Underground multi-parameter environment monitoring method

Abstract

The invention relates to the technical field of mine mining, and discloses an underground multi-parameter environment monitoring method, which comprises the following steps of: acquiring equipment pose, vibration spectrum and environment parameters in real time through a multi-source sensor, performing space-time alignment, filtering noise reduction and feature extraction on original data, outputting a standardized state vector, and performing data processing on the standardized state vector; fusing the pose data, the vibration signals and the environmental parameters under a unified space-time reference to construct a multi-dimensional feature matrix; and carrying out dimension reduction and redundant information elimination by adopting principal component analysis, and outputting a fused equipment state vector. According to the method, through real-time fusion of the equipment pose and the airflow dynamic state, the ventilation equipment is cooperatively regulated and controlled, the interference of mechanical vibration on the gas monitoring precision is eliminated and the reliability of gas concentration detection is guaranteed based on coupling modeling of vibration spectrum and pose drift and intelligent compensation of sensor reading errors, and through space correlation analysis and dynamic air volume optimization, the gas concentration detection accuracy is improved. Wind speed sudden drop caused by tramcar passing is actively eliminated, and roadway global continuous monitoring is achieved.

Description

Technical Field

[0001] The present invention relates to the field of mine mining, and more particularly to an underground multi-parameter environment monitoring method. Background Art

[0002] In deep mines, dynamic operations such as tunnel boring machines and transport vehicle movements cause wind disturbances and equipment interference, leading to three core problems for traditional environmental monitoring systems: The air volume at the tunneling face fluctuates by more than ±15% with equipment displacement; Equipment vibration causes sensor reading drift (gas detection error reaches ±0.15%); When the transport vehicle passes, the local wind speed drops by 40%, forming a monitoring blind spot.

[0003] Existing static monitoring models are difficult to cope with such dynamic interference, and there is an urgent need to develop monitoring methods with strong device-environment coupling. Summary of the Invention

[0004] The present invention provides an underground multi-parameter environment monitoring method to solve the technical problem in related technologies that existing static monitoring models are difficult to cope with such dynamic interference.

[0005] The present invention provides a method for monitoring an underground multi-parameter environment, comprising the following steps: S100, real-time device status perception: This system uses multi-source sensors to collect device posture, vibration spectrum, and environmental parameters in real time, performs spatiotemporal alignment, filtering, noise reduction, and feature extraction on the raw data, and outputs a standardized state vector. S200, multi-source data fusion processing: The posture data, vibration signals, and environmental parameters are integrated under a unified spatiotemporal reference to construct a multi-dimensional feature matrix. Principal component analysis is used to reduce the dimensionality and eliminate redundant information, and the fused device state vector is output. S300, digital twin disturbance modeling: This system establishes vibration sensitivity models and environmental disturbance models based on fused data, fits equations using the least squares method, verifies model accuracy online, and outputs pose drift prediction equations. S400, dynamic feedforward compensation control: predicts posture drift based on the disturbance model, generates control signals using a feedforward-PID composite controller, issues control commands to actuators in real time, and verifies compensation effects using measured data; S500, edge intelligent decision-making: evaluates compensation performance indicators, dynamically triggers model updates or parameter optimization, performs hierarchical responses based on comprehensive security situation, and synchronizes decisions to the digital twin platform; S600, control command output: reconstruct the control signal, perform encryption encoding and timestamp synchronization, transmit the command according to priority through the real-time network, receive execution feedback and calculate the control error and delay to form a closed-loop verification.

[0006] Furthermore, in S100, the following steps are included: S110, equipment posture acquisition: uses UWB base stations and tags to capture the 3D coordinates and posture angles of the tunnel boring machine and transport vehicle in real time, LiDAR scans the area point cloud data, and dynamically models the surface deformation of the equipment; S120, environmental parameter collection: synchronously reads gas / temperature and humidity / wind speed sensor data, and the air pressure sensor monitors the internal pressure of the wind tube; S130, vibration signal acquisition: accelerometers are deployed at key locations on the equipment to collect 0-500Hz raw vibration signals; S140, vibration spectrum calculation: perform FFT transformation on the original acceleration signal and calculate the power spectrum density; S150, data encapsulation and timestamp: package pose / environment / spectrum data and add PTP1588v2 protocol timestamp.

[0007] Furthermore, in S200, the following steps are included: S210, PTP1588v2 time synchronization: aligns all device / sensor time bases through a master-slave clock architecture, calculates clock offsets and corrects timestamps; S220, device coordinate system → roadway global coordinate system conversion: convert the device posture to the roadway global coordinate system, and integrate the LiDAR point cloud data to correct the positioning error; S230, environmental sensor space mapping: calculating global coordinates based on sensor installation locations and establishing a sensor-device spatial relationship matrix; S240, vibration spectrum spatiotemporal correlation: Correlate the vibration spectrum density vector to a specific device location and establish a frequency-space distribution model; S250, spatiotemporal index construction: Establish an R*-tree spatiotemporal index to achieve millisecond-level cross-modal data retrieval.

[0008] Furthermore, in S300, the following steps are included: S310, Vibration-Position Coupling Modeling: Analyze the quantitative relationship between vibration spectrum and equipment posture drift, and establish a frequency sensitivity matrix; S320, Environmental-Sensor Error Modeling: Quantify the impact of environmental parameters on positioning / measurement errors and construct a temperature, humidity, and gas composite interference model; S330, comprehensive prediction of posture drift: Integrates vibration and environmental factors to predict device posture changes and build a state space model; S340, disturbance parameter identification: training sensitivity coefficient and interference coefficient based on historical data; S350, Model Validation and Confidence Assessment: Calculate prediction residuals and confidence intervals, and dynamically update model parameters.

[0009] Furthermore, the calculation formula for vibration-posture coupling modeling is as follows: Posture drift caused by vibration: ; ; ; in, is the pose drift vector caused by vibration, is the frequency sensitivity matrix, is the vibration spectrum density vector, is the power spectrum density value of the first frequency point, is the power spectrum density value of the second frequency point, is the power spectrum density value of the Kth frequency point, is the transpose operation; Frequency sensitivity matrix: ; in, is the sensitivity coefficient of the j-th frequency band vibration to the i-th axial drift, is the sensitivity coefficient of the first frequency band vibration to the x-axis drift, is the sensitivity coefficient of the second frequency band vibration to the x-axis drift, is the sensitivity coefficient of the K-band vibration to the x-axis drift, is the sensitivity coefficient of the first frequency band vibration to the y-axis drift, is the sensitivity coefficient of the second frequency band vibration to the y-axis drift, is the sensitivity coefficient of the K-band vibration to the y-axis drift, is the sensitivity coefficient of the first frequency band vibration to the z-axis drift, is the sensitivity coefficient of the second frequency band vibration to the z-axis drift, is the sensitivity coefficient of the K-band vibration to the z-axis drift, is the spectrum vector dimension, is the three-dimensional coordinate of the device in the global coordinate system, is the change in position components on the x, y, and z axes in the global coordinate system, is the variation operator.

[0010] Furthermore, the calculation formula for environment-sensor error modeling is as follows: Environmental comprehensive error: ; ; ; in, is the comprehensive error caused by environmental factors, is the temperature deviation, is the current temperature, is the reference temperature, is the humidity deviation, is the current humidity, For reference humidity, is the gas / carbon monoxide concentration, is the temperature interference coefficient, is the humidity interference coefficient, is the gas concentration interference coefficient, is the interference coefficient of carbon monoxide concentration.

[0011] Furthermore, the calculation formula for comprehensive prediction of posture drift is as follows: State space equations: ; ; in, is the device state vector, is the state vector of the device at the next moment, is the three-dimensional coordinate of the device in the global coordinate system, is the velocity component in the global coordinate system, is the state transition matrix, is the vibration disturbance input matrix, is the environmental perturbation input matrix, is Gaussian process noise, is a discrete time index.

[0012] Furthermore, in S400, the following steps are included: S410, disturbance prediction calculation, predicts posture drift based on the state space model of S330, integrating vibration and environmental disturbance factors; S420, generating compensation control quantity: designing a PID controller to generate compensation instructions, introducing feedforward compensation to offset prediction drift; S430, multi-objective collaborative optimization: resolve multi-device compensation conflicts and meet safety and energy constraints; S440, compensation instruction issuance: The instruction is issued to the device PLC via the OPCUA protocol. The instruction format is standardized and the transmission delay requirement is <20ms. S450, Compensation Effect Verification: Compare the posture errors before and after compensation and calculate key performance indicators.

[0013] Furthermore, in S500, the following steps are included: S510, dynamic evaluation of compensation effect: calculate key performance indicators based on the verification results of S450, trigger model update and parameter optimization threshold; S520, online update of disturbance model: using recursive least squares method to update vibration sensitivity coefficient and environmental interference coefficient in real time; S530, collaborative optimization of control parameters: Optimize PID gains and feedforward gains through particle swarm optimization; S540, Security Situation Decision: Executes security policies based on integrated environmental and vibration data, generating device-level / system-level emergency commands; S550, Digital Twin Synchronization and Log: Synchronizes decision results to the digital twin platform, generates encrypted decision logs and security requirements, log integrity is SHA-256 hash check, and synchronization delay is synchronization delay time <50ms.

[0014] Furthermore, in S600, the following steps are included: S610, control signal reconstruction: Fusion of the control parameters optimized in S530 and the real-time posture error generates a control signal, embedding feedforward compensation; S620, security instruction embedding: Generate device-level security linkage instructions based on the security decision coverage control signal of S540; S630, Signal Security Coding: Encrypts and verifies the integrity of control signals and adds timestamp synchronization; S640, real-time transmission scheduling: transmits encrypted instructions via TSN and dynamically assigns transmission priorities; S650, execution status feedback: receiving the execution confirmation signal from the device PLC and calculating the control error and delay.

[0015] The beneficial effects of the present invention are: The present invention integrates the equipment posture and wind flow dynamics in real time, coordinates and controls the ventilation equipment, significantly smoothes the sudden change in air volume caused by equipment movement, ensures the stability of the airflow at the excavation working face, and based on the coupled modeling of vibration spectrum and posture drift, intelligently compensates for sensor reading errors, eliminates the interference of mechanical vibration on gas monitoring accuracy, and ensures the reliability of gas concentration detection. Through spatial correlation analysis and dynamic air volume optimization, it actively eliminates the sudden drop in wind speed caused by the passage of mine cars, realizes continuous monitoring of the entire tunnel, eliminates the risk of local data loss, and finally forms a "perception-modeling-compensation-decision-making" closed loop, thereby improving the robustness and safety of underground environmental monitoring. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 This is a flow chart of a method for downhole multi-parameter environmental monitoring proposed by the present invention; Figure 2 This is a sub-step flow chart of S100 of a downhole multi-parameter environment monitoring method proposed by the present invention; Figure 3 This is a sub-step flow chart of S200 of a downhole multi-parameter environment monitoring method proposed by the present invention; Figure 4 This is a sub-step flow chart of S300 of a downhole multi-parameter environment monitoring method proposed by the present invention; Figure 5 This is a sub-step flow chart of S400 of a downhole multi-parameter environment monitoring method proposed by the present invention; Figure 6 This is a sub-step flow chart of S500 of a downhole multi-parameter environment monitoring method proposed by the present invention; Figure 7 This is a sub-step flow chart of S600 of a downhole multi-parameter environment monitoring method proposed by the present invention. DETAILED DESCRIPTION

[0017] The subject matter described herein will now be discussed with reference to exemplary embodiments. It should be understood that these embodiments are discussed solely to enable those skilled in the art to better understand and implement the subject matter described herein, and that the functions and arrangements of the elements discussed may be varied without departing from the scope of this specification. Various examples may omit, substitute, or add various processes or components as needed. In addition, features described with respect to some examples may also be combined in other examples.

[0018] like Figure 1-Figure 7 As shown, a method for monitoring multi-parameter environment in a well includes the following steps: S100, real-time device status perception: This system uses multi-source sensors (IMU, LiDAR, and environmental sensors) to collect device posture, vibration spectrum, and environmental parameters (temperature, gas concentration, etc.) in real time. It then performs spatiotemporal alignment, filtering, noise reduction, and feature extraction (such as vibration spectrum peaks) on the raw data, outputting a standardized state vector. In one embodiment of the present invention, the following steps are specifically included: S110, equipment posture acquisition (UWB / LiDAR positioning): uses UWB base stations and tags to capture the 3D coordinates and posture angles of the tunnel boring machine and transport vehicle in real time, LiDAR scans the area point cloud data, and dynamically models the surface deformation of the equipment; formula: Device pose vector: ; in, are Cartesian coordinates, is the yaw angle (in rad), is the pitch angle (in rad), is the device posture state vector; Positioning accuracy constraints: ; in, is the pose measurement error vector, is the Euclidean norm, is the positioning accuracy threshold; S120, environmental parameter acquisition (multi-sensor fusion): synchronously reads gas / temperature and humidity / wind speed sensor data, and the air pressure sensor monitors the internal pressure of the wind tube; formula: Environmental parameter vector: ; in, is the temperature, For humidity, is the methane concentration, is the carbon monoxide concentration, is the wind speed, is the air pressure of the blower, is the time index, is the state vector of environmental parameters at time t; S130, vibration signal acquisition (three-axis MEMS accelerometer): accelerometers are deployed at key locations on the equipment to collect 0-500Hz raw vibration signals; formula: Acceleration time domain signal: ; in, is the time variable, is the triaxial acceleration, is the acceleration vector at time τ; Sampling Theorem: ; in, is the sampling frequency, is the highest frequency of the signal, is the Nyquist sampling coefficient; S140, vibration spectrum calculation (edge node preprocessing): Perform FFT transformation on the original acceleration signal and calculate the power spectrum density (PSD); formula: Power spectral density function: ; in, is the frequency, is the Hanning window function, is the number of segments, For each segment, is the discrete-time acceleration signal, is the imaginary unit, is the power spectral density, is the discrete time index, is the segment index; Output spectrum vector: ; in, is the kth frequency point, , is the spectrum vector dimension, Vibration spectral density vector, is the frequency range, The 1st to Kth power spectrum densities respectively; S150, data encapsulation and timestamp: package pose / environment / spectrum data and add PTP1588v2 protocol timestamp; Timestamp alignment: ; in, Including the time offset calculated by the PTP1588 protocol, which is used to correct the clock deviation between different devices. The local timestamp of the sensor, which indicates the local system time when the sensor collects data. The timestamp after synchronization, indicating the unified time base after clock synchronization; Packet serialization: ; in, It is a serialized JSON format data packet. JSON encoding function, converting the data structure into a JSON string. The packet timestamp field, the value is , Is the data field, containing the pose vector , environmental parameter vector and spectrum vector ; Output: structured data packet (including posture, environment, and spectrum data); S200, multi-source data fusion processing: fuses posture data, vibration signals, and environmental parameters under a unified spatiotemporal reference to construct a multi-dimensional feature matrix; uses principal component analysis (PCA) to reduce dimensionality and eliminate redundant information, and outputs the fused device state vector; In one embodiment of the present invention, the following steps are specifically included: S210, PTP1588v2 time synchronization: aligns all device / sensor time bases through a master-slave clock architecture, calculates clock offsets and corrects timestamps; formula: Clock offset calculation: ; in, The time when the master clock sends the Sync message, The time when the slave clock receives the Sync message. Send Delay_Req time from the clock, Receive Delay_Req time for the master clock, is the clock offset, which indicates the time difference between the master and slave clocks; Timestamp correction: ; in, The local original timestamp, the time recorded locally by the device, is the timestamp after synchronization, the unified time after offset correction, is the clock offset; S220, equipment coordinate system → roadway global coordinate system conversion: equipment posture Convert to the global coordinate system of the roadway and fuse LiDAR point cloud data to correct positioning errors; formula: Homogeneous transformation matrix: ; in, is the homogeneous transformation matrix used for coordinate system transformation, is the yaw angle rotation matrix, which represents the rotation of the device in the horizontal plane. is the pitch angle rotation matrix, which represents the vertical rotation of the device. is the translation vector, which represents the displacement of the coordinate origin. is the yaw angle, which indicates the rotation angle of the device in the horizontal plane. is the pitch angle, which indicates the inclination of the device relative to the horizontal plane. is the coordinate of the roadway reference point, is the cosine function, is a sine function, and is the matrix filling constant; Global coordinate calculation: ; ; in, is the device pose vector, which contains position and attitude information. is the homogeneous transformation matrix used for coordinate system transformation, is the global coordinate vector, the converted device position, is the three-dimensional coordinate of the device in the global coordinate system, is the transpose symbol; S230, environmental sensor space mapping: calculating global coordinates based on sensor installation locations and establishing a sensor-device spatial relationship matrix; formula: Sensor global coordinates: ; ; in, is the position of the sensor in the global coordinate system, is the position of the device in the global coordinate system, is the device rotation matrix, which is composed of the yaw angle and pitch angle Calculated, The installation offset of the sensor relative to the device reference point; Distance Matrix: ; in, is the sensor number index, is the device number index, is the Euclidean norm, which calculates the distance between two points. is the global coordinate of the i-th sensor, is the global coordinate of the jth device, is the distance matrix, which represents the distance between all sensors and devices; S240, vibration spectrum spatiotemporal correlation: transform the vibration spectrum density vector Associated with specific device locations, establishing frequency-space distribution models; formula: Spatial distribution of vibration energy: ; in, For spatial points The vibration energy density at is the global coordinate of the vibration source, is the global coordinate of the sensor, is the spatial attenuation coefficient of the kth frequency component, Frequency The power spectral density at is the dimension of the spectrum vector, indicating the number of frequency components, is the three-dimensional spatial position in the global coordinate system; S250, spatiotemporal index construction: Building an R*-tree spatiotemporal index to achieve millisecond-level cross-modal data retrieval; Index key generation: ; in, To synchronize timestamps, is the three-dimensional coordinate of the device in the global coordinate system, It is a four-dimensional index key used to uniquely identify a spatiotemporal data point in the R* tree; Combine timestamp and global coordinates as multidimensional index keys; R* tree insertion operation: ; in, To associate device / sensor metadata, It is a multidimensional index key, including timestamp and three-dimensional coordinates. The insertion method of the R* tree is used to insert key-value pairs Insert into the appropriate position in the tree; Range query optimization: ; in, To query the time point, is the query start time, The query end time. is the spatial position vector, is the spatial cube area, is the total number of index nodes, is the query time complexity, For range query conditions, the time and space query ranges are defined to retrieve data points that meet the conditions in the R* tree; S300, digital twin disturbance modeling: This system establishes a vibration sensitivity model (β parameter) and an environmental disturbance model (α parameter) based on fused data. It uses the least squares method to fit equations, verifies model accuracy online, and outputs a pose drift prediction equation. In one embodiment of the present invention, the following steps are specifically included: S310, Vibration-Position Coupling Modeling: Analyze the quantitative relationship between vibration spectrum and equipment posture drift, and establish a frequency sensitivity matrix; formula: Posture drift caused by vibration: ; ; ; in, is the pose drift vector caused by vibration, is the frequency sensitivity matrix, is the vibration spectrum density vector, is the power spectrum density value of the first frequency point, is the power spectrum density value of the second frequency point, is the power spectrum density value of the Kth frequency point, is the transpose operation; Frequency sensitivity matrix: ; in, is the sensitivity coefficient of the j-th frequency band vibration to the i-th axial drift, for example, is the sensitivity coefficient of the first frequency band vibration to the x-axis drift, is the sensitivity coefficient of the second frequency band vibration to the x-axis drift, is the sensitivity coefficient of the K-band vibration to the x-axis drift, is the sensitivity coefficient of the first frequency band vibration to the y-axis drift, is the sensitivity coefficient of the second frequency band vibration to the y-axis drift, is the sensitivity coefficient of the K-band vibration to the y-axis drift, is the sensitivity coefficient of the first frequency band vibration to the z-axis drift, is the sensitivity coefficient of the second frequency band vibration to the z-axis drift, is the sensitivity coefficient of the K-band vibration to the z-axis drift, is the spectrum vector dimension, is the three-dimensional coordinate of the device in the global coordinate system, is the change in position components on the x, y, and z axes in the global coordinate system, is the variation operator; S320, Environmental-Sensor Error Modeling: Quantify the impact of environmental parameters on positioning / measurement errors and construct a temperature, humidity, and gas composite interference model; formula: Environmental comprehensive error: ; ; ; in, is the comprehensive error caused by environmental factors, is the temperature deviation, is the current temperature, is the reference temperature, is the humidity deviation, is the current humidity, For reference humidity, is the gas / carbon monoxide concentration, is the temperature interference coefficient, is the humidity interference coefficient, is the gas concentration interference coefficient, is the interference coefficient of carbon monoxide concentration; S330, comprehensive prediction of posture drift: Integrates vibration and environmental factors to predict device posture changes and build a state space model; formula: State space equations: ; ; in, is the device state vector (position + velocity), is the state vector of the device at the next moment, is the three-dimensional coordinate of the device in the global coordinate system, is the velocity component in the global coordinate system, is the state transfer matrix (describing the kinematic characteristics of the device), is the vibration disturbance input matrix, is the environmental perturbation input matrix, is Gaussian process noise (mean is 0, covariance is Q, and the unit is consistent with the state vector), is the discrete time index; S340, Perturbation Parameter Identification (OLS-Ridge Regression): Training Sensitivity Coefficients Based on Historical Data and interference coefficient ; formula: Optimization objective function: ; in, is the measured pose drift, is the measured environmental error, is the spectral density matrix, is the environmental parameter matrix, is the vibration sensitivity coefficient vector, is the environmental interference coefficient vector, is the regularization coefficient, is the Euclidean norm; S350, Model Validation and Confidence Assessment: Calculate prediction residuals and confidence intervals, and dynamically update model parameters; formula: Prediction residuals: ; 95% confidence interval: ; in, is the forecast residual at time t, is the mean residual, is the residual standard deviation, To verify the number of samples, For model prediction drift, is the measured drift, is the critical value of the standard normal distribution at the 95% confidence level, is the 95% confidence interval; S400, dynamic feedforward compensation control: predicts the posture drift based on the disturbance model, generates control signals in conjunction with a feedforward-PID composite controller, issues control commands to the actuators in real time, and verifies the compensation effect using measured data (calculating the drift suppression rate η and the residual energy ratio r); In one embodiment of the present invention, the following steps are specifically included: S410, disturbance prediction calculation, predicts posture drift based on the state space model of S330, integrating vibration and environmental disturbance factors; formula: Perturbation prediction equation: ; ; ; ; in, To predict pose drift, is the perturbation input matrix, is the vibration sensitivity coefficient matrix, is the vibration spectrum, is the environmental interference coefficient, is the environmental parameter vector ( , from S230, units are ℃, %, vol%, ppm) are the predicted pose drifts in the x, y, and z directions, respectively. is the power spectrum density value of the first frequency point, is the power spectrum density value of the second frequency point, is the power spectrum density value of the Kth frequency point, is the transpose operation, is the sensitivity coefficient of the j-th frequency band vibration to the i-th axial drift; S420, generating compensation control quantity: designing a PID controller to generate compensation instructions, introducing feedforward compensation to offset prediction drift; formula: PID-feedforward compound controller: ; ; ; in, To compensate the control quantity, is the pose tracking error, is the reference trajectory, is the global pose, is the PID gain matrix, is the feedforward compensation gain matrix, To control the cycle, are the control quantities in the x, y, and z directions respectively; S430, multi-objective collaborative optimization: resolve multi-device compensation conflicts and meet safety and energy constraints; formula: Optimization objective function: ; Constraints: ; in, is the target compensation amount, is the energy consumption weight coefficient, is the minimum safe distance between devices. is the maximum control threshold, is the position vector of devices i and j, is the kth control component; S440, compensation instruction issuance: the instruction is issued to the device PLC via the OPCUA protocol. The instruction format is standardized and the transmission delay requirement is ,in is the transmission delay time; formula: Instruction format: ; Transmission delay constraints: ; in, To standardize the instruction format, Is the unique identifier of the instruction, Enumeration value for instruction type (compensation / stop / ventilation, etc.). is the control vector, To synchronize timestamps, Receive timestamp for PLC, Send a timestamp to the system, is the transmission delay time; OPCUA communication parameters: ; in, The frequency of issuing instructions, which means the number of times instructions are sent per second. is the service quality level, 1 means "at least once" delivery guarantee, The communication timeout period. If the time is exceeded, the communication is considered to have failed. S450, Compensation Effect Verification: Compare the posture errors before and after compensation and calculate key performance indicators; formula: Drift suppression rate: ; Residual energy ratio: ; in, is the uncompensated measured drift, To compensate for the measured drift, is the drift suppression rate (expressed in percentage), is the residual energy ratio; S500, edge intelligent decision-making: evaluates compensation performance indicators (η, r), dynamically triggers model updates (RLS algorithm optimizes β, α) or parameter optimization (PSO adjusts PID gain), comprehensively analyzes safety status (methane concentration, vibration peak) to execute graded responses (shutdown / ventilation), and synchronizes decisions to the digital twin platform; In one embodiment of the present invention, the following steps are specifically included: S510, dynamic evaluation of compensation effect: Calculate key performance indicators (KPIs) based on the verification results of S450, trigger model updates and parameter optimization thresholds; formula: Drift suppression rate and residual energy ratio: ; Among them, η is the drift suppression rate, which represents the percentage of compensation effect, r is the residual energy ratio, which represents the ratio of position error to reference trajectory, is the uncompensated measured drift, To compensate for the measured drift, is the device global coordinate, is the reference trajectory coordinate, is the vector norm, which represents the modulus of the vector. is the Euclidean norm, which represents the 2 norm of the vector; Decision logic: ; S520, online update of disturbance model: real-time update of vibration sensitivity coefficient using recursive least squares (RLS) method and environmental interference coefficient ; formula: Recursive least squares update equation: ; ; ; ; ; in, is the current parameter vector, including the vibration sensitivity coefficient and the environmental interference coefficient, is the parameter vector at the previous moment, is the Kalman gain matrix, used to update parameter estimates, is the characteristic matrix, including vibration and environmental parameters (from S230-S240), is the observation vector at the current moment, is the forgetting factor, which is used to adjust the weight of historical data (the default is 0.98). is the current covariance matrix, which represents the uncertainty of parameter estimation, is the parameter estimation covariance matrix of the previous moment, where the initial value , is the identity matrix, is the time step index, is the vibration sensitivity coefficient, is the environmental interference coefficient; S530, collaborative optimization of control parameters: Optimizing PID gains through particle swarm optimization (PSO) and feedforward gain ; Optimization goal: ; ; Constraints: ; in, is the objective function value, which represents the control performance evaluation index, is the tracking error, To control the amount, To control the energy consumption weight, is the PID gain constraint boundary, is the maximum control threshold, To optimize the time window length, is the infinite norm, which represents the absolute value of the maximum component of the vector, are the proportional, integral, and differential gains of the PID controller; S540, Security Situation Decision: Executes security policies based on integrated environmental and vibration data, generating device-level / system-level emergency commands; Decision rules: ; in, is the methane concentration, is the peak value of the vibration spectrum, is the wind speed, is the power spectrum density at the kth frequency point, is the frequency point index; S550, Digital Twin Synchronization and Log: Synchronizes decision results to the digital twin platform, generates encrypted decision logs and security requirements, log integrity is SHA-256 hash check, and synchronization delay is , is the synchronization delay time; Log hash check: ; in, is the generated hash value (256 bits), is the decision parameter set ( wait), To synchronize timestamps, is the data concatenation symbol, is the SHA-256 hash function; Incremental synchronization protocol: ; in, For incremental data packets, is the parameter identifier, is the new value of the parameter, is the old value of the parameter, Index of the parameter Synchronization period constraints: ; in, is the synchronization period, is the model update frequency, is the maximum synchronization period threshold, is the minimum value function; S600, control command output: reconstructs the control signal (embedding optimization parameters and safety instructions), performs encryption encoding (AES-256 + CRC32 checksum) and timestamp synchronization, transmits the command according to priority via the real-time network (TSN), receives execution feedback, and calculates control error and delay to form a closed-loop verification; In one embodiment of the present invention, the following steps are specifically included: S610, control signal reconstruction: Fusion of the control parameters optimized in S530 and the real-time posture error generates a control signal, embedding feedforward compensation; formula: PID controller with feedforward compensation: ; in, is the output control vector, is the optimized PID gain matrix (from S530), To optimize the feedforward gain matrix (from S530), is the real-time pose error, To predict pose drift, To control the cycle, is the error integral term, is the error differential term; S620, security instruction embedding: Generate device-level security linkage instructions based on the security decision coverage control signal of S540; Logical AND parameters: Security Control Vectors: ; in, is the security control vector, is the ventilation control gain (default 0.5), is the methane concentration, is the original control vector, Zero control vector, used for emergency stop, To obtain the maximum value function, ensure that the ventilation control quantity is non-negative; S630, Signal Security Coding: Encrypt and verify the integrity of the control signal and add timestamp synchronization; Protocol and parameters: Encryption and verification: ; ; in, is the encrypted control vector (binary data), Dynamic session key (update cycle 1 hour, 256 bits), timestamps for spatiotemporal alignment, is the data concatenation symbol, It is an AES encryption function with a key length of 256 bits. It is a 32-bit cyclic redundancy check function; S640, real-time transmission scheduling: transmits encrypted instructions through TSN (time-sensitive networking) and dynamically assigns transmission priorities; Scheduling strategy: ; Among them, the transmission constraints, maximum delay, and bandwidth allocation are ; in, is the maximum transmission delay, To allocate bandwidth, is the packet length, is the transmission priority (7 is the highest); S650, execution status feedback: receiving the execution confirmation signal from the device PLC and calculating the control error and delay; formula: Control error and delay calculation: ; ; ; in, is the control error, which indicates the deviation of instruction execution. is the transmission delay, which indicates the time difference between instruction transmission and execution. is the control vector issued, The executed control quantity confirmed by the equipment, is the device response timestamp, To synchronize timestamps, is the Euclidean norm, which is used to calculate the size of the control error. They are the control quantities executed in the x, y, and z directions respectively.

[0019] In one embodiment of the present invention, the following operation example of a downhole multi-parameter monitoring system is also provided; The scenario is described as follows: In a deep coal mine, a roadheader is cutting the coal wall, and transport trucks are frequently transporting coal slag. At this point, equipment vibration, wind disturbance, and gas release cause dramatic fluctuations in environmental parameters, and the system initiates a dynamic compensation monitoring process: 1. Real-time perception of device status Tunnel boring machine position acquisition: UWB positioning tags capture the tunnel boring machine's position in real time (102.3m, 45.7m, -15.2m in the tunnel). LiDAR scans the surface deformation of the equipment and detects the offset of the cutting head. Environmental parameter synchronization: The gas sensor detected a sudden increase in CH4 concentration near the cutoff point, and the wind speed sensor showed local wind stagnation; Vibration signal processing: The accelerometer in the cutting section of the tunnel boring machine collected high-frequency vibrations, and spectrum analysis identified 85 Hz as the dominant frequency.

[0020] 2. Multi-source data fusion processing The spatiotemporal alignment engine unifies the tunnel boring machine's position, gas concentration, and vibration spectrum into a global coordinate system and associates them with the same timestamp; The fused data revealed that the position of the cutting head coincided with the high-concentration gas area, and the 85Hz vibration intensity was positively correlated with the equipment posture drift.

[0021] 3. Digital Twin Disturbance Modeling Vibration-posture modeling: 85Hz vibration was identified as the main cause of posture drift, and the sensitivity coefficient β was marked as high risk; Environmental interference analysis: Exceeding the limit of gas concentration leads to LiDAR ranging error, and the environmental interference coefficient α is dynamically corrected; The model outputs a warning: the cutting head posture will continue to deviate and may hit the side wall of the tunnel.

[0022] 4. Dynamic feedforward compensation control Predict the posture offset of the roadheader 1 second later, and use the feedforward-PID controller to generate the reverse compensation force; The multi-objective optimization module coordinates the ventilation system: increasing local fan speeds to dilute gas and stabilize airflow; The command is sent to the tunnel boring machine PLC: adjust the hydraulic arm angle to offset posture drift.

[0023] 5. Edge Intelligent Decision-Making Verification after compensation: The pose offset suppression rate reached 88%, but the residual energy ratio still exceeded the standard, triggering a model update.

[0024] The RLS algorithm optimizes the vibration sensitivity coefficient β, and the PSO adjusts the PID gain parameters; Safety decision: Due to the continuous increase in gas concentration, the "ventilation enhancement" command was activated and the backup fan was linked.

[0025] 6. Control instruction output Reconstruct the control signal: embed the optimized PID gains and ventilation commands; AES-256 encrypted instructions are transmitted to the device with high priority via the TSN network; Execution feedback: The tunnel boring machine's posture is stable, the ventilation system reduces the gas concentration to a safe threshold, and the transmission delay is only 18ms.

[0026] The above examples achieved the following results through the downhole multi-parameter environmental monitoring method: Risk mitigation: Equipment collision was successfully avoided, and gas concentration dropped by 50%-80% within 10 seconds.

[0027] System Adaptation: After the model is updated online, the efficiency of subsequent vibration interference compensation is improved by 35%.

[0028] Panoramic visualization: The digital twin platform displays equipment trajectories, risk heat maps, and control instruction flows in real time.

[0029] The above describes the embodiments of the present invention, but the present invention is not limited to the above specific implementation methods. The above specific implementation methods are merely illustrative and not restrictive. Ordinary technicians in this field can also make many forms under the guidance of the present invention, all of which are protected by the present invention.

Claims

1. A method for monitoring underground multi-parameter environment, characterized in that: The following steps are involved: S100, real-time device status perception: This system uses multi-source sensors to collect device posture, vibration spectrum, and environmental parameters in real time, performs spatiotemporal alignment, filtering, noise reduction, and feature extraction on the raw data, and outputs a standardized state vector. S200, multi-source data fusion processing: fusing posture data, vibration signals, and environmental parameters under a unified spatiotemporal reference to construct a multi-dimensional feature matrix; Use principal component analysis to reduce dimensionality and eliminate redundant information, and output the fused device state vector; S300, digital twin disturbance modeling: This system establishes vibration sensitivity models and environmental disturbance models based on fused data, fits equations using the least squares method, verifies model accuracy online, and outputs pose drift prediction equations. S400, dynamic feedforward compensation control: predicts posture drift based on the disturbance model, generates control signals using a feedforward-PID composite controller, issues control commands to actuators in real time, and verifies compensation effects using measured data; S500, edge intelligent decision-making: evaluates compensation performance indicators, dynamically triggers model updates or parameter optimization, performs hierarchical responses based on comprehensive security situation, and synchronizes decisions to the digital twin platform; S600, control command output: reconstruct the control signal, perform encryption encoding and timestamp synchronization, transmit the command according to priority through the real-time network, receive execution feedback and calculate the control error and delay to form a closed-loop verification.

2. A method for monitoring underground multi-parameter environment according to claim 1, characterized in that: In S100, the following steps are included: S110, equipment posture acquisition: uses UWB base stations and tags to capture the 3D coordinates and posture angles of the tunnel boring machine and transport vehicle in real time, LiDAR scans the area point cloud data, and dynamically models the surface deformation of the equipment; S120, environmental parameter collection: synchronously reads gas / temperature and humidity / wind speed sensor data, and the air pressure sensor monitors the internal pressure of the wind tube; S130, vibration signal acquisition: accelerometers are deployed at key locations on the equipment to collect 0-500Hz raw vibration signals; S140, vibration spectrum calculation: perform FFT transformation on the original acceleration signal and calculate the power spectrum density; S150, data encapsulation and timestamp: packages pose / environment / spectrum data and adds PTP1588v2 protocol timestamp.

3. The method for monitoring underground multi-parameter environment according to claim 1, characterized in that: In S200, the following steps are included: S210, PTP1588v2 time synchronization: aligns all device / sensor time bases through a master-slave clock architecture, calculates clock offsets and corrects timestamps; S220, device coordinate system → roadway global coordinate system conversion: convert the device posture to the roadway global coordinate system, and integrate the LiDAR point cloud data to correct the positioning error; S230, environmental sensor space mapping: calculating global coordinates based on sensor installation locations and establishing a sensor-device spatial relationship matrix; S240, vibration spectrum spatiotemporal correlation: Correlate the vibration spectrum density vector to a specific device location and establish a frequency-space distribution model; S250, spatiotemporal index construction: Establish an R*-tree spatiotemporal index to achieve millisecond-level cross-modal data retrieval.

4. A method for monitoring underground multi-parameter environment according to claim 1, characterized in that: In S300, the following steps are included: S310, Vibration-Position Coupling Modeling: Analyze the quantitative relationship between vibration spectrum and equipment posture drift, and establish a frequency sensitivity matrix; S320, Environmental-Sensor Error Modeling: Quantify the impact of environmental parameters on positioning / measurement errors and construct a temperature, humidity, and gas composite interference model; S330, comprehensive prediction of posture drift: Integrates vibration and environmental factors to predict device posture changes and build a state space model; S340, disturbance parameter identification: training sensitivity coefficient and interference coefficient based on historical data; S350, Model Validation and Confidence Assessment: Calculate prediction residuals and confidence intervals, and dynamically update model parameters.

5. A method for monitoring underground multi-parameter environment according to claim 4, characterized in that: The calculation formula for vibration-posture coupling modeling is as follows: Posture drift caused by vibration: ； ； ； in, is the pose drift vector caused by vibration, is the frequency sensitivity matrix, is the vibration spectrum density vector, is the power spectrum density value of the first frequency point, is the power spectrum density value of the second frequency point, is the power spectrum density value of the Kth frequency point, is the transpose operation; Frequency sensitivity matrix: ； in, is the sensitivity coefficient of the j-th frequency band vibration to the i-th axial drift, is the sensitivity coefficient of the first frequency band vibration to the x-axis drift, is the sensitivity coefficient of the second frequency band vibration to the x-axis drift, is the sensitivity coefficient of the K-band vibration to the x-axis drift, is the sensitivity coefficient of the first frequency band vibration to the y-axis drift, is the sensitivity coefficient of the second frequency band vibration to the y-axis drift, is the sensitivity coefficient of the K-band vibration to the y-axis drift, is the sensitivity coefficient of the first frequency band vibration to the z-axis drift, is the sensitivity coefficient of the second frequency band vibration to the z-axis drift, is the sensitivity coefficient of the K-band vibration to the z-axis drift, is the spectrum vector dimension, is the three-dimensional coordinate of the device in the global coordinate system, is the change in position components on the x, y, and z axes in the global coordinate system, is the variation operator.

6. A method for monitoring underground multi-parameter environment according to claim 5, characterized in that: The calculation formula for environment-sensor error modeling is as follows: Environmental comprehensive error: ； ； ； in, is the comprehensive error caused by environmental factors, is the temperature deviation, is the current temperature, is the reference temperature, is the humidity deviation, is the current humidity, For reference humidity, is the gas / carbon monoxide concentration, is the temperature interference coefficient, is the humidity interference coefficient, is the gas concentration interference coefficient, is the interference coefficient of carbon monoxide concentration.

7. A method for monitoring downhole multi-parameter environment according to claim 6, characterized in that: The calculation formula for comprehensive prediction of pose drift is as follows: State space equations: ； ； in, is the device state vector, is the state vector of the device at the next moment, is the three-dimensional coordinate of the device in the global coordinate system, is the velocity component in the global coordinate system, is the state transition matrix, is the vibration disturbance input matrix, is the environmental perturbation input matrix, is Gaussian process noise, is a discrete time index.

8. A method for downhole multi-parameter environmental monitoring according to claim 7, characterized in that: In S400, the following steps are included: S410, disturbance prediction calculation, predicts posture drift based on the state space model of S330, integrating vibration and environmental disturbance factors; S420, generating compensation control quantity: designing a PID controller to generate compensation instructions, introducing feedforward compensation to offset prediction drift; S430, multi-objective collaborative optimization: resolve multi-device compensation conflicts and meet safety and energy constraints; S440, compensation instruction issuance: The instruction is issued to the device PLC via the OPCUA protocol. The instruction format is standardized and the transmission delay requirement is <20ms. S450, Compensation Effect Verification: Compare the posture errors before and after compensation and calculate key performance indicators.

9. A method for monitoring underground multi-parameter environment according to claim 8, characterized in that: In S500, the following steps are included: S510, dynamic evaluation of compensation effect: Calculate key performance indicators based on the verification results of S450, trigger model update and parameter optimization threshold; S520, online update of disturbance model: using recursive least squares method to update vibration sensitivity coefficient and environmental interference coefficient in real time; S530, collaborative optimization of control parameters: Optimize PID gains and feedforward gains through particle swarm optimization; S540, Security Situation Decision: Executes security policies based on environmental and vibration data, generating device-level / system-level emergency instructions; S550, Digital Twin Synchronization and Log: Synchronizes decision results to the digital twin platform, generates encrypted decision logs and security requirements, log integrity is SHA-256 hash check, and synchronization delay is synchronization delay time <50ms.

10. A method for monitoring underground multi-parameter environment according to claim 9, characterized in that: In S600, the following steps are included: S610, control signal reconstruction: Fusion of the control parameters optimized in S530 and the real-time posture error generates a control signal, embedding feedforward compensation; S620, security instruction embedding: Generate device-level security linkage instructions based on the security decision coverage control signal of S540; S630, Signal Security Coding: Encrypts and verifies the integrity of control signals and adds timestamp synchronization; S640, real-time transmission scheduling: transmits encrypted instructions via TSN and dynamically assigns transmission priorities; S650, execution status feedback: receiving the execution confirmation signal from the device PLC and calculating the control error and delay.

Images