ( t )) is the humidity modulation function, which can be a linear or exponential function, describing the effect of humidity on the creep rate; in order to characterize the effect of material damage on the creep process, anisotropic damage variables are introduced D ij ( t ), describing the degree of local structural degradation of the material; m ij Describe the damage inhibition effect.

[0044] In the framework of damage mechanics and creep theory, this method proposes a coupling model. ij The creep strain rate in the direction is described by: (30) in, B ij is the damage rate constant; p ,q , r A sensitivity index that describes the effects of stress and strain on damage accumulation; O D,ij is the damage activation energy; h ( RH ( t )) is the humidity modulation function. The model considers the coupling effect of stress, accumulated strain and residual good rate on damage rate.

[0045] Step 7: Construct a time-dependent deformation prediction model: Based on a comprehensive dataset of downhole transmission data , build a deep reinforcement learning prediction model to predict the strain increment in the future period. The model mainly consists of the following modules: First of all Perform convolution operations to extract local spatial features: (31) Then, LSTM is used to process the time series features output by the convolutional layer to capture the dynamic evolution of time-dependent deformation: (32) Map the LSTM output to the prediction target and output the strain increment Δ in the future period e pre ( t k + Δt ) is predicted as follows: (33) In the above three formulas, i CNN Represents the weight parameters of the convolutional neural network layer; i LSTM Represents the weight parameters of the long short-term memory network layer; i FC Represents the weight parameters of the fully connected layer (output layer).

[0046] The model uses the actual strain increment Δ e true ( t k + Δt ) and the difference between the predicted results to construct the reward function: (34) This reward function constitutes a feedback signal, which is used to update the model parameters online, achieve adaptive adjustment, and finally form a closed loop.

[0047] Step 8: Apply the prediction model to the control and safety support of deep tunnel surrounding rock and health monitoring of tunnel surrounding rock. By comparing the prediction results with the real-time monitoring data, the system pre-sets the strain increment safety threshold. When the model predicts that the future strain increment reaches or exceeds the safety threshold, the system automatically triggers an early warning. The early warning signal will be transmitted to the ground control center through the underground communication network and further sent to the on-site monitoring terminal. When the prediction shows that the strain increases sharply and there is a potential risk of instability, the system can recommend timely addition of support structures or local pressure relief treatment to reduce the risk of sudden accidents. In addition, the system also supports automatic recording of early warning events and related parameters to provide data basis for subsequent accident investigations and safety assessments. After receiving the underground data and early warning information, the ground control center generates an early warning report to help decision makers understand the safety status of the deep tunnel surrounding rock in a timely manner and organize on-site emergency disposal.

[0048] Of course, the present invention may have many other embodiments. Without departing from the spirit and essence of the present invention, technicians familiar with the field may make various corresponding changes and deformations based on the present invention, but these corresponding changes and deformations should all fall within the scope of protection of the claims attached to the present invention.

Claims

1. The anisotropic time-dependent deformation monitoring and prediction model of tunnel surrounding rock under mining disturbance is characterized by: Model building includes the following steps: Step 1, in-situ sample preparation, select the tunnel surrounding rock that is not disturbed by mining in the advanced support pressure area of ​​the coal mining face, obtain coal samples by coring, cut along the horizontal bedding 0°, vertical bedding 90°, and inclined bedding 45° of the coring coal samples, and obtain three kinds of samples. The three samples have the same size and are rectangular; Step 2: constructing a sample installation cavity, constructing an installation cavity in a surrounding rock area of ​​the mining tunnel that is not disturbed by mining in front of the coal mining face, wherein the size of the installation cavity is larger than the sample; Step 3: Sample installation and monitoring system arrangement: Place the sample in the installation cavity, arrange a hydraulic loading device at the open end of the installation cavity, support the sample and the five side walls of the installation cavity through the hydraulic loading device, and the hydraulic loading device is supported on the end surface of the sample, finally forming a three-axis support; The loading end of the hydraulic loading device is provided with a stress sensing device, a fiber Bragg grating stress sensor is embedded in the stress sensing device, a strain gauge is arranged on the telescopic arm of the hydraulic loading device, a temperature sensor and a humidity sensor are arranged in the installation cavity, the stress sensor, the strain gauge, the temperature sensor and the humidity sensor are all connected to a data acquisition device, and the data acquisition device is connected to a ground control center; The hydraulic loading device is controlled by a ground control center; Step 4: In-situ test environment construction and in-situ stress recovery: According to the field exploration data, the target value of the in-situ ground stress of the sample is determined, which is spatially represented as three directions: x, y, and z. A preload is applied to the sample through a hydraulic loading device to simulate the in-situ stress state; Step 5: Data collection: Under the disturbance of mining, real-time data from stress sensors, strain gauges, temperature sensors, and humidity sensors are collected to correct thermal stress effects and other environmental factors during the accelerated test; Step 6: construct a time-dependent deformation model based on the data obtained in step 5; Step 7: Construct a time-dependent deformation prediction model. Based on the data obtained in step 5, construct a deep reinforcement learning prediction model to predict the strain increment in the future period.

2. The anisotropic time-dependent deformation monitoring and prediction model of tunnel surrounding rock under mining disturbance according to claim 1 is characterized in that: After step 7, also include: Step 8: Apply the time-dependent deformation prediction model to mine tunnel support design, coal mining face advancement speed optimization, and rock burst prevention and control.

3. The anisotropic time-dependent deformation monitoring and prediction model of tunnel surrounding rock under mining disturbance according to claim 1 is characterized in that: In step 1: The sampling location was selected at a depth of 100m inside the advanced support pressure zone of the coal mining face, and the coal core was drilled using a drill rod with a diameter of 200mm. The dimensions of the sample are 50mm × 50mm × 100mm in width × length × height, with the width being in the x direction, the length being in the y direction, and the height being in the z direction. The dimensional error of the sample in each direction is no more than ±0.5mm.

4. The anisotropic time-dependent deformation monitoring and prediction model of tunnel surrounding rock under mining disturbance according to claim 1 is characterized in that: In step 2: There are at least 9 installation cavities in the same mining tunnel, and at least 3 of each type of specimen are installed; The size of the installation cavity is based on the size of the specimen with a safety margin added, and the depth, width, and height of the installation cavity satisfy the following formulas: ； in, L chamber , W chamber , H chamber They are the depth, width and height of the installation cavity respectively; L sample , W sample , H sample are the length, width and height of the specimen respectively, δ For installation allowance, L max After determining the width and height, the depth direction is determined by the achievable safe excavation depth. L max Not less than 1 m; The specimen installation cavity closest to the coal mining face is at least 400 m away from the coal mining face.

5. The anisotropic time-dependent deformation monitoring and prediction model of tunnel surrounding rock under mining disturbance according to claim 1 is characterized in that: Step 3 includes constructing a preload force feedback system, implementing regulation by real-time monitoring of the force sensors on each hydraulic loading system, and timely adjusting the parameters of the hydraulic device to control the loading error within 5%. The regulation method is as follows; Set the sampling frequency of the stress sensor to T s , at each sampling moment k Collect the current record actual load force F actual [ k ], the target preload is recorded as F target , then the error can be calculated e [ k ]: （2） Adopt closed-loop control algorithm based on PID control to generate control instructions u [ k ] is used to adjust the hydraulic device parameters in time to compensate for errors: （3） in, K p is the proportional gain, which controls the amplification of the immediate error; K i is the integral gain, which accumulates past errors to eliminate steady-state deviations; K d The system is used to calculate the differential gain and predict the future error trend to suppress oscillation. To cope with the nonlinearity and uncertainty of the on-site environment, the preload force feedback system introduces an adaptive correction algorithm. The system compares the real-time collected loading force data with the target preload force, calculates the error, and generates a control signal through the PID controller. u [ k ], this signal drives the hydraulic actuator to adjust the pressure parameters so that the loading force gradually approaches the target value: （4）。 6. The anisotropic time-dependent deformation monitoring and prediction model of tunnel surrounding rock under mining disturbance according to claim 5 is characterized in that: Step 4 includes fine-tuning control, which is as follows: When the error is smaller, a higher sampling frequency is used to capture the subtle changes in stress error and calculate: （5） According to the online error and change trend, the PID parameters are updated using the adaptive algorithm: （6） in σ target,x , σ target,y and σ target,z is the target stress value of the specimen in the x, y, and z directions, σ x , σ y、 σ z is the actual stress value of the specimen in the x, y, and z directions, α p , α i and α d is the adaptive factor, f [ k ] is the system dynamic prediction function.

7. The anisotropic time-dependent deformation monitoring and prediction model of tunnel surrounding rock under mining disturbance according to claim 1 is characterized in that: In step 5, the data collection method and signal processing for each sensor are as follows: At each sampling moment, data is collected from each sensor to obtain discrete sampling data: （9） The timestamp of each channel satisfies: （10） ξ The permissible deviation is extremely small; The data collected by each channel at time t k Composition of data vector: （11） For the data vector X[ k ] to perform low-pass filtering and set the filter impulse response to h [ n ], through the convolution operation, the filtered signal data is obtained as follows: （12） in, L is the filter length; The low-pass filtered data is subjected to discrete wavelet transform to decompose the signal into different scales and time positions. W i is the wavelet coefficient, according to the set threshold λ The wavelet coefficients are soft-thresholded and the criterion is: （13） parameter λ Depending on the noise level and number of samples n Determine as follows: （14） By inverse wavelet transform Reconstruct the signal and get the denoised signal as , perform inverse wavelet transform on the processed wavelet coefficients, and get the denoised signal as follows: （15） The Kalman filter is used to estimate the state of the denoised signal, where A is the state transfer matrix: （16） After preprocessing, the denoised signal is compressed, and the compression function is recorded as Q , then the final preprocessing output vector data is: （21） In summary, the overall operation model of the preprocessing module is: （22） The preprocessed data vector Y[k] is stored in the data acquisition unit and then transmitted in the form of a data packet. The data packet is defined as: （23） The entire data stream is represented as a continuous data set D: （24） For multi-channel configurations, each unit m Generate a data set D m , including their respective timestamps and preprocessed data, set the m The data set of units is recorded as: （25） The data integration system performs time alignment and fusion of data from each unit to form an overall data set, providing a unified data source for global monitoring and subsequent analysis: （26） When building data transmission, choose a low-latency transmission network to cooperate with the construction of the data transmission model, and the data return delay is less than 50 ms: （27） Construct the final return data at the ground control center, Indicates that the ground side is at time t The data received at the time is in the form of: （28）。 8. The anisotropic time-dependent deformation monitoring and prediction model of tunnel surrounding rock under mining disturbance according to claim 7 is characterized in that: Step 6: Constructing a time-dependent deformation model includes: Construct the cumulative strain as follows: （29） in, ε ij 0 is the initial strain, is the creep strain rate, A ij To reflect the material constants of the basic creep characteristics in different directions, σ ij ( t ) is the corresponding component of the stress tensor, expressed as σ x ( t ), σ y ( t ) σ z ( t ); n ij is the stress sensitivity index; O ij is the activation energy, J is the gas constant, T C ( t ) is temperature; g ( RH ( t )) is the humidity modulation function, D ij ( t ) is the anisotropic damage variable, m ij is the damage inhibition parameter; The sample is ij The creep strain rate in the direction is as follows: （30） in, B ij is the damage rate constant; p , q , r A sensitivity index that describes the effects of stress and strain on damage accumulation; O D,ij is the damage activation energy; h ( RH ( t )) is the humidity modulation function.

9. The anisotropic time-dependent deformation monitoring and prediction model of tunnel surrounding rock under mining disturbance according to claim 7 is characterized in that: In step 7, the data set is synthesized based on the downhole transmission data , build a deep reinforcement learning prediction model, which consists of the following modules: First of all Perform convolution operations to extract local spatial features: （31） Then, LSTM is used to process the time series features output by the convolutional layer to capture the dynamic evolution of time-dependent deformation: （32） Map the LSTM output to the prediction target and output the strain increment Δ in the future period ε pre ( t k + Δt ) is predicted as follows: （33） in, θ CNN Represents the weight parameters of the convolutional neural network layer; θ LSTM Represents the weight parameters of the long short-term memory network layer; θ FC Represents the weight parameters of the fully connected layer (output layer); The deep reinforcement learning prediction model uses the actual strain increment Δ acquired in real time ε true ( t k + Δt ) and the difference between the predicted results to construct the reward function: （34） This reward function constitutes a feedback signal, which is used to update the model parameters online, achieve adaptive adjustment, and finally form a closed loop.

10. The anisotropic time-dependent deformation monitoring and prediction model of tunnel surrounding rock under mining disturbance according to claim 1 is characterized in that: The stress sensing devices corresponding to the top and bottom of the specimen completely cover the top and bottom surfaces of the specimen.

Images