Monitoring and intelligent decision-making system for influence of holographic blasting vibration on slope stability

Abstract

The invention relates to the technical field of monitoring and intelligent decision analysis, and discloses a monitoring and intelligent decision system for the influence of holographic blasting vibration on slope stability, and the system comprises a data collection module, a data processing module, an intelligent decision module, and a feedback control module. The method comprises the following steps: dynamically deploying a three-dimensional vibration sensor array based on topographic features; collecting and preprocessing blasting vibration signals and geological data; fusing the physical model and the data driving model to predict vibration propagation; calculating a dynamic safety coefficient and generating a multi-objective optimization scheme; and outputting an early warning instruction and feeding back and adjusting the sensor network configuration. Through a three-dimensional dynamic monitoring network, real-time data processing, a hybrid drive prediction model, multi-target quantitative evaluation and hierarchical early warning regulation and control, the problems of incomplete sensor coverage, early warning delay, prediction deviation, evaluation deficiency and poor scene adaptability are solved, and accurate monitoring and intelligent decision making of blasting vibration influence are realized.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of monitoring and intelligent decision analysis, in particular to a monitoring and intelligent decision system for the influence of holographic blasting vibration on slope stability. BACKGROUND

[0002] Blasting operation is indispensable in mine exploitation, tunnel construction and other scenarios, and the vibration effect caused by blasting operation poses a potential threat to the stability of the surrounding rock mass. With the extension of engineering scale to deep and complex geological areas, traditional monitoring methods face higher requirements for precision and real-time performance. Existing technologies mainly focus on vibration waveform collection and simple threshold alarm, but lack support for key issues such as three-dimensional vibration field reconstruction and multi-physical field coupling analysis. Although there has been significant progress in the miniaturization of sensing nodes and wireless transmission technology in recent years, the core needs such as elimination of monitoring blind spots in complex rock mass structures and long-term stability prediction have not been effectively met.

[0003] The current mainstream solution involves deploying single-point high-precision vibration meters to infer the overall vibration distribution from local measurement data, but it may overlook the influence of slope structure heterogeneity. The vibration attenuation model based on empirical formula has significant prediction deviation in layered rock mass, multi-fracture zone and other scenarios. The introduction of machine learning algorithms for optimization of prediction is limited by small sample data and insufficient coupling with physical laws, and has weak generalization ability.

[0004] The rigid sensor layout leads to frequent monitoring dead angles, such as missing data in key areas such as slope shoulders and potential slip surfaces. The data processing process is lengthy, and the risk assessment often lags behind by tens of minutes from signal collection, which cannot support real-time regulation and control in blasting operations. The prediction model relies too much on simplified assumptions, and the description of nonlinear effects such as joint surface slip and rock mass damage accumulation is rough, resulting in low long-term prediction reliability. These defects result in the contradiction between "complete monitoring and delayed warning" or "sufficient data and inaccurate decision-making" in engineering, which seriously hinders the safety control efficiency in high-risk scenarios.

[0005] Therefore, the present application proposes a monitoring and intelligent decision system for the influence of holographic blasting vibration on slope stability to solve the problems of the prior art. SUMMARY

[0006] In view of the deficiencies of the prior art, the present application provides a monitoring and intelligent decision system for the influence of holographic blasting vibration on slope stability, which solves the problems of data fragmentation, real-time warning response delay, low prediction accuracy in complex geology, missing quantitative evaluation model and poor adaptability of blasting parameters to slope characteristics caused by incomplete sensor deployment in the prior art.

[0007] To achieve the above purpose, the present application realizes the following technical scheme: a monitoring and intelligent decision system for the influence of holographic blasting vibration on slope stability, comprising:

[0008] A data acquisition module is configured to acquire slope blasting vibration signals in real time through a three-dimensional vibration sensor array.

[0009] A data processing module is connected to the data acquisition module and configured to perform multi-scale feature extraction and fusion analysis on the vibration signals.

[0010] An intelligent decision-making module is connected to the data processing module and configured to generate a blasting parameter optimization scheme and a safety warning instruction based on the fusion analysis results.

[0011] A feedback control module is connected to the intelligent decision-making module and configured to execute the warning instruction and dynamically adjust a sensor network deployment strategy.

[0012] Preferably, the data acquisition module comprises:

[0013] A sensor deployment unit is configured to generate a honeycomb monitoring grid on the slope surface based on a Delaunay triangulation algorithm, dynamically deploy three-dimensional piezoelectric vibration sensors, and dynamically adjust the sensor density.

[0014] A signal conditioning unit is configured to filter and process the original vibration signals in the 0.5Hz-200Hz frequency band using a band-pass filter.

[0015] A multi-mode transmission unit is configured to transmit data using a RS-485 and LoRa dual-channel heterogeneous communication protocol.

[0016] Preferably, the sensor density in the sensor deployment unit satisfies the formula:

[0017] ;

[0018] wherein, is the number of sensors, is the area of the monitoring region, is the real-time risk assessment coefficient, is the effective monitoring radius of the sensor.

[0019] Preferably, the data processing module comprises:

[0020] A feature extraction unit is configured to extract time-frequency energy features of the vibration signals using wavelet packet decomposition.

[0021] A physical modeling unit is configured to construct a blasting vibration propagation model based on an improved Sadovsky formula.

[0022] A data fusion unit is configured to fuse the vibration features and ground penetrating radar data using a weighted tensor.

[0023] Preferably, the improved Sadovsky formula in the physical modeling unit is:

[0024] ;

[0025] in, For site coefficient, This is the equivalent amount of explosive. For the distance between the centers, The decay exponent, The elevation difference influence coefficient. The medium absorption coefficient, For joint correction factor, This is a rock quality indicator.

[0026] Preferably, the data fusion unit achieves vibration-geological feature fusion through a learnable weight tensor, satisfying the following:

[0027] ;

[0028] in, This is a geological feature matrix. The characteristic matrix of the vibration spectrum. and The weight tensor is obtained through training an adversarial generative network. It represents the Hadamah accumulation. This indicates a channel splicing operation.

[0029] Preferably, the intelligent decision-making module includes:

[0030] Risk assessment unit, calculates dynamic safety factor :

[0031] ;

[0032] in, For the cohesion of rock strata, For normal stress, It is the internal friction angle. For shear stress components, To accumulate vibrational energy, This is the critical energy threshold.

[0033] Preferably, the risk assessment unit further includes:

[0034] The multi-objective optimization unit uses the NSGA-III algorithm to solve the following objective function:

[0035] ;

[0036] Constraints include safety factor thresholds Lower limit of hole distance ;

[0037] in, , These are the weighting coefficients. The peak vibration velocity of the particle. This represents the blasting cost function. For the amount of explosives, The hole spacing, To delay time, This is the cost coefficient.

[0038] Preferably, the feedback control module includes:

[0039] The early warning classification unit calculates the early warning level based on the fuzzy membership function:

[0040] ;

[0041] in, This represents the peak vibrational velocity of the particle.

[0042] Dynamically deployed units based on real-time risk coefficients Adjust the sensor density and position.

[0043] This invention also provides a method for monitoring and intelligent decision-making regarding the impact of holographic blasting vibration on slope stability, comprising the following steps:

[0044] S1. Dynamically deploy a three-dimensional vibration sensor array based on terrain features;

[0045] S2. Collect and preprocess blasting vibration signals and geological data;

[0046] S3. Predicting vibration propagation by integrating physical models and data-driven models;

[0047] S4. Calculate the dynamic safety factor and generate a multi-objective optimization scheme;

[0048] S5 outputs warning commands and provides feedback to adjust the sensor network configuration.

[0049] This invention provides a monitoring and intelligent decision-making system for the impact of holographic blasting vibration on slope stability. It has the following beneficial effects:

[0050] 1. This invention employs a dynamic deployment method for a three-dimensional sensor array based on terrain features, combined with an adaptive density adjustment mechanism, to achieve comprehensive coverage of the area affected by blasting vibrations. Compared to existing fixed sensor layout schemes, this effectively solves the problems of incomplete coverage of monitoring blind spots and severe data fragmentation, significantly improving the comprehensiveness and accuracy of data acquisition, and providing a complete data foundation for slope stability analysis.

[0051] 2. This invention achieves real-time acquisition and rapid analysis of blasting vibration signals through the collaborative design of a high-performance embedded processor and a multi-level signal processing link. Compared with the data analysis mode that relies on offline processing in the prior art, it breaks through the technical bottlenecks of high data processing latency and poor early warning timeliness, and can identify potential risks in real time during blasting operations, greatly improving on-site safety management capabilities.

[0052] 3. This invention constructs a blasting vibration propagation prediction model that integrates physical laws and data characteristics, combined with a dynamic correction mechanism for complex geological parameters, achieving a high-precision prediction of vibration propagation patterns. Compared to existing prediction methods using single empirical formulas, this significantly improves the prediction bias problem under complex rock mass structures and multi-fracture geological conditions, providing a scientific basis for long-term slope stability assessment.

[0053] 4. This invention innovates a dynamic safety factor calculation model and a multi-objective optimization algorithm, enabling precise quantitative assessment of the blasting impact range and intensity. Compared with the limitations of existing technologies that rely on human experience for decision-making, it solves the technical defects of traditional methods that struggle to balance safety constraints and economic costs, generating optimized blasting parameter schemes that take into account both safety and construction efficiency, thus improving the scientific nature of engineering decision-making.

[0054] 5. This invention constructs a customized early warning rule base based on slope safety level and geological characteristics. Through a dynamic parameter adjustment mechanism, it achieves the technical effect of differentiated and precise control. Compared with the general early warning strategies in existing technologies, it effectively solves the problem of insufficient adaptability to special slope scenarios, significantly enhances the system's response capability to complex engineering environments, and provides targeted protection measures for slopes of different safety levels. Attached Figure Description

[0055] Figure 1 This is a system architecture diagram of the present invention;

[0056] Figure 2 This is a flowchart of the method of the present invention. Detailed Implementation

[0057] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0058] Please see Figure 1 This invention provides a monitoring and intelligent decision-making system for the impact of holographic blasting vibration on slope stability, comprising:

[0059] The data acquisition module is used to acquire slope blasting vibration signals in real time through a three-dimensional vibration sensor array;

[0060] In this embodiment, the data acquisition module achieves accurate capture of slope blasting vibration signals through a multi-dimensional sensing architecture and dynamic control mechanism. Its technical solution is technically coupled with the feature extraction unit of the subsequent data processing module, providing the system with raw data input with a high signal-to-noise ratio.

[0061] For sensor deployment units, the installation location of a three-dimensional piezoelectric vibration sensor is typically determined using the Delaunay triangulation algorithm, generating a honeycomb-shaped monitoring grid. Specifically, the side length of each grid cell... The calculation formula is:

[0062] ;

[0063] in, The area of ​​the monitoring area (unit: m) 2 ), The initial number of sensors (based on slope height) according to (Estimation).

[0064] As an alternative, when significant cracks or weak interlayers exist on the slope surface, the high-risk areas are locally densified during the mesh generation process, with a densification coefficient... .

[0065] The dynamic density adjustment formula is:

[0066] ;

[0067] in: The real-time risk assessment coefficient is calculated using the following formula:

[0068] ;

[0069] in, This represents the historical landslide probability (based on statistics of accident records in the area over the past 10 years). Current cumulative vibration energy (unit: mm) 2 / s 2 ), The critical energy threshold (calibrated through rock fatigue tests); The effective monitoring radius of the sensor (unit: m), and the sensor sensitivity. and environmental noise level Related:

[0070] ;

[0071] Example parameters: mV / g At dB, m.

[0072] In one possible implementation, the sensor mounting bracket is equipped with a tilt compensation mechanism, the deflection angle of which... satisfy:

[0073] ;

[0074] in, The local elevation difference of the mounting surface (unit: mm). The bracket length is fixed at 120mm. This design ensures that the sensor's three axes are aligned with the slope normal, minimizing installation errors. .

[0075] The signal conditioning unit, specifically, the signal processing chain comprises four cascaded circuits:

[0076] Preamplifier: AD620 instrumentation amplifier, gain dB, input noise density ;

[0077] Bandpass filter: 8th order Butterworth active filter, passband range 0.5-200Hz, stopband attenuation slope dB / oct, passband ripple dB;

[0078] Power frequency notch filtering: dual Type of active notch filter, center frequency Hz (adaptive switching), quality factor notch depth dB;

[0079] Analog-to-digital conversion: The AD7606 chip enables 16-bit resolution sampling.

[0080] The sampling rate during the blasting phase is set as follows:

[0081] ;

[0082] in, Real-time vibration velocity (unit: mm / s), calculated using a moving average algorithm (window length). Calculated in milliseconds (ms).

[0083] As an option, a multi-mode transmission unit employs a dual-channel heterogeneous communication protocol combining RS-485 and LoRa, with the following parameter configuration:

[0084] RS-485 link:

[0085] Baud rate adaptive adjustment rules:

[0086] ;

[0087] in, To activate the number of sensors, bits (data size per sample). Data acquisition period (default 100ms);

[0088] Data frame structure: Start bit (1 bit) + Data bits (8 bits) + Check bit (CRC-16) + Stop bit (2 bits), single frame transmission delay ms;

[0089] LoRa link:

[0090] Physical layer parameters: center frequency 470MHz, bandwidth kHz, spreading factor coding rate ;

[0091] Transmit power Dynamic adjustment:

[0092] ;

[0093] in, dBm is the reference power. Transmission distance (unit: m). m is the reference distance;

[0094] Data packet encapsulation format: preamble (4 bytes) + frame header (2 bytes) + sensor ID (4 bytes) + timestamp (8 bytes) + triaxial data (3×4 bytes) + CRC (2 bytes), with a fixed packet length of 24 bytes.

[0095] In an extended embodiment, the sensor deployment unit integrates a temperature drift compensation module, whose correction formula is:

[0096] ;

[0097] in: This is the vibration velocity value after temperature compensation; The original vibration velocity measurement is the uncompensated value. The linear temperature coefficient is -0.02% / ℃, determined through sensor calibration tests. The second temperature coefficient (value 5 × 10) -6 / ℃ 2 ); Real-time temperature (unit: °C). ℃ is the reference temperature.

[0098] In another embodiment, the multimode transmission unit supports adaptive routing switching based on link quality: when the signal-to-noise ratio (SNR) of the LoRa link is <6dB and the RS-485 bit error rate (BER) is >10... -4 When relay node forwarding is enabled, the relay selection rule is as follows:

[0099] ;

[0100] in, For the first Received signal strength of each relay node (unit: dBm). The expected number of transmissions (estimated using historical transmission success rates).

[0101] The data processing module, connected to the data acquisition module, is used for multi-scale feature extraction and fusion analysis of vibration signals;

[0102] In this embodiment, the data processing module performs in-depth analysis of the vibration signals and geological parameters input by the data acquisition module through multi-level feature analysis and multi-physics field coupling analysis, forming a fusion feature space for slope stability assessment.

[0103] The feature extraction unit typically employs wavelet packet decomposition to extract the time-frequency energy distribution features of the vibration signal. Specifically, the db4 wavelet basis function is selected for 5-level decomposition, dividing the 0-200Hz frequency band into 16 sub-bands (each level of decomposition generates 2...). n There are 4 subbands (n=4). The energy characteristics of each subband are calculated as follows:

[0104] ;

[0105] in, sensor nodes The cumulative value of vibration energy at the location; Indicates the first The sensor at the first Individual band ( The decomposition coefficients of ). This is the length of the time window, which is set to 100ms by default.

[0106] In one possible implementation, the energy characteristics are dynamically normalized:

[0107] ;

[0108] in, and The first The mean and standard deviation of the energy of each element over a sliding time window (24 hours) are updated every minute.

[0109] As an alternative, the physical modeling unit employs a modified Sadovsky formula to construct a blasting vibration propagation model, the expression of which is:

[0110] ;

[0111] in, To predict the peak vibration velocity of particles; This represents the maximum amount of explosives in a single section. The distance between the centers of the explosion; (Site coefficient): By arranging 5 sets of reference explosion sources in the monitoring area ( The calibration is performed using the following formula:

[0112] ;

[0113] in, This represents the measured peak vibration velocity.

[0114] (Decrease Index) (Elevation difference influence coefficient): Determined through multivariate logarithmic regression of data from different distance sensors, satisfying:

[0115] ;

[0116] in, These are the fitting constants;

[0117] (Medium absorption coefficient): The value is taken according to the rock mass type; for granite, it is taken as 0.004m. -1 0.006m of sandstone was taken. -1 The loose accumulation layer is taken as 0.008m. -1 ;

[0118] (Joint correction factor): determined by joint density (Unit: strips / m) 2 The calculation is as follows:

[0119] ;

[0120] (Rock quality indicators): Obtained by calculating the fracture ratio using ground-penetrating radar scanning data, and meeting the following requirements:

[0121] ;

[0122] In the layered rock mass scenario, an equivalent vibration velocity model is adopted:

[0123] ;

[0124] in, The vibration velocity is calculated using a layered weighted average. For the first Thickness of rock layers (unit: m). The total thickness of the rock mass The predicted vibration velocity for the corresponding layer; This represents the total number of rock mass layers (integer type), based on geological profiles.

[0125] The data fusion unit, specifically, constructs a geological-vibration feature tensor fusion space, mathematically expressed as:

[0126] ;

[0127] in: This is a geological feature matrix, containing Longitudinal wave velocity Porosity Moisture content Joint density Geostress (Unit: MPa); This represents the vibration characteristic matrix, which includes the peak value of the triaxial vibration velocity and the dominant frequency. Frequency band energy ratio (Energy ratio of high-frequency band 50-200Hz to low-frequency band 0.5-50Hz); The learnable weight tensor is obtained through training an adversarial generative network. The generator network consists of three fully connected layers (128-64-d neurons) with the SELU activation function. The discriminator network consists of 1D convolutional layers (kernel size 5, channels 64) and max pooling layers. This represents the Hadamard product (element-by-element multiplication). This indicates a concatenation operation along the feature dimension.

[0128] In one extended embodiment, a multi-head attention mechanism is introduced to dynamically adjust the weights:

[0129] ;

[0130] in, , To query the matrix and the key matrix, This is a trainable parameter matrix.

[0131] As an enhancement to the core solution, the data processing module supports the following extended functions:

[0132] An online incremental learning mechanism triggers online parameter updates when the proportion of newly added monitoring data exceeds 15% of historical data.

[0133] ;

[0134] in, No. The model parameters (tensors) for each iteration are updated using gradient descent. For prediction error, The total number of training samples, The peak vibration velocity of the particle predicted by the model. This refers to the peak vibration velocity of the particle measured on-site. For physical constraints, The physical field output by the neural network. For spatial gradient operators, These are the right-hand terms of equations relating known physical laws (such as wave equations and constitutive equations), with units matching the physical field. The adaptive learning rate is set to 0.001 initially, decreasing by 50% every 1000 steps.

[0135] Anomaly suppression module, outlier detection based on Mahalanobis distance:

[0136] ;

[0137] in, The Mahalanobis distance; For real-time feature vectors; The historical mean of the eigenvectors, The covariance matrix is ​​calculated using a sliding window (1 hour in length); when At that time, apply a truncation weight to the eigenvector:

[0138] ;

[0139] in, The corrected feature vector; weight coefficients , For degrees of freedom Confidence level The critical value of the chi-square distribution.

[0140] The intelligent decision-making module, connected to the data processing module, is used to generate blasting parameter optimization schemes and safety warning instructions based on the fusion analysis results.

[0141] In this embodiment, the intelligent decision-making module generates blasting control commands based on multi-source feature fusion results and real-time monitoring data through quantitative risk assessment and multi-objective optimization algorithms, forming a closed-loop decision-making chain from data analysis to execution feedback.

[0142] Risk assessment unit, specifically, dynamic safety factor The calculation formula is:

[0143] ;

[0144] in, For the first The cohesion of the rock mass (unit: MPa) was obtained through in-situ direct shear tests. The normal stress range of the tests was 0.5-2 MPa, and the loading rate was 0.5 mm / min. The normal stress component (unit: MPa) is measured directly by an embedded geostress sensor or calculated based on finite element inversion, with mesh size... m; The internal friction angle (unit: °) is determined by fitting the Mohr-Coulomb envelope based on triaxial test data. The confining pressure range is 1-5 MPa, and the fitting residual is... ; The shear stress component (unit: MPa) is calculated based on the spatial gradient of the vibration velocity field. The calculation formula is as follows:

[0145] ;

[0146] in, This is the dynamic friction coefficient of the rock mass (default 0.35-0.55). In terms of spatial coordinate direction; Cumulative vibration energy (unit: kJ / m) 3 The calculation formula is:

[0147] ;

[0148] in, Density of rock mass (unit: kg / m³) 3 ), ms is the length of the integration window; This refers to the real-time vibration velocity of the particles. This is the start time of the integration process; Critical energy threshold (unit: kJ / m³) 3 Calibration was achieved through rock fatigue testing, with test conditions of 10 Hz frequency and 10 cycles. 6 Next, granite was taken. Sandstone .

[0149] In one possible implementation, when the joint surface slips and displaces... When the cohesion is at the mm level, dynamic reduction is applied:

[0150] ;

[0151] in, Measurements were taken using a distributed fiber optic sensor (sampling rate 1kHz), with an accuracy of ±0.1mm.

[0152] As an alternative, the NSGA-III algorithm is used to solve the following multi-objective optimization problems:

[0153] ;

[0154] in, The weight coefficients satisfy the normalization constraint. The default configuration is The safety level of the project will be dynamically adjusted according to its requirements.

[0155] The blasting cost function;

[0156] in: Single-hole charge (unit: kg), range 50-500 kg, step size 10 kg; The hole spacing (unit: m) has a lower limit constraint. The upper limit constraint is the maximum span of the drilling equipment; The differential delay time (unit: ms) ranges from 25 to 100 ms, with an accuracy of ±0.5 ms. Yuan / kg Yuan / m Price per millisecond (RMB / ms), updated regularly based on market unit price.

[0157] Constraints:

[0158] Safety constraints: When the slope has a bedding structure, the value is increased to 1.35; vibration constraint: mm / s, corresponding to GB-6722-2014 Explosion Safety Regulations;

[0159] Geometric constraints: , Drilling depth (unit: m).

[0160] Algorithm configuration:

[0161] Population size: 200, individual coding dimension: 3 ( );

[0162] Reference point generation: 12 reference directions are defined using the Das-Dennis method, and the hyperplane parameters are... ;

[0163] Crossover operation: Simulate binary crossover, with a distribution exponent of 20 and a probability of 0.9;

[0164] Mutation operation: multinomial mutation, distribution exponent 30, probability 0.1;

[0165] Termination criteria: Pareto frontier improvement rate <1% for 10 generations or maximum iterations of 200 generations.

[0166] In one extended embodiment, a fuzzy early warning rule base is constructed, whose membership function is defined as:

[0167] ;

[0168] ;

[0169] in, For dynamic safety factor , Peak particle velocity (unit: mm / s). As an enhancement to the core scheme, a dynamic adjustment strategy for blasting parameters is supported:

[0170] ;

[0171] in, For dynamic adjustment of the charge amount; This represents the current charge per hole; For each warning level (Level 1: Blue, Level 2: Orange, Level 3: Red), the step size limit is adjusted to ±30%; if the optimization fails to converge after three consecutive iterations, the mesh parameters are reconstructed.

[0172] ;

[0173] in, The optimized blast hole spacing; This represents the current spacing between blasting holes; This represents the current safety factor; the factor is... For experience scaling factor, exponent Used to balance the safety factor and the sensitivity of the hole spacing.

[0174] The feedback control module, connected to the intelligent decision-making module, is used to execute early warning commands and dynamically adjust the sensor network deployment strategy;

[0175] In this embodiment, the feedback control module executes a multi-level response strategy and dynamically optimizes the sensor network topology based on real-time risk assessment results and early warning instructions, forming a closed-loop control system of monitoring-execution-adaptation.

[0176] The early warning classification unit, specifically, calculates the early warning level using an S-shaped membership function:

[0177] ;

[0178] in, The peak vibration velocity of the mass particles (unit: mm / s) is extracted from the fused feature matrix output by the data processing module, with an accuracy of ±0.1 mm / s; The slope adjustment coefficient of the function is determined by fitting historical accident data (which determines the steepness of the transition zone). This is the threshold baseline value.

[0179] In one possible implementation, the rules for classifying warning levels are as follows:

[0180] Red Alert: or dynamic safety factor ;

[0181] Orange alert: and ;

[0182] Blue Alert: and .

[0183] threshold , Based on ROC curve analysis, the corresponding false alarm rate is ≤5% and the false negative rate is ≤2%.

[0184] As an option, dynamically deploying units allows for dynamic adjustment of sensor network density using the following formula:

[0185] ;

[0186] in, The real-time risk coefficient is calculated as follows: The value range is [0, 1.5], and the coefficient is... , Determined through principal component analysis; The effective monitoring radius of the sensor (unit: m), and the signal-to-noise ratio. The relationship (unit: dB) is as follows:

[0187] ;

[0188] in, dB is the lower limit of the identifiable signal. At dB, m; The area of ​​the monitoring area (unit: m) 2 The grid edge length correction formula is calculated in real time using the vertex coordinates of the Delaunay triangular mesh:

[0189] ;

[0190] in, The adjusted spacing between blast holes; The initial grid side length, This represents the real-time risk coefficient.

[0191] Sensor location update strategies include:

[0192] High-risk areas focus: local authorities At that time, a Voronoi diagram is generated with the risk point as the center. The newly added sensors are deployed at the vertices of the Thiessen polygon, and the vertex spacing satisfies:

[0193] ;

[0194] in, Set minimum spacing (in meters) for the sensors to ensure the effectiveness of the monitoring network coverage; The radius of influence of blasting vibration (unit: m) represents the maximum effective range predicted based on the vibration attenuation model. The conservative spacing threshold is set under an empirical safety factor to prevent monitoring blind spots.

[0195] Inefficient node recovery: energy recovery from vibration over 3 consecutive hours For edge nodes, triggering sleep commands reduces energy consumption, with sleep nodes accounting for no more than 20%;

[0196] Communication link optimization: based on signal strength between nodes (Unit: dBm) The minimum spanning tree is dynamically constructed based on the terrain elevation difference, and the link weights are calculated as follows:

[0197] ;

[0198] in, For nodes and Elevation difference (unit: m), coefficient The test calibration is affected by the terrain undulations.

[0199] In one extended embodiment, a fault self-recovery mechanism is configured:

[0200] Node failure detection: When the node heartbeat packet loss rate is >30% and lasts for 5 minutes, or the baseline offset is... mm / s, identified as a faulty node;

[0201] Backup node activation: Wake up the nearest backup node based on the priority list. The priority calculation formula is as follows:

[0202] ;

[0203] in, The coordinates of the fault node are shown in meters. For the coordinates of the spare node, This represents the real-time risk coefficient.

[0204] As an enhancement to the core solution, it supports multi-level early warning and response strategies:

[0205] Red Alert: Immediately cut off the power supply circuit of the blasting detonator (response delay ≤ 50ms), activate the audible and visual alarm (frequency 2Hz, sound intensity 90dB@1m), and activate the slope surface displacement monitoring radar (sampling rate 100Hz).

[0206] Orange alert: Limit single-hole charge to 70% of current value; differential delay time adjusted to... And activate the drone inspection (flight altitude 30m, resolution 5cm).

[0207] Blue alert: Sensor calibration command generated, baseline offset detection triggered.

[0208] ;

[0209] in, The average absolute deviation of the reference vibration velocity; The reference vibration signal generated by the exciter (amplitude 0.5 mm / s, frequency 10 Hz). For the first Vibration velocity at the secondary verification measurement point.

[0210] In another embodiment, temperature compensation and humidity correction are introduced:

[0211] Temperature compensation: The sensor sensitivity correction formula is:

[0212] ;

[0213] in, The nominal sensitivity is... For temperature coefficient, For reference temperature;

[0214] Humidity correction: Signal attenuation is calculated as follows:

[0215] ;

[0216] in, Humidity compensation factor; The relative humidity of the environment (unit: %), when The relay amplification gain is increased by 3dB at this time.

[0217] Please see Figure 2This invention also provides a method for monitoring and intelligent decision-making regarding the impact of holographic blasting vibration on slope stability, comprising the following steps:

[0218] S1. Dynamically deploy a three-dimensional vibration sensor array based on terrain features; construct a digital elevation model of the slope surface using three-dimensional laser scanning technology, and adopt a honeycomb grid division strategy to adaptively adjust the sensor distribution density according to the degree of terrain undulation and rock mass structure characteristics. Nodes are added at key locations such as fracture development areas and slope shoulders. After deployment, baseline calibration is performed using a standard vibration source to ensure the consistency of response of each sensor.

[0219] S2. Acquire and preprocess blasting vibration signals and geological data; simultaneously acquire triaxial vibration acceleration and velocity signals, and eliminate environmental noise and power frequency interference through multi-stage filtering to extract effective frequency band features. Combined with ground-penetrating radar detection data, obtain rock mass integrity, joint density, and groundwater level parameters, construct a multimodal database, and achieve real-time data transmission and time synchronization alignment through heterogeneous communication protocols.

[0220] S3. A fusion of physical and data-driven models is used to predict vibration propagation. An improved blasting vibration attenuation formula characterizes the energy propagation law, and a rock mass damage factor is introduced to correct the theoretical prediction value. Simultaneously, a deep neural network is used to learn the nonlinear characteristics in historical data, and an attention mechanism is employed to fuse physical laws and data features, generating high-precision vibration field distribution prediction results and identifying potential risk propagation paths.

[0221] S4. Calculate the dynamic safety factor and generate a multi-objective optimization scheme; combine the real-time vibration energy accumulation effect and geotechnical parameters to dynamically evaluate the slope stability safety factor. Employ a multi-objective optimization algorithm to balance blasting efficiency and safety constraints, generating a Pareto optimal solution set for explosive quantity, hole spacing, and detonation sequence, providing a blasting scheme recommendation that balances economy and risk control.

[0222] S5 outputs early warning commands and provides feedback to adjust the sensor network configuration; based on the safety assessment results, it triggers a tiered early warning mechanism and executes control commands such as emergency shutdown and charge quantity limits. It simultaneously analyzes monitoring blind spots and abnormal data nodes, dynamically activates backup sensors or adjusts the network topology, and optimizes data transmission reliability through adaptive relay routing to form a closed-loop control link.

[0223] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A monitoring and intelligent decision-making system for the impact of holographic blasting vibration on slope stability, characterized in that, include: The data acquisition module is used to acquire slope blasting vibration signals in real time through a three-dimensional vibration sensor array; The data processing module, connected to the data acquisition module, is used for multi-scale feature extraction and fusion analysis of vibration signals; The intelligent decision-making module, connected to the data processing module, is used to generate blasting parameter optimization schemes and safety warning instructions based on the fusion analysis results. The feedback control module, connected to the intelligent decision-making module, is used to execute early warning commands and dynamically adjust the sensor network deployment strategy.

2. The monitoring and intelligent decision-making system for the impact of holographic blasting vibration on slope stability according to claim 1, characterized in that, The data acquisition module includes: The sensor deployment unit generates a honeycomb monitoring grid on the slope surface based on the Delaunay triangulation algorithm, dynamically deploys three-dimensional piezoelectric vibration sensors, and dynamically adjusts the sensor density. The signal conditioning unit is equipped with a bandpass filter to filter the original vibration signal in the 0.5Hz-200Hz frequency band; The multi-mode transmission unit uses RS-485 and LoRa dual-channel heterogeneous communication protocols to transmit data.

3. The monitoring and intelligent decision-making system for the impact of holographic blasting vibration on slope stability according to claim 2, characterized in that, The sensor density in the sensor deployment unit satisfies the formula: ； in, For the number of sensors, To monitor the area, For real-time risk assessment coefficients, The effective monitoring radius of the sensor.

4. The monitoring and intelligent decision-making system for the impact of holographic blasting vibration on slope stability according to claim 1, characterized in that, The data processing module includes: The feature extraction unit uses wavelet packet decomposition to extract the time-frequency energy features of the vibration signal; The physical modeling unit constructs a blasting vibration propagation model based on the improved Sadovsky formula; The data fusion unit fuses vibration characteristics with ground-penetrating radar data using weighted tensors.

5. The monitoring and intelligent decision-making system for the impact of holographic blasting vibration on slope stability according to claim 4, characterized in that, The improved Sadovsky formula in the physical modeling unit is as follows: ； in, For site coefficient, This is the equivalent amount of explosive. For the distance between the centers, The decay exponent, The elevation difference influence coefficient. The medium absorption coefficient, For joint correction factor, This is a rock quality indicator.

6. The monitoring and intelligent decision-making system for the impact of holographic blasting vibration on slope stability according to claim 4, characterized in that, The data fusion unit achieves vibration-geological feature fusion through a learnable weight tensor, satisfying the following: ； in, This is a geological feature matrix. The characteristic matrix of the vibration spectrum. and The weight tensor is obtained through training an adversarial generative network. It represents the Hadamah accumulation. This indicates a channel splicing operation.

7. The monitoring and intelligent decision-making system for the impact of holographic blasting vibration on slope stability according to claim 1, characterized in that, The intelligent decision-making module includes: Risk assessment unit, calculates dynamic safety factor : ； in, For the cohesion of rock strata, For normal stress, It is the internal friction angle. For shear stress components, To accumulate vibrational energy, This is the critical energy threshold.

8. The monitoring and intelligent decision-making system for the impact of holographic blasting vibration on slope stability according to claim 7, characterized in that, The risk assessment unit also includes: The multi-objective optimization unit uses the NSGA-III algorithm to solve the following objective function: ； Constraints include safety factor thresholds Lower limit of hole distance ; in, , These are the weighting coefficients. The peak vibration velocity of the particle. This represents the blasting cost function. For the amount of explosives, The hole spacing, To delay time, This is the cost coefficient.

9. The monitoring and intelligent decision-making system for the impact of holographic blasting vibration on slope stability according to claim 1, characterized in that, The feedback control module includes: The early warning classification unit calculates the early warning level based on the fuzzy membership function: ； in, This represents the peak vibrational velocity of the particle. Dynamically deployed units, based on real-time risk coefficients Adjust the sensor density and position.

10. A method for monitoring and intelligent decision-making regarding the impact of holographic blasting vibration on slope stability, applied to the monitoring and intelligent decision-making system for the impact of holographic blasting vibration on slope stability as described in any one of claims 1-9, characterized in that, Includes the following steps: S1. Dynamically deploy a three-dimensional vibration sensor array based on terrain features; S2. Collect and preprocess blasting vibration signals and geological data; S3. Predicting vibration propagation by integrating physical models and data-driven models; S4. Calculate the dynamic safety factor and generate a multi-objective optimization scheme; S5 outputs warning commands and provides feedback to adjust the sensor network configuration.

Images