Multi-parameter dynamic cooperative early warning method and system for instability and failure of engineering rock mass under unloading condition

Abstract

The present application relates to the technical field of monitoring and early warning of unloading instability of engineering rock mass, and discloses a multi-parameter dynamic cooperative early warning method and system for unloading instability and failure of engineering rock mass under unloading conditions. The method obtains the geology, initial stress, unloading operation and physical and mechanical parameters of the rock mass of the unloading engineering surface to be monitored, and determines the stress path of the true triaxial unloading test; four types of physical parameters, i.e. stress, acoustic emission, electromagnetic radiation and infrared imaging, are synchronously collected in the test, the unloading stages are divided according to the inflection point of the volumetric strain curve, a multi-parameter stage response database is established and the dynamic weights of each stage are calculated; during field monitoring, the real-time variation law of the four types of parameters is matched with the stage response law in the database, the current unloading stage is identified, a comprehensive early warning index is calculated, and graded early warning and disaster prevention feedback measures are output.

Description

Technical Field

[0001] This invention relates to the field of monitoring and early warning technology for unloading instability of engineering rock mass, specifically to a multi-parameter dynamic collaborative early warning method and system for instability and failure of engineering rock mass under unloading conditions. It can be used for monitoring and early warning of instability risks in mining, roadway excavation, tunnel excavation, underground cavern construction, slope excavation, and other engineering face operations involving unloading. Background Technology

[0002] In engineering activities such as mining, tunnel excavation, underground cavern construction, and slope excavation, the original stress equilibrium state of the rock mass can be altered by operations such as excavation, mining, blasting, pressure relief, support adjustment, or local constraint release. During unloading operations, the local constraint stress of the rock mass decreases, and stress concentration, superposition of disturbed stress, and stress redistribution may occur in adjacent areas. As the unloading process continues, the original fissures within the rock mass may open, close, or shift, and microcracks may gradually initiate, expand, and interconnect. In severe cases, this can induce sudden rock instability, spalling, roof collapse, rock bursts, rockbursts, and local slope landslides, threatening the safety of on-site personnel, equipment, and construction work. Therefore, monitoring and early warning of the rock mass instability and failure process under unloading conditions is a crucial technical aspect of engineering safety control.

[0003] In existing technologies, rock mass stability is typically evaluated on-site using methods such as stress monitoring, microseismic monitoring, acoustic emission monitoring, electromagnetic radiation monitoring, and infrared thermal imaging. Stress monitoring reflects changes in the stress state of the rock mass or supporting structure; acoustic emission or microseismic monitoring reflects crack activity and energy release processes within the rock mass; electromagnetic radiation monitoring reflects electromagnetic anomalies during rock fracturing; and infrared thermal imaging reflects abnormal temperature changes on the rock mass surface or in localized areas. These various monitoring methods provide a basis for assessing rock mass stability from the perspectives of mechanical response, fracturing activity, energy release, and thermal anomalies.

[0004] However, in actual unloading operations, rock mass instability and failure are often not accurately characterized by a single parameter mutation. The stress state, crack propagation degree, energy release mode, and local thermal anomalies of the rock mass may differ at different unloading stages, and different monitoring parameters also have varying sensitivities to instability precursors. Relying solely on a single monitoring parameter, a fixed threshold, or a uniform weight for judgment may make it difficult to identify early damage evolution during the unloading process in a timely manner, and may also weaken the influence of key anomalies near the instability stage, thus affecting the timeliness and reliability of early warning judgments.

[0005] Furthermore, on-site monitoring in engineering projects is easily affected by factors such as monitoring space, sensor deployment conditions, construction disturbances, dust and water vapor, equipment protection, signal transmission, and background noise, leading to problems such as missing monitoring data, delayed abnormal signals, and difficulty in unified analysis of different monitoring parameters. Although indoor rock mechanics tests can obtain various response data during the unloading and failure process of rock mass under controlled conditions, the connection between indoor test results and on-site online monitoring and early warning is still insufficient in current technologies. The rock mass unloading and failure laws obtained from the tests are difficult to effectively transform into a basis for stage identification and early warning judgment that can be used on-site.

[0006] Therefore, existing monitoring and early warning technologies for rock mass instability under unloading conditions still have problems such as insufficient utilization of multi-parameter information, insufficient consideration of response differences at different unloading stages, weak adaptability of fixed thresholds or fixed weights, and insufficient connection between indoor test patterns and field monitoring applications. These problems make it difficult to meet the needs of dynamic identification and graded early warning of instability risks in unloading engineering surfaces. Summary of the Invention

[0007] The technical problem to be solved by this invention is: the existing monitoring and early warning technology for engineering rock mass instability under unloading conditions has problems such as insufficient utilization of multi-parameter information, insufficient consideration of response differences at different unloading stages, weak adaptability of fixed thresholds or fixed weights, and insufficient connection between indoor test patterns and field monitoring applications.

[0008] To address the above problems, this invention provides a multi-parameter dynamic collaborative early warning method for engineering rock mass instability and failure under unloading conditions, comprising the following steps: S1. Obtain the geological, initial stress, unloading operation and rock physical and mechanical parameters of the unloading surface to be monitored, and determine the stress path of the true triaxial unloading test corresponding to the on-site unloading process accordingly. S2. Perform a true triaxial unloading test on the rock mass specimen according to the stress path, and simultaneously collect four types of physical parameters: stress, acoustic emission, electromagnetic radiation and infrared imaging. S3. Based on the characteristic inflection points of the volumetric strain curves in the experiment, the unloading stages are divided into the initial unloading stabilization period, the damage evolution acceleration period, and the instability and failure precursor period. The stage response laws of four types of parameters under each stage are extracted, and a multi-parameter stage response database is established. S4. For each unloading stage, the characteristic indicators of the four types of parameters are standardized, and the dynamic weights of each type of parameter in this stage are calculated based on principal component analysis. S5. During on-site monitoring, four types of parameters of the surface to be monitored are collected in real time, and their change patterns are matched with the phased response patterns in the database to identify the current unloading stage. The corresponding dynamic weights are called to calculate the comprehensive early warning index, and graded early warning information and disaster prevention feedback measures are output accordingly.

[0009] Further, in step S1, the unloading engineering face to be monitored is an engineering face in which the constraint stress in at least one direction of the rock mass is reduced and stress redistribution occurs due to excavation, mining, blasting, pressure relief, support adjustment or local constraint release; the unloading engineering face to be monitored includes at least one of the following: mine roadway excavation face, mining face, coal pillar edge area, tunnel excavation face, underground cavern excavation face or slope excavation face.

[0010] Further, in step S1, the geological parameters include at least one of the following: stratum position, thickness, dip angle, geothermal temperature, structural distribution, degree of fracture development, water-bearing state, and permeability; the initial stress parameters include at least one of the following: stress distribution direction and stress magnitude; the unloading operation parameters include at least one of the following: unloading location, unloading direction, unloading rate, unloading duration, and unloading influence range; and the rock mass physical and mechanical parameters include at least one of the following: density, water content, porosity, uniaxial compressive strength, uniaxial tensile strength, elastic modulus, and Poisson's ratio.

[0011] Furthermore, in step S1, determining the stress path for the true triaxial unloading test includes: S11. Determine the direction and magnitude of the ground stress distribution on the unloading surface to be monitored based on the field test results; S12. Based on the stress variation law in different directions during the on-site unloading operation, determine the loading, holding or unloading relationship of the rock mass in different directions; S13. Determine the stress path for indoor true triaxial unloading tests based on the direction of ground stress distribution, the magnitude of ground stress, and the loading, holding, or unloading relationships in different directions.

[0012] Further, in step S2, the true triaxial unloading test includes: The rock mass specimen is loaded with stress levels corresponding to the initial three-dimensional stress state in the field in three mutually perpendicular loading directions. After entering the unloading stage, the stress in at least one loading direction is unloaded according to the stress change relationship corresponding to the actual unloading process in the field, and the stress in the other loading directions is loaded, maintained or adjusted according to the stress evolution law in the field until the rock mass specimen undergoes macroscopic instability failure or reaches the preset termination condition.

[0013] Furthermore, the stress path of the true triaxial unloading test includes at least one of the following: a graded unloading path, a continuous unloading path, and a rapid unidirectional unloading path; The graded unloading path is as follows: after loading the rock mass specimen to a three-dimensional stress state corresponding to the initial stress state on site, the stress in the target unloading direction is unloaded step by step according to the unloading amplitude and holding time corresponding to the gradual unloading process on site. The continuous unloading path is as follows: after loading the rock mass specimen to a three-dimensional stress state corresponding to the initial stress state on site, the stress in the target unloading direction is continuously unloaded according to the unloading rate corresponding to the continuous unloading process on site. The rapid unidirectional unloading path is as follows: after loading the rock mass specimen to a three-dimensional stress state corresponding to the initial stress state on site, the stress in the target unloading direction is rapidly unloaded according to the unloading rate corresponding to the rapid unloading process or the sudden release of local constraints on site.

[0014] Further, in step S3, the characteristic indices of the stress parameters include at least one of axial stress, circumferential strain, volumetric strain, and rate of change of elastic modulus; the characteristic indices of the acoustic emission parameters include at least one of count rate, energy rate, cumulative count, cumulative energy, and b-value; the b-value is the slope of the acoustic emission signal amplitude distribution statistical curve, used to quantitatively characterize the scale evolution law of microcracks inside the material; when the b-value decreases, it indicates that large-scale crack activity intensifies, small-scale cracks decrease, and the risk of near failure of the rock mass or rock mass specimen increases; when the b-value increases or remains stable, it indicates that small-scale micro-fracture activity is dominant, and the rock mass or rock mass specimen as a whole is in a relatively stable damage evolution state. The characteristic indices of the electromagnetic radiation parameters include at least one of instantaneous amplitude, pulse number, cumulative pulse number, and average amplitude; the characteristic indices of the infrared imaging parameters include at least one of temperature gradient, maximum temperature difference, high-temperature zone area ratio, and temperature variation coefficient.

[0015] Furthermore, in step S3, the unloading stage is divided according to the characteristic inflection point of the volumetric strain curve in the experiment, including: S31. In a true triaxial unloading test, based on the volumetric strain data collected or output by the true triaxial testing equipment, establish a volumetric strain curve that varies with time. S32. Perform slope change analysis or first derivative analysis on the volumetric strain curve to identify the location where the volumetric strain change rate continuously changes; S33. The location where the rate of change of volumetric strain changes continuously is taken as the characteristic inflection point; S34. Based on the aforementioned characteristic inflection point, the true triaxial unloading process is divided into the initial unloading stabilization period, the damage evolution acceleration period, and the instability and failure precursor period; S35. Based on the variation patterns of stress parameters, acoustic emission parameters, electromagnetic radiation parameters, and infrared imaging parameters in different unloading stages, the division results of the unloading stages are verified, and the stage response patterns corresponding to different unloading stages are formed.

[0016] Furthermore, steps S4 and S5 include: S41. Perform Z-score standardization on the characteristic indices of the four types of parameters within the same unloading stage; S42. Calculate the correlation coefficient matrix based on the standardized feature indicators; S43. Solve for the eigenvalues ​​and eigenvectors of the correlation coefficient matrix, and select principal components with a cumulative contribution rate of not less than 85% according to the magnitude of the eigenvalues; S44. Calculate the comprehensive weight of each feature index based on the loading coefficients and variance contribution rates of the selected principal components; S45. Sum the comprehensive weights of the characteristic indices belonging to the same type of physical parameters to obtain the dynamic weights of stress parameters, acoustic emission parameters, electromagnetic radiation parameters and infrared imaging parameters under the corresponding unloading stage. S46. Normalize the weighted composite values ​​of the four types of parameters to the same numerical range; S51. During on-site monitoring, the real-time collected data on the changing patterns of the four types of parameters are matched with the phased response patterns of different unloading stages in the multi-parameter phased response database to determine the current unloading stage. S52. Based on the dynamic weights corresponding to the current unloading stage, the standardized comprehensive values ​​of the four types of parameters are weighted and summed to obtain the comprehensive early warning index; S53. Output blue, yellow, orange, or red warning information based on the comprehensive early warning index, and invoke the disaster prevention feedback measures corresponding to the warning information.

[0017] The present invention also provides a multi-parameter dynamic collaborative early warning system for instability and failure of engineering rock mass under unloading conditions, including a multi-parameter sensing unit, a data acquisition unit, a transmission unit, a data processing unit, an early warning output unit, and a feedback control unit; The multi-parameter sensing unit is deployed on the unloading surface to be monitored. The multi-parameter sensing unit includes a stress monitoring component, an acoustic emission monitoring component, an electromagnetic radiation monitoring component, and an infrared thermal imaging monitoring component. The output terminals of the stress monitoring component, acoustic emission monitoring component, electromagnetic radiation monitoring component, and infrared thermal imaging monitoring component are respectively connected to the input terminal of the data acquisition unit; The output of the data acquisition unit is connected to the data processing unit through the transmission unit; The data processing unit is connected to the early warning output unit and the feedback control unit, respectively. The data processing unit is used to call the multi-parameter stage response database established by the true triaxial unloading test, and to execute the multi-parameter dynamic collaborative early warning method for the instability and failure of engineering rock mass under the unloading condition. Furthermore, the stress monitoring component includes a stress sensor and / or a strain sensor; the acoustic emission monitoring component includes an acoustic emission sensor; the electromagnetic radiation monitoring component includes an electromagnetic radiation monitor; and the infrared thermal imaging monitoring component includes an infrared thermal imager.

[0018] Compared with existing technologies that mainly rely on a single parameter and a fixed threshold for judgment, the present invention has the following advantages: First, this invention addresses the early warning needs of engineering rock mass instability and failure under unloading conditions. It determines the stress path of the true triaxial unloading test based on the geology, initial stress, unloading operation, and physical and mechanical parameters of the unloading surface to be monitored, so that the indoor test can be as close as possible to the on-site unloading process and avoid the test conditions from being out of touch with the actual engineering conditions.

[0019] Secondly, this invention divides the unloading stage in a true triaxial unloading test by using volumetric strain curves and extracts the response patterns of four types of physical parameters—stress, acoustic emission, electromagnetic radiation, and infrared imaging—at each stage, forming a multi-parameter stage response database. During on-site monitoring, the real-time collected variation patterns of the four types of parameters can be matched with the stage response patterns in the database, thus enabling the identification of the current unloading stage even when it is difficult to directly and completely obtain the volumetric strain curve on-site.

[0020] Furthermore, this invention calculates dynamic weights for different unloading stages and calls the corresponding stage weights to calculate a comprehensive early warning index when issuing an on-site warning. This allows the early warning judgment to be adjusted according to the stage changes of the unloading and damage process, reducing the risk of misjudgment or missed judgment caused by fixed thresholds or fixed weights.

[0021] Furthermore, this invention outputs tiered early warning information based on a comprehensive early warning index and links it with corresponding disaster prevention feedback measures, enabling the monitoring results to be further transformed into actionable measures on-site. Attached Figure Description

[0022] To more clearly illustrate the technical solution of the present invention, the accompanying drawings used in this specification are briefly described below.

[0023] Figure 1 A schematic diagram of a multi-parameter dynamic collaborative early warning method for rock mass instability and failure under unloading conditions; Figure 2 This is a schematic diagram of the volumetric strain versus time curve in a true triaxial unloading test; Figure 3 This is a schematic diagram of the rate of change of volumetric strain over time in a true triaxial unloading test. Figure 4 A schematic diagram of a multi-parameter dynamic collaborative early warning system for rock mass instability and failure under unloading conditions; Figure 5 The volumetric strain-time curve of the gas-bearing coal and rock sample in Example 3; Figure 6 The volumetric strain growth rate curve of the gas-bearing coal and rock sample in Example 3 is shown. Detailed Implementation

[0024] The technical solution of the present invention will now be clearly and completely described with reference to the accompanying drawings. It should be understood that the following embodiments are intended to enable those skilled in the art to understand and implement the present invention, and the specific implementation conditions described can be adaptively adjusted according to the geological conditions of the engineering site, the conditions of the test equipment, and the conditions of the monitoring layout.

[0025] Example 1 like Figure 1 As shown in the figure, this embodiment provides a multi-parameter dynamic collaborative early warning method for the instability and failure of engineering rock mass under unloading conditions.

[0026] This method first acquires the geological, initial stress, unloading operation, and rock mass physical and mechanical parameters of the unloading surface to be monitored, and determines the true triaxial unloading test stress path corresponding to the on-site unloading process based on these parameters. Then, a true triaxial unloading test is conducted on the rock mass specimen according to the stress path, and four types of physical parameters—stress, acoustic emission, electromagnetic radiation, and infrared imaging—are collected simultaneously. Subsequently, the unloading stages are divided according to the characteristic inflection points of the volumetric strain curves in the test, and the stage response laws of the four types of parameters under each stage are extracted to establish a multi-parameter stage response database. Then, the dynamic weights of the four types of parameters are calculated for each unloading stage. During on-site monitoring, the real-time collected change laws of the four types of parameters are matched with the stage response laws in the database to identify the current unloading stage, and the corresponding dynamic weights are called to calculate the comprehensive early warning index, outputting graded early warning information and disaster prevention feedback measures.

[0027] During the on-site parameter acquisition phase, engineering faces that are undergoing unloading operations or may be affected by unloading are selected as the unloading engineering faces to be monitored. The unloading engineering faces to be monitored are those where the constraint stress in at least one direction of the rock mass has decreased and stress redistribution has occurred due to excavation, mining, blasting, pressure relief, support adjustment, or local constraint release. These unloading engineering faces can be mine roadway excavation faces, working faces in mining areas, coal pillar edge areas, tunnel excavation faces, underground cavern excavation faces, or slope excavation faces. Monitoring points can also be set up in areas within the unloading influence range that have not yet shown obvious anomalies to identify potential instability risks in advance.

[0028] At the unloading surface to be monitored, geological, initial stress, unloading operation, and rock mass physical and mechanical parameters are collected. Geological parameters may include strata position, thickness, dip angle, geothermal temperature, structural distribution, fracture development, water content, and permeability; initial stress may include the direction and magnitude of geostress distribution; unloading operation parameters may include unloading location, unloading direction, unloading rate, unloading duration, and unloading influence range; rock mass physical and mechanical parameters may include density, water content, porosity, uniaxial compressive strength, uniaxial tensile strength, elastic modulus, and Poisson's ratio.

[0029] The above parameters can be obtained through engineering geological data, core drilling, field testing, geostress testing, rock mechanics experiments, construction records, or numerical simulation analysis. For gas-bearing coal and rock masses, water-rich rock masses, or high-temperature engineering rock masses, in addition to obtaining stress state and rock mass physical and mechanical parameters, environmental parameters such as gas pressure, gas content, adsorption and desorption characteristics, pore water pressure, water content, permeability, and background geothermal temperature should also be obtained in conjunction with the engineering environment. These environmental parameters are used to characterize the gas, moisture, and temperature conditions of the rock mass during the unloading process and serve as the basis for setting gas pressure, hydraulic boundaries, or temperature boundaries in indoor true triaxial unloading tests. In one implementation, the direction and magnitude of in-situ stress distribution can be obtained using hydraulic fracturing, hollow inclusion strain gauge methods, or other in-situ stress testing methods. For mine roadway scenarios, the in-situ stress test results can be correlated with vertical stress, horizontal stress perpendicular to the roadway axis, and horizontal stress parallel to the roadway axis by combining the roadway orientation and unloading direction. For other unloading engineering faces such as tunnels, underground caverns, or slopes, the initial stress state in the corresponding direction can be determined based on the engineering face orientation, excavation direction, or unloading direction. If necessary, existing numerical simulation software can be used to assist in analyzing the stress change patterns caused by on-site unloading operations and to transform the on-site stress evolution process into an executable stress path for indoor true triaxial unloading tests.

[0030] When determining the stress path for a true triaxial unloading test, the direction and magnitude of the geostress distribution on the unloading surface to be monitored are determined based on the field test results. Furthermore, the loading, holding, or unloading relationships of the rock mass in different directions are determined based on the stress variation patterns during the field unloading operation. For engineering rock masses affected by gas, water, or temperature, the aforementioned environmental parameters should be combined with the stress path to ensure that the laboratory test not only reproduces the three-dimensional stress state and unloading path on-site but also selectively reproduces the influence of environmental factors such as gas, water, and temperature on the rock mass damage evolution and multi-parameter response, depending on the engineering type.

[0031] For example, for engineering faces with significant excavation unloading or local constraint release, the direction closer to the free excavation face can be used as the target unloading direction. For stress redistribution processes caused by working face advancement, mining backfilling, tunnel excavation, underground cavern expansion, or slope excavation, the unloading direction can be combined with the stress concentration direction, so that the stress path of the true triaxial unloading test can reflect the actual stress state on site where some directions are unloaded while others continue to be loaded or maintained. The loading rate, unloading rate, unloading amplitude, and holding time in the true triaxial unloading test are not limited to a single fixed value, but are comprehensively determined based on the on-site unloading process, the control capability of the test equipment, and the bearing conditions of the rock mass specimen, with the aim of reproducing the main stress evolution process of the unloaded engineering face to be monitored.

[0032] In the indoor true triaxial unloading test phase, a true triaxial testing machine was used to conduct loading and unloading tests on the rock mass specimens. First, the rock mass specimens were loaded in three mutually perpendicular loading directions to stress levels corresponding to the initial three-dimensional stress state in the field. Then, in the unloading phase, the stress in at least one loading direction was unloaded according to the stress change relationship corresponding to the actual unloading process in the field, while the stress in the other loading directions was loaded, maintained, or adjusted according to the stress evolution law in the field until the rock mass specimens experienced macroscopic instability and failure or reached the preset termination condition.

[0033] The stress path for the true triaxial unloading test can include at least one of the following: a graded unloading path, a continuous unloading path, and a rapid unidirectional unloading path. The graded unloading path is used to simulate the gradual excavation, step-by-step pressure relief, or phased constraint release process on-site; the continuous unloading path is used to simulate the continuous stress release process caused by the continuous advancement of unloading operations; and the rapid unidirectional unloading path is used to simulate the rapid unloading process caused by blasting disturbance, sudden exposure of local rock mass, or rapid release of local constraints. The specific unloading rate, unloading amplitude, and holding time of the above three types of unloading paths can be equivalently set according to the on-site unloading operation process and are not limited to a single fixed value.

[0034] During the true triaxial unloading test, four types of physical parameters—stress, acoustic emission, electromagnetic radiation, and infrared imaging—are simultaneously acquired. Stress parameters are acquired using the true triaxial testing machine and its associated sensors; acoustic emission parameters are acquired using acoustic emission monitoring equipment; electromagnetic radiation parameters are acquired using electromagnetic radiation monitoring equipment; and infrared imaging parameters are acquired using an infrared thermal imager. All monitoring devices use a unified acquisition start time and corresponding timestamp, enabling corresponding analysis of the four types of physical parameters during the same unloading process. If different devices have different sampling frequencies, existing data processing methods such as time window aggregation, resampling, or timestamp alignment can be used to achieve data synchronization.

[0035] In the feature extraction stage, at least one of axial stress, circumferential strain, volumetric strain, and elastic modulus change rate is extracted from stress parameters; at least one of count rate, energy rate, cumulative count, cumulative energy, and b-value is extracted from acoustic emission parameters; at least one of instantaneous amplitude, pulse number, cumulative pulse number, and average amplitude is extracted from electromagnetic radiation parameters; and at least one of temperature gradient, maximum temperature difference, high-temperature area ratio, and temperature variation coefficient is extracted from infrared imaging parameters.

[0036] In the unloading stage segmentation, volumetric strain serves as a crucial basis for defining the unloading stage in indoor true triaxial unloading tests. During a true triaxial unloading test, volumetric strain data can be directly output by the true triaxial testing equipment during the test, or it can be output after processing the deformation or strain data collected by the testing equipment. The specific method for obtaining volumetric strain can be determined based on the data acquisition method of the true triaxial testing equipment.

[0037] After establishing the curve of volumetric strain versus time, slope change analysis or first derivative analysis is performed on the curve to identify the location where the rate of volumetric strain change continuously changes, and this location is taken as the characteristic inflection point. Based on the characteristic inflection point, the true triaxial unloading process can be divided into the initial unloading stabilization period, the damage evolution acceleration period, and the instability and failure precursor period.

[0038] In one experimental example, the rock mass specimen was unloaded according to the above-described true triaxial unloading test path, and the volumetric strain variation data over time were recorded during the test, thereby obtaining the following results: Figure 2 The volumetric strain evolution curve is shown. To further observe the rate of change of volumetric strain, the volumetric strain data is subjected to differential processing of adjacent sampling points or first-order derivative analysis, resulting in the following... Figure 3 The volumetric strain rate curve is shown. Based on... Figure 2 and Figure 3 It can be seen that the rock mass specimens exhibited phased characteristics during the unloading process, namely, gradual change in volumetric strain, gradual increase in the rate of change of volumetric strain, and rapid increase in the rate of change of volumetric strain. This can be used to help determine the initial unloading stabilization period, the damage evolution acceleration period, and the instability and failure precursor period.

[0039] In practical analysis, the volumetric strain curve typically reflects the overall deformation evolution of a rock mass specimen during unloading. In the initial unloading phase, the volumetric strain changes relatively smoothly over time, indicating that the rock mass specimen is mainly in a state of stress adjustment and relatively stable deformation. As unloading continues, the curve slope gradually increases, indicating that internal damage begins to accumulate at an accelerated rate. When the specimen approaches instability, the curve shows a significant nonlinear rise, indicating that cracks are rapidly expanding and tending to penetrate. The rate of change of volumetric strain is a further characterization of the slope change of the volumetric strain curve. It remains at a low level with small fluctuations during the initial unloading stabilization period, continues to rise during the accelerated damage evolution period with enhanced local fluctuations, and rapidly increases and forms a concentrated peak during the pre-instability failure period. Therefore, by mutually verifying the trend of volumetric strain change and the rate of change of volumetric strain, the staged changes of the rock mass specimen from stable deformation and accelerated damage evolution to the pre-instability failure stage can be clearly identified.

[0040] After completing the stage division based on the volumetric strain curve, the response characteristics of stress parameters, acoustic emission parameters, electromagnetic radiation parameters, and infrared imaging parameters in different stages were further analyzed. The multi-parameter coordinated response characteristics of different unloading stages are shown in Table 1.

[0041] Table 1. Multi-parameter coordinated response characteristics at different stages of unloading failure. Based on the multi-parameter synergistic response characteristics shown in Table 1, the characteristic indices, stage boundaries, response trends, and dynamic weight calculation results of different stages in the true triaxial unloading test are stored to establish a multi-parameter stage response database. This database uses the true triaxial unloading test data as the basis for stage division and pattern extraction. The stage response patterns stored in the database are used for stage matching in the field monitoring stage. The field monitoring stage does not require direct and complete acquisition of the volumetric strain curves corresponding to the indoor rock mass specimens.

[0042] In the phased dynamic weight calculation stage, principal component analysis is used to calculate the dynamic weights of different physical parameters for each unloading stage. First, the characteristic index sample data within a specific unloading stage are Z-score standardized to eliminate differences in dimensions and numerical scales. The standardization formula is: ; in, Indicates the first The first sample The standardized values ​​of each characteristic indicator Indicates the first The first sample The original values ​​of each characteristic indicator, Indicates the first [number]th ... The average of all samples for each characteristic indicator. Indicates the first [number]th ... The standard deviation of all samples for each characteristic index; when When the value is zero, this characteristic index will not participate in the dynamic weight calculation during the corresponding unloading stage, or it will be removed as a constant index.

[0043] After standardizing all feature indicators in this stage, calculate the correlation coefficient matrix R between the feature indicators. Any element in the correlation coefficient matrix R... Represented as: ; in, Indicates the first The first sample The standardized variable values ​​of each indicator, Indicates the first The first sample The standardized variable values ​​of each indicator, This indicates the total number of samples involved in the calculation during this stage. Indicates the first The first indicator and the first The correlation coefficients between the indicators.

[0044] After obtaining the correlation coefficient matrix R, solve for the eigenvalues ​​of this matrix. and its corresponding feature vector The solved eigenvalues Arrange the components in descending order of variance, calculate the cumulative variance contribution rate, and select the top N principal components with a cumulative variance contribution rate of not less than 85%. Then, calculate the loading coefficients of each feature index in the selected principal components. The calculation relationship is as follows: ; in, Indicates the first The first characteristic index in the th Loading coefficients in each principal component Indicates the first The eigenvector corresponding to the principal component is the th ? One portion, Indicates the first The eigenvalues ​​corresponding to each principal component.

[0045] After obtaining the loading coefficients, a weighted sum is performed using the variance contribution rate of each principal component as the weight to obtain the comprehensive weight of each feature index at this stage. The calculation formula is: ; in, Indicates the first The combined weight of each characteristic indicator Indicates the number of selected principal components. Indicates the first [participant] in the summation. The eigenvalues ​​corresponding to each principal component are calculated, and then normalized.

[0046] To obtain the dynamic weights of four types of physical parameters—stress, acoustic emission, electromagnetic radiation, and infrared imaging—the comprehensive weights of characteristic indices belonging to the same type of physical parameter are summed. Specifically, the comprehensive weights of the corresponding characteristic indices among axial stress, circumferential strain, volumetric strain, and rate of change of elastic modulus are summed to obtain the dynamic weight of the stress parameter at that stage. ; The dynamic weights of the acoustic emission parameters at this stage are obtained by summing the weights of the corresponding characteristic indicators among the count rate, energy rate, cumulative count, cumulative energy, and b-value. ; The dynamic weights of the electromagnetic radiation parameters at this stage are obtained by summing the weights of the corresponding characteristic indicators among the instantaneous amplitude, pulse number, cumulative pulse number, and average amplitude. ; The dynamic weights of the infrared imaging parameters at this stage are obtained by summing the weights of the corresponding characteristic indicators among temperature gradient, maximum temperature difference, high-temperature zone area ratio, and temperature variation coefficient. .

[0047] The aforementioned four types of parameter weights serve as dynamic weights for the corresponding unloading stages. To ensure comparability of different physical parameters in the final calculation of the comprehensive early warning index, the original comprehensive values ​​obtained by weighting each type of physical parameter according to its characteristic index weights are mapped to the same numerical range using a normalization method, resulting in the standardized comprehensive values ​​S for stress, AE, EME, and IR for acoustic emission, electromagnetic radiation, and infrared imaging, respectively. The principal component analysis, normalization, and weight calculation described above can be performed using a general-purpose computer, engineering monitoring server, or existing data processing software.

[0048] During the on-site monitoring phase, multi-parameter sensing devices are deployed on the unloading surface to be monitored to collect stress parameters, acoustic emission parameters, electromagnetic radiation parameters, and infrared imaging parameters in real time. Since it is usually difficult to directly and completely obtain the volumetric strain curve corresponding to the indoor rock mass specimens from the on-site engineering surface, the volumetric strain curve is not used as the sole identification criterion for the current unloading stage. Instead, the changing patterns of the four types of parameters collected in real time are matched with the stage-specific response patterns corresponding to different unloading stages in the multi-parameter stage response database.

[0049] When the on-site stress parameters change steadily and no obvious abnormalities are found in acoustic emission, electromagnetic radiation and infrared imaging parameters, it is determined that the current unloading engineering surface is in the initial unloading stabilization period, and the dynamic weight corresponding to the initial unloading stabilization period is called to calculate the comprehensive early warning index. When the rate of change of on-site stress or strain increases, the acoustic emission count rate and energy rate increase, the number of electromagnetic radiation pulses and the average amplitude increase, and the infrared temperature difference or temperature gradient begins to increase, it is determined that the current unloading engineering surface is in the damage evolution acceleration period, and the dynamic weight corresponding to the damage evolution acceleration period is called to calculate the comprehensive early warning index. When strong synergistic anomalies, abrupt increases, or concentrated peak values ​​appear in the on-site stress, acoustic emission, electromagnetic radiation, and infrared imaging parameters, it is determined that the current unloading engineering surface is in the pre-instability and failure stage, and the comprehensive early warning index is calculated by calling the dynamic weight corresponding to the pre-instability and failure stage.

[0050] After identifying the current unloading stage, retrieve the dynamic weights corresponding to the current unloading stage from the database. , , and The standardized composite values ​​S, AE, EME, and IR of each parameter are weighted and summed to obtain the comprehensive early warning index. Comprehensive Early Warning Index Represented as: .

[0051] Since the standardized composite values ​​of stress parameters, acoustic emission parameters, electromagnetic radiation parameters, and infrared imaging parameters are all normalized to the interval [0,1], and the dynamic weights corresponding to each parameter, after normalization, satisfy a weight sum of 1, the comprehensive early warning index I ranges from [0,1]. The closer the value is to 1, the more obvious the multi-parameter anomaly response and the higher the risk of rock mass instability.

[0052] In the stage of tiered early warning output and disaster prevention feedback linkage, the first early warning threshold, the second early warning threshold, and the third early warning threshold are set as follows: , , And satisfying 0 < < < <1. , , The comprehensive early warning index during the normal on-site monitoring period can be determined using the mean-standard deviation method.

[0053] Specifically, the comprehensive early warning index during the normal on-site monitoring period is selected as the benchmark sample, denoted as... , , ..., Calculate its average value : ; Calculate its standard deviation : .

[0054] Three baseline thresholds were then set: ; ; .

[0055] For areas with high ground stress and high temperature, a ground stress correction factor can be introduced. and geothermal correction factor The baseline threshold and the geostress correction factor are combined. and geothermal correction factor By combining these factors, the actual dynamic early warning threshold is obtained. The actual dynamic early warning threshold is... Represented as: ; in, Indicates the actual execution of the first Level dynamic early warning threshold, Indicates the first Level-based early warning threshold. Ground stress correction factor. and geothermal correction factor The value can be determined based on the current burial depth, background ground temperature, historical monitoring data, and field test data of the monitoring area. Without introducing a correction factor, it can be taken as... =1、 =1.

[0056] when < At that time, the system outputs a blue warning message; when ≤ < When this happens, the system outputs a yellow warning message; when ≤ < At that time, the system outputs an orange warning message; when ≥ At that time, the system outputs a red warning message.

[0057] The disaster prevention and response measures corresponding to different warning levels are shown in Table 2 below.

[0058] Table 2 Disaster prevention feedback measures corresponding to different warning levels After the system outputs different levels of early warning information, it calls the corresponding disaster prevention feedback measures according to Table 2. When the comprehensive early warning index continuously exceeds a certain level threshold, the system raises the early warning level and triggers the corresponding response measures; when the response measures are implemented and the comprehensive early warning index continuously decreases and stabilizes below the corresponding threshold, the system can gradually lower the early warning level based on the on-site verification results.

[0059] Through the implementation of this embodiment, the unloading stage division results, the stage response laws of the four types of parameters, and the dynamic weights obtained in the true triaxial unloading test can be applied to the field monitoring process. Even without directly relying on the volumetric strain curve in the field, this method can still identify the current unloading stage through the combined variation laws of the four types of parameters—stress, acoustic emission, electromagnetic radiation, and infrared imaging—and output corresponding levels of early warning information and disaster prevention feedback measures based on the dynamic weights of the corresponding stages.

[0060] Example 2 like Figure 4 As shown in the figure, this embodiment provides a multi-parameter dynamic collaborative early warning system for the instability and failure of engineering rock mass under unloading conditions.

[0061] The system comprises a multi-parameter sensing unit, a data acquisition unit, a transmission unit, a data processing unit, an early warning output unit, and a feedback control unit. The multi-parameter sensing unit is deployed on the unloading surface to be monitored. The data acquisition unit is connected to the multi-parameter sensing unit, and the transmission unit connects to both the data acquisition unit and the data processing unit. The data processing unit is connected to both the early warning output unit and the feedback control unit. Thus, the multi-parameter monitoring data collected on-site can be sent to the data processing unit via the data acquisition and transmission units. The data processing unit then performs stage matching, dynamic weighting, and comprehensive early warning index calculation, outputting tiered early warning information and corresponding feedback control information.

[0062] The multi-parameter sensing unit includes a stress monitoring component, an acoustic emission monitoring component, an electromagnetic radiation monitoring component, and an infrared thermal imaging monitoring component. The stress monitoring component includes stress sensors and / or strain sensors to collect stress and / or strain change information in the unloaded engineering surface or its adjacent rock mass and support structure. The acoustic emission monitoring component includes acoustic emission sensors to collect acoustic emission signals generated during the initiation, propagation, or penetration of rock cracks. The electromagnetic radiation monitoring component includes an electromagnetic radiation monitor to collect electromagnetic radiation signals generated during the loading and fracturing or crack propagation of rock mass. The infrared thermal imaging monitoring component includes an infrared thermal imager to acquire infrared thermal images of the monitored area and extract infrared imaging parameters such as temperature gradient, maximum temperature difference, area ratio of high-temperature zones, or temperature variation coefficient from the infrared thermal images. The output terminals of the stress monitoring component, acoustic emission monitoring component, electromagnetic radiation monitoring component, and infrared thermal imaging monitoring component are respectively connected to the input terminals of the data acquisition unit, allowing stress parameters, acoustic emission parameters, electromagnetic radiation parameters, and infrared imaging parameters to be input into the data acquisition unit.

[0063] The data acquisition unit receives raw data collected by the multi-parameter sensing unit and performs aggregation, buffering, and time stamping on the raw data. For analog signals, digital signals, or image data output from different types of sensors, the data acquisition unit can process them using existing signal conditioning, analog-to-digital conversion, protocol conversion, and data buffering methods. By using a unified time stamp, stress, acoustic emission, electromagnetic radiation, and infrared imaging data can be mapped to the same monitoring time window, providing a data foundation for subsequent stage matching and comprehensive early warning index calculation.

[0064] The transmission unit is used to transmit the multi-parameter monitoring data processed by the data acquisition unit to the data processing unit. The transmission unit can adopt wired transmission, wireless transmission or transmission method that connects to the existing monitoring communication network at the engineering site.

[0065] The data processing unit can consist of a server, industrial control computer, monitoring host, or a computing platform with data processing capabilities. The data processing unit receives multi-parameter monitoring data sent by the transmission unit and calls upon the multi-parameter stage response database established from indoor true triaxial unloading tests. Specifically, the data processing unit first performs time synchronization, noise filtering, outlier removal, normalization, and feature extraction on the stress parameters, acoustic emission parameters, electromagnetic radiation parameters, and infrared imaging parameters collected in real time on-site; then, it matches the combined variation patterns of the four types of parameters on-site with the stage response patterns of different unloading stages in the multi-parameter stage response database to determine the current unloading stage; finally, it calls upon the dynamic weight corresponding to the current unloading stage to calculate the comprehensive early warning index.

[0066] The early warning output unit is connected to the data processing unit and is used to output tiered early warning information. The early warning output unit may include a display terminal, an audible and visual alarm, a dispatch room monitoring interface, a mobile communication terminal, a field broadcasting system, or a safety management platform. The early warning output content may include the early warning level, comprehensive early warning index, location of abnormal areas, key abnormal parameters, early warning time, and recommended handling measures.

[0067] The feedback control unit connects to the data processing unit to generate on-site handling instructions or information based on the warning level. The feedback control unit can connect to engineering scheduling systems, equipment control systems, broadcasting systems, personnel positioning systems, or safety linkage systems to send information such as increasing sampling frequency, focusing on key inspections, restricting operations, suspending operations, evacuating personnel, or activating emergency plans. For engineering measures such as pressure relief, reinforcement, support adjustment, and hazardous area closure, the feedback control unit can output handling suggestions or execution instructions to the on-site management terminal.

[0068] During system operation, the multi-parameter sensing unit continuously collects four types of physical parameters—stress, acoustic emission, electromagnetic radiation, and infrared imaging—from the unloaded engineering surface under monitoring. The data acquisition unit aggregates and timestamps the data, and the transmission unit transmits the data to the data processing unit. The data processing unit identifies the current unloading stage and calculates the comprehensive early warning index based on the multi-parameter stage response database established from indoor true triaxial unloading tests. The early warning output unit issues early warning information, and the feedback control unit generates corresponding response measures based on the early warning level. Figure 4 As shown in the system structure, this system achieves dynamic and coordinated early warning of the instability and failure process of engineering rock mass under unloading conditions through the coordination of multi-parameter sensing, data acquisition, data transmission, stage identification, early warning output and feedback control.

[0069] Example 3 To further illustrate the specific implementation process of stage division, standardization, principal component analysis, dynamic weight calculation, and comprehensive early warning index calculation in this invention, this embodiment uses a gas-bearing coal and rock unloading instability early warning scenario in a coal mine as an example. During the advancement of the working face in this coal mine, the coal and rock mass ahead is affected by mining, resulting in stress redistribution. There are risks of reduced lateral constraint stress, gas involvement, and accelerated evolution of coal and rock damage in local areas. To simulate the loading and failure process of gas-bearing coal and rock under unloading disturbance in this coal mine, coal and rock samples were collected near the working face of the coal mine, and the corresponding stress state was reproduced in a laboratory true triaxial multi-parameter test system. Multi-parameter response data of the coal and rock samples during loading and unloading failure processes were collected simultaneously.

[0070] In this embodiment, considering the coal and rock occurrence conditions at the coal mine site and the experimental reproduction conditions, the actual test conditions were determined as follows: the gas pressure was 4 MPa, and the two lateral stresses were respectively... =5MPa =15MPa. Under this condition, the volumetric strain, axial stress, cumulative acoustic emission energy window value, and maximum absolute amplitude of charge signal of the coal and rock samples are collected simultaneously. The volumetric strain-time curve is used to delineate the deformation and failure stages of the coal and rock; axial stress serves as a mechanical response index, the cumulative acoustic emission energy window value as an acoustic response index, and the maximum absolute amplitude of charge signal as an electrical response index, used to calculate the dynamic weights and comprehensive early warning index at different stages. It should be noted that this embodiment is used to illustrate the implementation methods of stage division, stage standardization, principal component analysis, dynamic weight calculation, and comprehensive early warning index calculation in this invention. In this embodiment, only one characterization index is selected for each type of parameter; therefore, the selection of multiple indices within the same parameter is no longer performed. Instead, axial stress is directly used as the mechanical response index, the cumulative acoustic emission energy window value as the acoustic response index, and the maximum absolute amplitude of charge signal as the electrical response index. When simultaneously collecting infrared imaging parameters at the actual engineering site, the infrared imaging parameters can be incorporated into the comprehensive early warning index calculation according to the same standardization, correlation coefficient matrix calculation, principal component analysis, and dynamic weight calculation process.

[0071] Since each parameter in this embodiment contains only one characterization index, we no longer filter multiple indices within a parameter or extract the first N principal components. Instead, after completing the principal component analysis, we retain all principal components to calculate the overall contribution of each parameter. At this point, the weight of each index within its respective parameter is 1, and the final index weight is the corresponding parameter weight.

[0072] First, the failure process is divided into stages based on the volumetric strain curve. According to the volumetric strain-time curve and the variation law of the volumetric strain rate of the coal and rock samples, the failure process is divided into three stages. For example...Figure 5 As shown, the volumetric strain increases slowly in the early stage, then the rate of increase gradually increases, and finally shows a rapid upward trend in the later stage; as Figure 6 As shown, the volumetric strain growth rate was generally low in the early stage, began to rise in the middle and late stages, and significantly increased in the pre-failure stage. Therefore, the entire process is divided into three stages: the first stage, the second stage, and the third stage. In the first stage, the volumetric strain growth rate was generally low, the multi-parameter response was relatively stable, and the coal and rock mass was in a stable deformation state. In the second stage, the volumetric strain growth rate began to rise, and acoustic emission and charge signals gradually became more active, indicating accelerated internal fracture propagation and damage evolution. In the third stage, the volumetric strain growth rate continued to increase, and acoustic emission and charge signals showed significant enhancement, indicating that the coal and rock mass had entered the pre-failure stage. The stage division results are shown in Table 3 below.

[0073] Table 3 Stage Division Secondly, standardization is performed in stages. The obtained data is standardized using Z-scores for each indicator in each stage, calculated as follows: ; in, Indicates the first The first sample The standardized values ​​of each characteristic indicator Indicates the first The original value of the j-th feature index of a sample. Indicates the first [number]th ... The average of all samples for each characteristic indicator. This represents the standard deviation of all samples for the j-th feature index within this stage; when When the value is zero, this characteristic index does not participate in the dynamic weight calculation during the corresponding unloading stage, or it is removed as a constant index. The mean and standard deviation of each stage are shown in Table 4 below.

[0074] Table 4. Mean and Standard Deviation for Each Stage As shown in Table 4, as the coal and rock samples progressed from the first stage to the third stage, the mean and standard deviation of the cumulative value of the acoustic emission energy window and the maximum absolute amplitude of the charge signal both increased significantly, indicating that the discreteness, abrupt changes, and abnormality of the acoustic and electrical responses were significantly enhanced in the pre-destruction stage.

[0075] Then, the correlation coefficient matrix is ​​calculated. Based on the standardized data, the correlation coefficient matrices for each stage are calculated. any element Represented as: ; in, Indicates the first The first sample The standardized variable values ​​of each indicator, Indicates the first The standardized variable value of the j-th indicator in a sample. This indicates the total number of samples involved in the calculation during this stage. Indicates the first The first indicator and the first The correlation coefficients between the indicators are shown below. The correlation coefficient matrices for the three stages are calculated as follows: First-stage correlation coefficient matrix : ; Second-stage correlation coefficient matrix : ; Third-stage correlation coefficient matrix : ; Then, principal component analysis and dynamic weight calculation are performed. The correlation coefficient matrix is ​​then obtained. Then, solve for the eigenvalues ​​of the matrix. and its corresponding feature vector The solved eigenvalues Arrange the components in descending order and calculate the cumulative variance contribution rate. Since there are only three indicators in this embodiment, the cumulative variance contribution rate of the first three principal components in each stage is 100%, satisfying the principal component selection condition of a cumulative contribution rate of not less than 85%. If the minimum number of principal components required to achieve a cumulative contribution rate of 85% is used, the cumulative contribution rates of the first two principal components in the first, second, and third stages are 86.20%, 87.87%, and 90.63%, respectively, all meeting the requirement of not less than 85%. This embodiment retains all three principal components to calculate the comprehensive contribution of each parameter in order to fully demonstrate the contribution calculation process of each indicator under all principal components; this processing method is a special implementation method with a cumulative contribution rate of not less than 85%, and does not affect the rule in the general scheme of this invention to select the first N principal components according to a cumulative contribution rate of not less than 85%. The calculated eigenvalues ​​and contribution rates of each stage are shown in Table 5 below.

[0076] Table 5. Characteristic values ​​and contribution rates at each stage After obtaining the eigenvalues ​​and eigenvectors, the loading coefficients of each index on different principal components are calculated. The loading coefficients reflect the contribution of a particular index to a specific principal component, and their calculation relationship is as follows: ; in, Indicates the first The first characteristic index in the th Loading coefficients in each principal component Indicates the first The eigenvector corresponding to the principal component is the th ? One portion, Indicates the first The eigenvalues ​​corresponding to each principal component are shown in Table 6 below.

[0077] Table 6 Load Factors for Each Stage It should be noted that in principal component analysis, the direction of the eigenvectors is arbitrary, and the loading coefficients may be positive or negative. The sign of the loading coefficient indicates the relative direction of the corresponding indicator on a certain principal component direction, not whether the indicator has a positive or negative contribution to the early warning result; the weight calculation focuses on the strength of the contribution, therefore the absolute value of the loading coefficient is used in the calculation of the overall contribution.

[0078] Since this embodiment retains all three principal components, the loading coefficients of each index on the three principal components are weighted and summed according to the principal component contribution rate to obtain the comprehensive contribution of each index. The calculation formula is as follows: ; in, Indicates the first The combined weight of each characteristic indicator Indicates the number of selected principal components. Indicates the first [participant] in the summation. The eigenvalues ​​corresponding to each principal component. The combined contribution of the three indicators at each stage is shown in Table 7 below.

[0079] Table 7. Overall Contribution of the Three Indicators at Each Stage In Table 7, a larger overall contribution indicates a greater contribution of the index to the principal component integrated information at that stage. After obtaining the overall contributions of axial stress, acoustic emission energy, and charge signal, the three were normalized to obtain the dynamic weights at different stages, as shown in Table 8 below.

[0080] Table 8. Phased Dynamic Weights Since each of the mechanical, acoustic, and electrical parameters in this embodiment contains only one characterization index, the weight of each index within its respective parameter is 1, and the final index weight is the corresponding parameter weight.

[0081] Finally, the comprehensive early warning index is calculated. After obtaining the dynamic weights of force, acoustic, and electrical parameters at different instability stages, the axial stress, cumulative acoustic emission energy window value, and maximum absolute amplitude of the charge signal at each sampling time are normalized to 0-1, transforming the three indices into dimensionless values ​​under the same dimension. The normalization formula is: ; in, Indicates the first At the [time]th moment The original values ​​of each indicator Indicates the first The minimum value of each indicator throughout the entire process. Indicates the first The maximum value of each indicator throughout the entire process. Indicates the first At the [time]th moment The normalized values ​​of each index are used to obtain the standardized composite value of the stress parameter. Standardized composite values ​​of acoustic emission parameters Standardized composite value of electrical response parameters .

[0082] Based on the aforementioned dynamic weight calculation results, the specific calculation formula for the comprehensive early warning index for the three stages is as follows: Phase 1: ; Phase Two: ; Phase Three: ; Taking t=204s as an example, according to the stage division in Table 3, this moment belongs to the second stage. The original data at this moment are axial stress 34.875, cumulative acoustic emission energy window value 679.954, and maximum absolute amplitude of charge signal 0.404; after normalization, we can obtain... =0.91552, =0.07553, =0.22716. By applying the dynamic weights from the second stage and substituting them into the comprehensive early warning index formula, we can obtain... =0.39861. This result indicates that the internal damage of the coal and rock has begun to accelerate at this point, and the acoustic emission and charge signals have increased compared to the stable stage, but have not yet reached the high-risk level of the precursor stage of destruction.

[0083] Therefore, this embodiment, based on real experimental data, forms a complete calculation chain from the stage division of the volumetric strain curve, staged standardization processing, construction of the correlation coefficient matrix, extraction of principal component contributions, determination of dynamic weights, to the output of the comprehensive early warning index. This calculation chain can transform the staged response laws obtained from indoor true triaxial tests into quantifiable dynamic weights and a comprehensive early warning index. This enables the present invention not only to describe the multi-parameter response laws during the unloading failure process, but also to provide calculable and updatable judgment criteria for stage identification, risk classification, and disaster prevention feedback of unloading engineering surfaces in the field.

Claims

1. A multi-parameter dynamic collaborative early warning method for instability and failure of engineering rock mass under unloading conditions, characterized in that, Includes the following steps: S1. Obtain the geological, initial stress, unloading operation and rock physical and mechanical parameters of the unloading surface to be monitored, and determine the stress path of the true triaxial unloading test corresponding to the on-site unloading process accordingly. S2. Perform a true triaxial unloading test on the rock mass specimen according to the stress path, and simultaneously collect four types of physical parameters: stress, acoustic emission, electromagnetic radiation and infrared imaging. S3. Based on the characteristic inflection points of the volumetric strain curves in the experiment, the unloading stages are divided into the initial unloading stabilization period, the damage evolution acceleration period, and the instability and failure precursor period. The stage response laws of four types of parameters under each stage are extracted, and a multi-parameter stage response database is established. S4. For each unloading stage, the characteristic indicators of the four types of parameters are standardized, and the dynamic weights of each type of parameter in this stage are calculated based on principal component analysis. S5. During on-site monitoring, four types of parameters of the surface to be monitored are collected in real time, and their change patterns are matched with the phased response patterns in the database to identify the current unloading stage. The corresponding dynamic weights are called to calculate the comprehensive early warning index, and graded early warning information and disaster prevention feedback measures are output accordingly.

2. The multi-parameter dynamic collaborative early warning method for engineering rock mass instability and failure under unloading conditions as described in claim 1, characterized in that, In step S1: The unloading engineering surface to be monitored is the engineering surface in which the constraint stress in at least one direction of the rock mass is reduced and stress redistribution occurs due to excavation, mining, blasting, pressure relief, support adjustment or local constraint release. The unloading engineering face to be monitored includes at least one of the following: mine roadway excavation face, mining face, coal pillar edge area, tunnel excavation face, underground cavern excavation face, or slope excavation face.

3. The multi-parameter dynamic collaborative early warning method for engineering rock mass instability and failure under unloading conditions as described in claim 1, characterized in that, In step S1: Geological parameters include at least one of the following: stratum position, thickness, dip angle, geothermal temperature, structural distribution, degree of fracture development, water-bearing state, and permeability; The initial stress parameters include at least one of the stress distribution direction and stress magnitude; The unloading operation parameters include at least one of the following: unloading location, unloading direction, unloading rate, unloading duration, and unloading influence range; The physical and mechanical parameters of rock mass include at least one of density, water content, porosity, uniaxial compressive strength, uniaxial tensile strength, elastic modulus, and Poisson's ratio.

4. The multi-parameter dynamic collaborative early warning method for engineering rock mass instability and failure under unloading conditions as described in claim 1, characterized in that, In step S1, determining the stress path for the true triaxial unloading test includes: S11. Determine the direction and magnitude of the ground stress distribution on the unloading surface to be monitored based on the field test results; S12. Based on the stress variation law in different directions during the on-site unloading operation, determine the loading, holding or unloading relationship of the rock mass in different directions; S13. Determine the stress path for indoor true triaxial unloading tests based on the direction of ground stress distribution, the magnitude of ground stress, and the loading, holding, or unloading relationships in different directions.

5. The multi-parameter dynamic collaborative early warning method for engineering rock mass instability and failure under unloading conditions as described in claim 1, characterized in that, In step S2, the true triaxial unloading test includes: The rock mass specimen is loaded with stress levels corresponding to the initial three-dimensional stress state in the field in three mutually perpendicular loading directions. After entering the unloading stage, the stress in at least one loading direction is unloaded according to the stress change relationship corresponding to the actual unloading process in the field, and the stress in the other loading directions is loaded, maintained or adjusted according to the stress evolution law in the field until the rock mass specimen undergoes macroscopic instability failure or reaches the preset termination condition.

6. The multi-parameter dynamic collaborative early warning method for engineering rock mass instability and failure under unloading conditions as described in claim 5, characterized in that, The stress path of the true triaxial unloading test includes at least one of the following: staged unloading path, continuous unloading path, and rapid unidirectional unloading path; The graded unloading path is as follows: after loading the rock mass specimen to a three-dimensional stress state corresponding to the initial stress state on site, the stress in the target unloading direction is unloaded step by step according to the unloading amplitude and holding time corresponding to the gradual unloading process on site. The continuous unloading path is as follows: after loading the rock mass specimen to a three-dimensional stress state corresponding to the initial stress state on site, the stress in the target unloading direction is continuously unloaded according to the unloading rate corresponding to the continuous unloading process on site. The rapid unidirectional unloading path is as follows: after loading the rock mass specimen to a three-dimensional stress state corresponding to the initial stress state on site, the stress in the target unloading direction is rapidly unloaded according to the unloading rate corresponding to the rapid unloading process or the sudden release of local constraints on site.

7. The multi-parameter dynamic collaborative early warning method for engineering rock mass instability and failure under unloading conditions as described in claim 1, characterized in that, In step S3: Characteristic indicators of stress parameters include volumetric strain; Characteristic indicators of acoustic emission parameters include at least one of count rate, energy rate, cumulative count, cumulative energy, and b-value; The characteristic parameters of electromagnetic radiation include at least one of instantaneous amplitude, pulse number, cumulative pulse number, and average amplitude; The characteristic indicators of infrared imaging parameters include at least one of the following: temperature gradient, maximum temperature difference, area ratio of high-temperature region, and temperature variation coefficient.

8. The multi-parameter dynamic collaborative early warning method for engineering rock mass instability and failure under unloading conditions according to claim 1, characterized in that, In step S3, the unloading stage is divided according to the characteristic inflection point of the volumetric strain curve in the experiment, including: S31. In a true triaxial unloading test, based on the volumetric strain data collected or output by the true triaxial testing equipment, establish a volumetric strain curve that varies with time. S32. Perform slope change analysis or first derivative analysis on the volumetric strain curve to identify the location where the volumetric strain change rate continuously changes; S33. The location where the rate of change of volumetric strain changes continuously is taken as the characteristic inflection point; S34. Based on the aforementioned characteristic inflection point, the true triaxial unloading process is divided into the initial unloading stabilization period, the damage evolution acceleration period, and the instability and failure precursor period; S35. Based on the variation patterns of stress parameters, acoustic emission parameters, electromagnetic radiation parameters, and infrared imaging parameters in different unloading stages, the division results of the unloading stages are verified, and the stage response patterns corresponding to different unloading stages are formed.

9. The multi-parameter dynamic collaborative early warning method for engineering rock mass instability and failure under unloading conditions according to claim 1, characterized in that, Steps S4 and S5 include: S41. Perform Z-score standardization on the characteristic indices of the four types of parameters within the same unloading stage; S42. Calculate the correlation coefficient matrix based on the standardized feature indicators; S43. Solve for the eigenvalues ​​and eigenvectors of the correlation coefficient matrix, and select principal components with a cumulative contribution rate of not less than 85% according to the magnitude of the eigenvalues; S44. Calculate the comprehensive weight of each feature index based on the loading coefficients and variance contribution rates of the selected principal components; S45. Sum the comprehensive weights of the characteristic indices belonging to the same type of physical parameters to obtain the dynamic weights of stress parameters, acoustic emission parameters, electromagnetic radiation parameters and infrared imaging parameters under the corresponding unloading stage. S46. Normalize the weighted composite values ​​of the four types of parameters to the same numerical range; S51. During on-site monitoring, the real-time collected data on the changing patterns of the four types of parameters are matched with the phased response patterns of different unloading stages in the multi-parameter phased response database to determine the current unloading stage. S52. Based on the dynamic weights corresponding to the current unloading stage, the standardized comprehensive values ​​of the four types of parameters are weighted and summed to obtain the comprehensive early warning index; S53. Output blue, yellow, orange, or red warning information based on the comprehensive early warning index, and invoke the disaster prevention feedback measures corresponding to the warning information.

10. A multi-parameter dynamic collaborative early warning system for instability and failure of engineering rock mass under unloading conditions, characterized in that, It includes a multi-parameter sensing unit, a data acquisition unit, a transmission unit, a data processing unit, an early warning output unit, and a feedback control unit; The multi-parameter sensing unit is deployed on the unloading surface to be monitored. The multi-parameter sensing unit includes a stress monitoring component, an acoustic emission monitoring component, an electromagnetic radiation monitoring component, and an infrared thermal imaging monitoring component. The output terminals of the stress monitoring component, acoustic emission monitoring component, electromagnetic radiation monitoring component, and infrared thermal imaging monitoring component are respectively connected to the input terminal of the data acquisition unit; The output of the data acquisition unit is connected to the data processing unit through the transmission unit; The data processing unit is connected to the early warning output unit and the feedback control unit, respectively. The data processing unit is used to call the multi-parameter stage response database established by the true triaxial unloading test, and to execute the multi-parameter dynamic collaborative early warning method for engineering rock mass instability and failure under unloading conditions as described in any one of claims 1 to 9. The stress monitoring component includes a stress sensor and / or a strain sensor; The acoustic emission monitoring component includes an acoustic emission sensor; The electromagnetic radiation monitoring component includes an electromagnetic radiation monitor. The infrared thermal imaging monitoring component includes an infrared thermal imager.

Images