A rock degradation mode discrimination method and system considering energy time effect

Abstract

The present application relates to the technical field of rock dynamic mechanical test and impact failure mechanism analysis, and provides a rock degradation mode discrimination method and system considering energy time effect, and relates to the technical field of rock dynamic mechanical test and impact failure mechanism analysis. The method obtains incident wave, reflected wave and transmitted wave signals through Hopkinson pressure bar test, calculates dissipation energy time history and generates energy time density spectrum; according to peak time, the peak energy accumulation stage and the peak energy dissipation stage are divided, and energy time characteristics are extracted; combined with nuclear magnetic resonance detection, three-dimensional section scanning and fragment image acquisition, multi-scale degradation characteristics are obtained, a cross-scale correlation model between energy time characteristics and degradation characteristics is established, and damage dominant type, fracture dominant type and crushing dominant type criteria are generated. The method can improve the objectivity and accuracy of rock impact degradation mode discrimination.

Description

Technical Field

[0001] This invention relates to the field of rock dynamic mechanics testing and impact failure mechanism analysis, and in particular to a method and system for identifying rock degradation modes that considers energy-time effects. Technical Background

[0002] Rocks are often subjected to transient impact loads in engineering environments such as deep mining, tunnel blasting, and impact protection. Their failure modes may manifest as different states such as internal damage accumulation, crack propagation and penetration, or overall fragmentation.

[0003] Current rock impact failure analysis relies heavily on indicators such as peak stress, strain rate, total dissipated energy, fragment size, or post-test visual observation. While these indicators can reflect the intensity of failure, they are insufficient in expressing the variability in the failure process. Furthermore, for samples with similar energy dissipation levels, different lithologies, initial defects, and loading conditions can lead to significantly different degradation results. A single macroscopic indicator is insufficient to accurately distinguish the dominant failure type. Simultaneously, existing discrimination methods often emphasize post-test statistical results, lacking a comprehensive consideration of the correlation between internal damage, fracture development, and fragment morphology, easily leading to biased evaluation results. Some methods also rely on manual observation or empirical thresholds, which are highly subjective and fail to meet the needs for quantitative and stable discrimination of rock impact degradation patterns. Summary of the Invention

[0004] In view of this, embodiments of the present invention provide a method and system for identifying rock degradation modes that considers energy-time effects, in order to solve or alleviate the technical problems existing in the prior art.

[0005] The technical solution of this invention is implemented as follows: a method for identifying rock degradation modes considering energy-time effects, comprising the following steps:

[0006] S1. Perform a split Hopkinson bar impact test on the rock sample to obtain incident wave signal, reflected wave signal and transmitted wave signal, and obtain the dissipated energy time history and energy time density spectrum based on the incident wave signal, reflected wave signal and transmitted wave signal; perform nuclear magnetic resonance detection, three-dimensional cross-section scanning and fragment image acquisition on the rock sample after impact to obtain microscopic pore data, fracture surface point cloud data and fragment morphology data.

[0007] In this embodiment, specifically, a split Hopkinson bar test system is used to apply an impact load to the rock sample, and the incident wave signal is acquired by strain gauges set on the incident bar and the transmission bar. , reflected wave signal and transmitted wave signal Based on the acquired stress wave signals, the incident energy time history was calculated. Time history of reflection energy and transmission energy time history And obtain the dissipated energy time history. .

[0008] The incident energy time history, reflected energy time history, transmitted energy time history, and dissipated energy time history are calculated according to the following formulas:

[0009]

[0010] ;

[0011] ;

[0012] ;

[0013] in, The cross-sectional area of ​​the compression member. The elastic wave velocity in the compression bar. The elastic modulus of the compression bar. It is the integral variable.

[0014] In the process of obtaining dissipated energy Subsequently, time-frequency analysis was performed on the dissipated energy time history. A dynamic adaptive time window algorithm was established to divide the loading process into multiple time-domain sub-intervals, and the effective computational domain was selected by combining the stress uniformity quantification index. Within the effective computational domain, the fundamental energy time density function was defined, and a time scale variable was introduced. Constructing the multivariable energy-time density function using a normalized window function And based on the multivariate energy-time density function, the energy-time density spectrum is obtained.

[0015] In this embodiment, the multi-element energy time density function This can be extended to zero-order, first-order, and second-order tensor forms. Zero-order energy-time density tensor The double integral of the multivariable energy-time density function over the effective time domain and scale domain:

[0016]

[0017] Among them, the integration domain It is the Cartesian product of the effective time domain and scale domain.

[0018] First-order energy-time density tensor The gradient characteristic used to characterize energy change is a vector field composed of the partial derivatives of the multivariable energy-time density function with respect to time and time scale:

[0019]

[0020] in, The time gradient is used to characterize the rate of change of energy density over time. The scale gradient is used to characterize the rate of change of energy density with respect to the observation scale.

[0021] Second-order energy-time density tensor The Hessian matrix, composed of second-order partial derivatives, is used to characterize the distribution structure of energy in the time-scale space.

[0022]

[0023] in, This indicates the degree of concentration of energy density over time. This indicates the degree of concentration of energy density along the scale direction. This indicates the coupling strength between the time dimension and the scale dimension.

[0024] S2. Identify the peak moment of the energy interaction process based on the energy time density spectrum, and divide the energy interaction process into the pre-peak energy accumulation stage and the post-peak energy dissipation stage using the peak moment as the dividing point. Extract the energy time feature vector including the peak moment, half width at half maximum (WHM), average pre-peak energy injection rate, post-peak energy decrease rate, pre-peak and post-peak energy ratio, pre-peak and post-peak dominant frequency ratio, and time-frequency aggregation degree.

[0025] In this embodiment, specifically, the peak time This represents the time coordinate corresponding to when the energy density spectrum reaches its global maximum in the time domain. The peak time is used as the reference point. Using this as a dividing point, the energy interaction process is divided into the pre-peak energy accumulation stage and the post-peak energy dissipation stage. The pre-peak energy accumulation stage reflects the process of impact energy being concentrated and injected into the interior of the sample, driving crack initiation and propagation; the post-peak energy dissipation stage reflects the process of rapid energy release after the peak value, crack penetration, or fragment formation.

[0026] The energy time feature vector Including peak times Half height and width Average rate of energy injection before peak Post-peak energy decrease rate Pre-peak and post-peak energy ratio pre-peak and post-peak dominant frequency ratio and time-frequency clustering , can be represented as:

[0027]

[0028] Among them, half height width The duration of the peak value in the energy density spectrum reflects the duration of energy concentration; the average rate of energy injection before the peak. Used to characterize the rate of energy input during the pre-peak energy accumulation phase; and the post-peak energy decline rate. Used to characterize the rate of energy decay during the post-peak energy dissipation stage; pre-peak to post-peak energy ratio. Used to characterize the ratio of pre-peak accumulated energy to post-peak dissipated energy; pre-peak to post-peak dominant frequency ratio. Used to characterize the difference between the dominant frequency in the pre-peak and post-peak phases; time-frequency clustering. Used to characterize the concentration of energy in the peak region of the energy density spectrum relative to the total energy.

[0029] S3. Based on the microscopic pore data, fracture surface point cloud data and fragment morphology data described in S1, extract the microscopic damage characteristic parameters, macroscopic fracture characteristic parameters and overall fragmentation characteristic parameters respectively, and summarize them to form a multi-scale deterioration feature vector.

[0030] In this embodiment, specifically, the microscopic damage characteristic parameters are obtained through nuclear magnetic resonance (NMR) detection. An NMR spectrometer is used to perform relaxation time spectrum testing on the impacted rock sample to obtain the internal pore distribution data. The three-dimensional digital model of the rock sample is divided into multiple voxel units, and a three-dimensional voxelized pore model is obtained by inverting the relaxation time spectrum using a simulated annealing algorithm. Distance transformation and local maximum search are performed on the three-dimensional voxelized pore model, and the maximum sphere algorithm is used to extract pore volumes and throats, generating a pore network topology. Based on the three-dimensional voxelized pore model and the pore network topology, the porosity change is extracted. Pore ​​connectivity increment pore volume distribution skewness coefficient and mean coordination number , as a microscopic damage characteristic parameter.

[0031] The macroscopic fracture characteristic parameters were obtained through three-dimensional cross-sectional scanning. A three-dimensional laser cross-sectional scanner was used to acquire three-dimensional point cloud data of the fracture surface of the rock sample after impact, and the three-dimensional point cloud data was interpolated onto a uniform grid to generate a fracture surface height matrix. A two-dimensional variational mode decomposition method was used to decompose the fracture surface height matrix into multiple two-dimensional intrinsic mode components at different spatial scales. Corrected Grasselli roughness parameters and anisotropy indices were calculated on each two-dimensional intrinsic mode component, and the multi-scale decomposition weights were determined based on the energy proportion of each scale mode. The multi-scale fractal dimension of roughness was obtained based on the roughness parameter sequence and anisotropy index sequence at each scale. and the mean of the anisotropy index , as a macroscopic fracture characteristic parameter.

[0032] The overall fragmentation characteristic parameters are obtained through fragment image acquisition and processing. All fragments generated from the rock sample after the impact test are collected, and powdery debris is supplemented and measured. All fragments are laid flat on a background plate, and fragment images are acquired using a high-resolution industrial camera. The pixel size is converted to the actual size using a calibration ruler. The fragment images undergo grayscale conversion, filtering, automatic threshold binarization segmentation, and local centroid segmentation to divide adjacent or slightly adhered fragment image regions into multiple independent fragment regions. Based on these independent fragment regions, the major axis dimension, minor axis dimension, major axis-to-minor axis ratio distribution, edge angle, fragment particle size distribution, and fragment fractal dimension are extracted. Characteristic particle size Mean ratio of major axis to minor axis Standard deviation of major axis to minor axis , as a characteristic parameter of overall fragmentation.

[0033] The multi-scale degraded feature vector It can be represented as:

[0034]

[0035] The aforementioned multi-scale degradation feature vectors can simultaneously express the degradation state of rock samples at the micro-pore scale, macro-fracture surface scale, and overall fragment scale.

[0036] S4. Using multivariate statistical regression analysis, establish the mapping relationship between the energy time feature vector and the multiscale degradation feature vector to form a cross-scale correlation model.

[0037] In this embodiment, specifically, multiple sets of energy-time feature vectors under different impact conditions are collected. and multi-scale degraded feature vectors As training data, for each set of experimental data, the energy-time feature vector is used as the input variable, and the microscopic damage feature parameters, macroscopic fracture feature parameters, and overall fragmentation feature parameters are used as the output variables to establish microscopic damage mapping models, macroscopic fracture mapping models, and overall fragmentation mapping models, respectively.

[0038] In one implementation, the mesoscopic damage mapping model includes:

[0039]

[0040] Macroscopic fracture mapping models include:

[0041]

[0042] The overall fragmentation mapping model includes:

[0043]

[0044] in, This is the mapping function obtained through multivariate statistical regression analysis.

[0045] Simultaneously, a cross-scale correlation was established to show the transmission of microscopic damage characteristics to macroscopic fracture characteristics, and the transmission of macroscopic fracture characteristics to overall fragmentation characteristics. This cross-scale correlation can reflect the evolutionary path of rock impact degradation, from the expansion of pores to crack formation, and then from crack penetration to overall fragmentation.

[0046] S5. Based on the cross-scale correlation model, establish a degradation mode criterion based on energy time characteristics, and classify rock impact degradation modes into damage-dominated, fracture-dominated, and fragmentation-dominated types, and determine the threshold conditions for each type of criterion.

[0047] In this embodiment, specifically, the degradation mode obtained from experimental observation is used as the discrimination criterion. Contribution weight analysis is performed on each parameter in the energy-time characteristics to determine the core discrimination parameters and auxiliary discrimination parameters. Threshold conditions for each discrimination parameter are determined through receiver operating characteristic (ROC) curve analysis. The core discrimination parameters include the peak time. Average rate of energy injection before peak Time-frequency clustering and the ratio of pre-peak to post-peak energy The auxiliary discrimination parameters include half-width and half-height. Post-peak energy decrease rate and the ratio of pre-peak to post-peak dominant frequencies .

[0048] When the peak time is late, the average energy injection rate before the peak is low, and the time-frequency concentration is low, it is identified as damage-dominated. Damage-dominated type indicates that the impact energy process is relatively dispersed, the energy injection rate is low, and it mainly induces the expansion of pores and the initiation of microcracks inside the sample. The sample as a whole remains intact or only suffers slight spalling.

[0049] When the peak time, the average rate of energy injection before the peak, and the time-frequency concentration are all within the middle range, and the difference between the dominant frequencies before and after the peak is within the preset tolerance range, it is judged to be fracture-dominant. Fracture-dominant type means that the impact energy can promote crack propagation and form a through fracture surface, but does not cause the entire sample to be fully broken.

[0050] When the peak time is early, the average rate of energy injection before the peak is high, the time-frequency concentration is high, and the energy before and after the peak is relatively high, it is identified as a fracture-dominant type. The fracture-dominant type indicates that the impact energy is highly concentrated in a short period of time, leading to rapid crack penetration, propagation, and convergence, ultimately forming multiple fragments or an overall fractured state.

[0051] A rock degradation mode discrimination system considering energy time effects is provided for use in any of the rock degradation mode discrimination methods considering energy time effects described above. The system includes a multi-scale information acquisition subsystem, an energy-time density feature extraction subsystem, a cross-scale correlation modeling subsystem, and a degradation mode criterion generation subsystem.

[0052] The multi-scale information acquisition subsystem is used to acquire microscopic pore data, fracture surface point cloud data, and fragment morphology data of rock samples after impact, and output multi-scale degradation characteristics.

[0053] The energy-time density feature extraction subsystem is used to receive the energy-time density spectrum and the dissipated energy time history, perform stage division of the energy action process and energy-time feature extraction, and output the energy-time features.

[0054] The cross-scale correlation modeling subsystem is used to receive multiple sets of energy-time characteristics and multi-scale degradation characteristics under different impact conditions, and to establish a mapping relationship between energy-time characteristics and multi-scale degradation characteristics.

[0055] The degradation mode criterion generation subsystem is used to determine the criterion threshold based on the analysis results of the cross-scale correlation modeling subsystem, generate three types of degradation mode criteria: damage-dominated, fracture-dominated, and fragmentation-dominated, and output the rock degradation mode type and confidence score.

[0056] Furthermore, the multi-scale information acquisition subsystem includes a nuclear magnetic resonance detection unit, a three-dimensional cross-sectional scanning unit, and a fragment image acquisition and processing unit.

[0057] The nuclear magnetic resonance (NMR) detection unit includes an NMR analyzer main unit, a sample probe, a pulse sequence controller, a T2 spectrum inversion module, a three-dimensional voxelization inversion module with a built-in simulated annealing algorithm, a pore network model construction module with a built-in maximum sphere algorithm, and a micro-parameter calculation module. The NMR detection unit is used to output the porosity change. Pore ​​connectivity increment pore volume distribution skewness coefficient and mean coordination number .

[0058] The three-dimensional cross-section scanning unit includes a three-dimensional laser cross-section scanner, a point cloud preprocessing module, a multi-scale decomposition module with a built-in two-dimensional variational mode decomposition algorithm, a modified Grasselli roughness calculation module, and a macroscopic parameter calculation module. The macroscopic parameter calculation module is used to output the multi-scale fractal dimension of roughness. and the mean of the anisotropy index .

[0059] The fragment image acquisition and processing unit includes a high-resolution industrial camera, a calibration ruler, an image preprocessing module, a local centroid segmentation module, a geometric parameter extraction module based on second-order central moments, a maximum likelihood estimation module, a fractal dimension calculation module, a feature particle size calculation module, and an overall fragmentation parameter output module. The overall fragmentation parameter output module is used to output the fractal dimension of the fragments. Characteristic particle size Mean ratio of major axis to minor axis Standard deviation of major axis to minor axis .

[0060] Furthermore, the energy-time density feature extraction subsystem includes an energy-time density spectrum input interface, a peak time identification module, a stage division module, a pre-peak feature extraction module, a post-peak feature extraction module, a comprehensive feature calculation module, and an energy-time feature output module. The energy-time density spectrum input interface is used to receive the energy-time density spectrum. and dissipation energy time history The peak moment identification module is used to identify the peak moment of the energy interaction process. The phase division module is used to divide peak time. The system divides the peak into a pre-peak energy accumulation stage and a post-peak energy dissipation stage. The pre-peak feature extraction module extracts the average pre-peak energy injection rate, pre-peak accumulated energy, and pre-peak dominant frequency. The post-peak feature extraction module extracts the post-peak energy decrease rate, post-peak dissipated energy, and post-peak dominant frequency. The comprehensive feature calculation module calculates the full width at half maximum (FWHM), pre-peak to post-peak energy ratio, pre-peak to post-peak dominant frequency ratio, and time-frequency clustering. The energy-time feature output module outputs the energy-time feature vector. .

[0061] Furthermore, the cross-scale correlation modeling subsystem includes a training data management module, a microscopic damage mapping modeling module, a macroscopic fracture mapping modeling module, a global fragmentation mapping modeling module, a cross-scale correlation establishment module, and a model reliability evaluation module. The training data management module is used to store data under different impact conditions. The data pairs are as follows: the micro-damage mapping modeling module is used to establish the mapping relationship between energy-time characteristics and micro-damage characteristic parameters; the macro-fracture mapping modeling module is used to establish the mapping relationship between energy-time characteristics and macro-fracture characteristic parameters; the overall fragmentation mapping modeling module is used to establish the mapping relationship between energy-time characteristics and overall fragmentation characteristic parameters; the cross-scale correlation establishment module is used to establish the transmission relationship between micro-damage characteristics, macro-fracture characteristics, and overall fragmentation characteristics; and the model reliability evaluation module is used to evaluate the fitting degree and discrimination reliability of the mapping model.

[0062] Furthermore, the degradation mode criterion generation subsystem includes a feature weight analysis module, a threshold determination module, a damage-dominant criterion module, a fracture-dominant criterion module, a fragmentation-dominant criterion module, and a comprehensive discrimination module. The feature weight analysis module determines the contribution of each parameter in the energy-time characteristics to degradation mode discrimination; the threshold determination module determines the threshold conditions for each discrimination parameter based on receiver operating characteristic (ROC) curve analysis; the damage-dominant criterion module outputs damage-dominant discrimination results; the fracture-dominant criterion module outputs fracture-dominant discrimination results; the fragmentation-dominant criterion module outputs fragmentation-dominant discrimination results; and the comprehensive discrimination module sequentially processes the input data through the three criterion modules and outputs the degradation mode type and confidence score.

[0063] Compared with the prior art, the present invention has the following advantages:

[0064] This invention incorporates the energy interaction time into the rock impact degradation mode discrimination process. It not only focuses on the amount of dissipated energy, but also further considers energy time characteristics such as peak time, half width at half maximum (FWHM), average rate of energy injection before the peak, rate of energy decrease after the peak, energy ratio before and after the peak, ratio of dominant frequencies before and after the peak, and time-frequency aggregation. This allows for a more accurate distinction of the impact of energy interaction processes on rock degradation modes under different impact conditions.

[0065] This invention obtains microscopic damage characteristic parameters, macroscopic fracture characteristic parameters, and overall fragmentation characteristic parameters through nuclear magnetic resonance detection, three-dimensional cross-sectional scanning, and fragment image acquisition and processing. It realizes multi-scale degradation characterization from microscopic pores and macroscopic fracture surfaces to overall fragment morphology, avoiding the discrimination bias caused by single-scale evaluation.

[0066] This invention establishes a mapping relationship between energy time feature vector and multi-scale degradation feature vector by using multivariate statistical regression analysis, forming a cross-scale correlation model that can reflect the transmission relationship between energy input time features and porosity expansion, fracture surface formation and overall fragmentation.

[0067] This invention establishes three criteria for deterioration modes: damage-dominated, fracture-dominated, and fragmentation-dominated. By using threshold conditions and confidence scores to output the discrimination results, it improves the objectivity and repeatability of rock impact deterioration mode discrimination.

[0068] The system provided by this invention consists of a multi-scale information acquisition subsystem, an energy-time-density feature extraction subsystem, a cross-scale correlation modeling subsystem, and a degradation mode criterion generation subsystem. The data flow relationship between each subsystem is clear, and it can realize a complete process from experimental data acquisition, feature extraction, cross-scale modeling to mode discrimination output, which is convenient for integration and application in rock impact testing platforms. Attached Figure Description

[0069] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0070] Figure 1 This is a flowchart of the method of the present invention;

[0071] Figure 2 This is a flowchart illustrating the multi-scale degradation information acquisition process of the present invention.

[0072] Figure 3 This is a schematic diagram of the energy-time density spectrum stage division process of the present invention;

[0073] Figure 4 This is a schematic diagram illustrating the relationship between the nuclear magnetic resonance T2 spectrum and aperture distribution of the present invention;

[0074] Figure 5 This is a schematic diagram of the simulated annealing convergence curve for the three-dimensional voxelization inversion of the present invention.

[0075] Figure 6 This is a schematic diagram of the pore network model of the present invention;

[0076] Figure 7 This invention relates to 2D-VMD multiscale decomposition;

[0077] Figure 8 This is a schematic diagram of Grasselli roughness and anisotropy analysis of the fracture surface in this invention.

[0078] Figure 9 This is a schematic diagram of the fragment morphology statistical process of the present invention;

[0079] Figure 10 This is a comparative illustration of the three degradation modes of the present invention: damage-dominated, fracture-dominated, and fragmentation-dominated.

[0080] Figure 11 This is the system module structure of the present invention. Detailed Implementation

[0081] The embodiments of the present invention will now be described in detail with reference to the accompanying drawings. It should be understood that the following embodiments are for illustrative purposes only and are not intended to limit the scope of protection of the present invention. Equivalent substitutions or modifications made by those skilled in the art without departing from the inventive concept should fall within the scope of protection of the present invention.

[0082] The embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0083] Example 1

[0084] This embodiment provides a rock degradation mode discrimination method that considers energy-time effects, including five main steps: impact test data acquisition, energy-time feature extraction, multi-scale degradation feature extraction, cross-scale correlation model establishment, and degradation mode criterion generation.

[0085] First, a split Hopkinson bar impact test was conducted on the rock sample. During the test, strain gauges on the incident bar collected incident and reflected wave signals, while strain gauges on the transmission bar collected transmitted wave signals. The collected incident, reflected, and transmitted wave signals were then conditioned, filtered, and converted from analog to digital, serving as the basis for energy time history calculations.

[0086] Based on the incident wave signal, reflected wave signal, and transmitted wave signal, the incident energy time history, reflected energy time history, transmitted energy time history, and dissipated energy time history are calculated. Then, time-frequency analysis is performed on the dissipated energy time history to obtain the energy-time density spectrum. The energy-time density spectrum can express the distribution of energy density in the time and frequency dimensions, thus providing a basis for identifying the concentrated input stage, dissipation and decay stage, and frequency component changes in the energy interaction process.

[0087] After obtaining the energy time density spectrum, the peak moments of the energy interaction process are identified. At peak time Using this as a dividing point, the energy interaction process is divided into a pre-peak energy accumulation stage and a post-peak energy dissipation stage. The pre-peak energy accumulation stage mainly reflects the process of impact energy gradually being injected into the rock sample and causing pore expansion and microcrack initiation; the post-peak energy dissipation stage mainly reflects the energy release process during crack expansion, penetration, and fragment formation.

[0088] Based on the above stage division, energy-time feature vectors are extracted. Peak time Used to characterize the temporal location of energy concentration; half-width at half-maximum. Used to characterize the duration of energy concentration; average rate of energy injection before the peak. Used to reflect the rate of energy accumulation before the peak and the rate of energy decrease after the peak. Used to reflect the rate of energy decay after the peak; pre-peak to post-peak energy ratio. Used to characterize the ratio of pre-peak accumulated energy to post-peak dissipated energy; pre-peak to post-peak dominant frequency ratio. Used to reflect changes in the dominant timescale of pre- and post-peak energy effects; time-frequency aggregation. It is used to reflect the degree of energy concentration in the time-frequency plane.

[0089] Subsequently, multi-scale degradation information was collected from the impacted rock samples. At the mesoscale, the T2 relaxation time spectra of the impacted samples were obtained using nuclear magnetic resonance (NMR). A pore model was obtained through three-dimensional voxelization inversion, and the pore network topology was constructed based on the maximum sphere algorithm to obtain the porosity change. Pore ​​connectivity increment pore volume distribution skewness coefficient and mean coordination number For macroscopic scales, a 3D laser cross-section scanner is used to acquire 3D point cloud data of the fracture surface. After point cloud preprocessing, mesh interpolation, and 2D variational mode decomposition, the multi-scale fractal dimension of roughness is obtained. and the mean of the anisotropy index For the overall scale, all fragments after the impact were collected and image-processed. The fractal dimension of the fragments was obtained through image segmentation and geometric parameter extraction. Characteristic particle size Mean ratio of major axis to minor axis Standard deviation of major axis to minor axis .

[0090] After feature extraction is completed, a multi-scale degraded feature vector is constructed. and multiple groups of different impact conditions The data input cross-scale correlation modeling subsystem employs multivariate statistical regression analysis to establish mapping relationships between energy-time characteristics and microscopic damage characteristics, macroscopic fracture characteristics, and overall fragmentation characteristics, and to establish cross-scale correlation relationships for the transmission of microscopic damage characteristics to macroscopic fracture characteristics and vice versa.

[0091] In this embodiment, the cross-scale correlation model can specifically take the following form:

[0092] Microscopic damage mapping model:

[0093] ;

[0094] ;

[0095] ;

[0096] Macroscopic fracture mapping model:

[0097] ;

[0098] ;

[0099] Overall fragmentation mapping model:

[0100] ;

[0101] ;

[0102] ;

[0103] Cross-scale correlation equations:

[0104] ;

[0105] ;

[0106] ;

[0107] The above mapping relationship is used to illustrate the correlation between energy time characteristics and multi-scale degradation characteristics. In practical applications, the regression coefficients can be trained and updated according to the scale of experimental data, rock type and impact conditions.

[0108] Finally, based on the analysis results of the cross-scale correlation model and the degradation patterns obtained from experimental observations, the criterion thresholds were determined using receiver operating characteristic (ROC) curve analysis, and three types of degradation pattern criteria were established: damage-dominated, fracture-dominated, and fragmentation-dominated.

[0109] In this embodiment, specifically, the criterion threshold includes a first threshold. Second threshold Third threshold The sixth threshold The seventh threshold and convergence tolerance .

[0110] Specifically, = ,

[0111] = ,

[0112] = ,

[0113] = ,

[0114] = ,

[0115] = .

[0116] When the peak time of the sample to be judged is late, the average rate of energy injection before the peak is low, and the time-frequency concentration is low, it is comprehensively judged as damage-dominant type; when the peak time, the average rate of energy injection before the peak, and the time-frequency concentration are all in the middle range, and the difference between the ratio of the dominant frequency before and after the peak and 1 is less than the preset tolerance, it is comprehensively judged as fracture-dominant type; when the peak time is early, the average rate of energy injection before the peak is high, the time-frequency concentration is high, and the ratio of the energy before and after the peak is high, it is comprehensively judged as fragmentation-dominant type.

[0117] Example 2

[0118] This embodiment provides a rock degradation mode discrimination system that considers energy-time effects, including a multi-scale information acquisition subsystem, an energy-time-density feature extraction subsystem, a cross-scale correlation modeling subsystem, and a degradation mode criterion generation subsystem.

[0119] The multi-scale information acquisition subsystem is used to acquire degradation information of impact-damaged rock samples at three scales: microscopic, macroscopic, and global. The multi-scale information acquisition subsystem includes a nuclear magnetic resonance (NMR) detection unit, a three-dimensional cross-sectional scanning unit, and a fragment image acquisition and processing unit.

[0120] The nuclear magnetic resonance (NMR) detection unit includes an NMR analyzer main unit, a sample probe, a pulse sequence controller, a T2 spectrum inversion module, a three-dimensional voxelization inversion module, a pore network model construction module, and a mesoscopic parameter calculation module. The NMR analyzer main unit generates radio frequency pulses and receives echo signals; the sample probe is used to hold the impacted rock sample; the pulse sequence controller controls the NMR test parameters; the T2 spectrum inversion module converts the echo signal into a T2 relaxation time distribution curve; the three-dimensional voxelization inversion module incorporates a simulated annealing algorithm to obtain a three-dimensional voxelized pore model based on the T2 spectrum; the pore network model construction module incorporates a maximum sphere algorithm to extract pore volumes and throats from the three-dimensional voxelized pore model and generate the pore network topology; and the mesoscopic parameter calculation module outputs the porosity change, pore connectivity increment, pore volume distribution skewness coefficient, and average coordination number.

[0121] The 3D cross-section scanning unit includes a 3D laser cross-section scanner, a point cloud preprocessing module, a multi-scale decomposition module, a modified Grasselli roughness calculation module, and a macroscopic parameter calculation module. The 3D laser cross-section scanner acquires 3D point cloud data of the fracture surface of a rock sample after impact; the point cloud preprocessing module performs denoising, registration, and interpolation on the point cloud data; the multi-scale decomposition module incorporates a two-dimensional variational mode decomposition algorithm to decompose the fracture surface height matrix into multiple two-dimensional intrinsic mode components at different spatial scales; the modified Grasselli roughness calculation module calculates roughness parameters at different scale modes; and the macroscopic parameter calculation module outputs the multi-scale fractal dimension and the mean anisotropy index of the roughness.

[0122] The fragment image acquisition and processing unit includes a high-resolution industrial camera, a calibration ruler, an image preprocessing module, a local centroid segmentation module, a geometric parameter extraction module based on the second-order central moment, a maximum likelihood estimation module, a fractal dimension calculation module, a feature particle size calculation module, and an overall fragmentation parameter output module. The high-resolution industrial camera is used to acquire fragment images; the calibration ruler is used to perform ratio conversion between pixel size and actual size; the image preprocessing module is used for grayscale conversion, filtering, and binarization segmentation; the local centroid segmentation module is used to segment adjacent or slightly adhered fragment image regions; the geometric parameter extraction module is used to extract the major axis dimension, minor axis dimension, major-minor axis ratio, and edge angles; the maximum likelihood estimation module is used to estimate the major-minor axis ratio distribution parameters; the fractal dimension calculation module is used to calculate the fragment fractal dimension; the feature particle size calculation module is used to calculate the feature particle size; and the overall fragmentation parameter output module outputs the fragment fractal dimension, feature particle size, mean major-minor axis ratio, and standard deviation of the major-minor axis ratio.

[0123] The energy-time density feature extraction subsystem receives the energy-time density spectrum and dissipated energy time history, and outputs an energy-time feature vector. The subsystem includes an energy-time density spectrum input interface, a peak time identification module, a stage division module, a pre-peak feature extraction module, a post-peak feature extraction module, a comprehensive feature calculation module, and an energy-time feature output module. The energy-time density spectrum input interface receives the energy-time density spectrum. and dissipation energy time history Peak moment identification module identifies peak moment. The phase division module is based on the peak time. The system divides the peak into a pre-peak energy accumulation stage and a post-peak energy dissipation stage. The pre-peak feature extraction module extracts the average pre-peak energy injection rate, pre-peak accumulated energy, and pre-peak dominant frequency. The post-peak feature extraction module extracts the post-peak energy decrease rate, post-peak dissipated energy, and post-peak dominant frequency. The comprehensive feature calculation module calculates the half-width at half-maximum (FWHM), pre-peak to post-peak energy ratio, pre-peak to post-peak dominant frequency ratio, and time-frequency clustering. The energy-time feature output module outputs the energy-time feature vector. .

[0124] The cross-scale correlation modeling subsystem is used to establish the mapping relationship between energy-time characteristics and multi-scale degradation characteristics. This subsystem includes a training data management module, a micro-damage mapping modeling module, a macro-fracture mapping modeling module, a global fragmentation mapping modeling module, a cross-scale correlation establishment module, and a model reliability evaluation module. The training data management module stores data under different impact conditions. The data sets include: a micro-damage mapping modeling module to establish the mapping relationship between energy-time characteristics and porosity changes, pore connectivity increments, pore volume distribution skewness coefficients, and average coordination number; a macro-fracture mapping modeling module to establish the mapping relationship between energy-time characteristics and roughness multi-scale fractal dimension and the mean anisotropy index; a global fragmentation mapping modeling module to establish the mapping relationship between energy-time characteristics and fragment fractal dimension, characteristic particle size, mean long-axis to short-axis ratio, and standard deviation of long-axis to short-axis ratio; a cross-scale correlation establishment module to establish the correlation relationship between the transmission of micro-damage characteristics to macro-fracture characteristics and the transmission of macro-fracture characteristics to global fragmentation characteristics; and a model reliability evaluation module to evaluate the model fitting effect and discrimination reliability.

[0125] The degradation mode criterion generation subsystem is used to determine the criterion thresholds and output the degradation mode type. This subsystem includes a feature weight analysis module, a threshold determination module, a damage-dominant criterion module, a fracture-dominant criterion module, a breakage-dominant criterion module, and a comprehensive discrimination module. The feature weight analysis module determines the contribution weight of each energy-time characteristic parameter to the degradation mode; the threshold determination module determines the threshold conditions for each discrimination parameter through receiver operating characteristic (ROC) curve analysis; the damage-dominant criterion module determines whether the input sample belongs to the damage-dominant type; the fracture-dominant criterion module determines whether the input sample belongs to the fracture-dominant type; the breakage-dominant criterion module determines whether the input sample belongs to the breakage-dominant type; and the comprehensive discrimination module integrates the results of the three criterions and outputs the rock degradation mode type and confidence score.

[0126] Example 3

[0127] This embodiment uses the Hopkinson bar test of a granite sample under an impact pressure of 0.5 MPa as an example to illustrate the complete implementation process of the present invention and the identification of the fracture-dominant degradation mode.

[0128] Step 1: Impact Testing and Energy-Time Feature Extraction

[0129] A standard granite specimen with a diameter of 50 mm and a height of 50 mm was prepared, with an initial porosity of... Initial pore connectivity A split Hopkinson bar testing system was used, with an impact pressure of 0.5 MPa. The bar diameter was 50 mm, the elastic modulus was 210 GPa, and the elastic wave velocity was 5090 m / s. The sampling frequency was set to 1 MHz.

[0130] Acquire incident wave signal , reflected wave signal and transmitted wave signal According to the one-dimensional stress wave theory, the total incident energy is 27.08 J, the total reflected energy is 12.67 J, the total transmitted energy is 3.90 J, and the total dissipated energy is 10.51 J.

[0131] Based on existing multidimensional characterization methods for energy-time density, the energy-time density spectrum is calculated. and dissipation energy time history Peak time of the energy density spectrum Half height and width .

[0132] by The energy action phase is divided into stages based on the boundary point. Pre-peak accumulation phase ( Pre-peak energy accumulation Average rate of energy injection before peak

[0133]

[0134] Pre-peak dominant frequency Post-peak dissipation phase ( Post-peak energy dissipation rate of energy decrease after peak Post-peak dominant frequency Calculate the pre-peak and post-peak energy ratio. The pre-peak and post-peak dominant frequencies are compared Time-frequency clustering .

[0135] Constructing energy-time feature vectors

[0136]

[0137] Step 2: Multi-scale degradation information collection

[0138] (1) Microscale—Nuclear magnetic resonance detection and three-dimensional voxelization inversion

[0139] CPMG pulse sequence testing was performed on the impact-exposed specimens using a nuclear magnetic resonance spectrometer. Echo interval. Waiting time The number of echoes was 6000, and the number of scans was 64. The result was obtained through regularized inversion. The spectrum exhibits a distinct bimodal characteristic: a small pore peak. The center time is approximately 2.8 ms, corresponding to an aperture of approximately 28 mm. m; large pore peak The center time is approximately 63 ms, corresponding to an aperture of approximately 630 nm. m.

[0140] The sample voxel resolution is set to 100. m, the grid is divided into Individual prime, total prime number .through The target porosity is estimated to be 10.55% based on spectral integration. Simulated annealing inversion parameters: initial temperature. Termination temperature The cooling rate is 0.95, with 500 iterations per temperature, for a total of 10,000 iterations. The objective function eventually converges to 0.0023.

[0141] The maximum sphere algorithm was used to extract the pore network: After distance transformation, 2847 local maxima were detected as initial spheres. After hierarchical filtering (overlap threshold of 50%), 856 principal spheres were retained. After pore body merging (merging threshold of 1.3 times the radius sum), 312 pore bodies were obtained. 718 throats were identified.

[0142] Detailed parameter calculation results: Porosity after impact , Pore ​​connectivity , ; pore volume distribution skewness coefficient (Positive skewness indicates the formation of a large number of small-sized microcracks); average coordination number .

[0143] (2) Three-dimensional scanning of the fracture surface at the macro scale

[0144] Using a 3D laser profile scanner at 10 Obtain fracture surface point cloud data at m resolution, total One measurement point, scanning area Range of height fluctuations to After removing the sloping trend surface, 2D-VMD decomposition is performed. Each scale mode.

[0145] Results of analysis at various scales:

[0146] 1 25μm 8.3% 12.4 0.15 2 85μm 18.7% 18.9 0.28 3 240μm 31.2% 24.6 0.42 4 520 μm 24.5% 28.3 0.51 5 1.2 mm 12.8% 31.7 0.38 6 3.5 mm 4.5% 35.2 0.22

[0147] Roughness multiscale fractal dimension obtained through double log-linear regression Weighted by energy, mean anisotropy index .

[0148] (3) Overall scale - fragment morphology statistics

[0149] The sample was completely crushed, and a total of 47 fragments were collected, of which 31 fragments had a particle size greater than 2 mm, and the fine fragments smaller than 2 mm were combined and weighed. Images of the fragments were taken using an industrial camera (resolution...). (pixels), after local centroid segmentation and geometric parameter extraction, the main statistical results are as follows:

[0150] Major axis to minor axis ratio Based on the maximum likelihood estimation of the log-normal distribution, , , , Fractal dimension of fragments (Double log-fit slope) Characteristic particle size .

[0151] Step 3: Construction of Multi-Scale Degraded Feature Vectors

[0152] By summing the detection parameters at three scales, a multi-scale degradation feature vector is constructed:

[0153] .

[0154] Step 4: Energy-Degradation Cross-Scale Correlation Modeling

[0155] Based on 15 sets of experimental data (3 sets each) under five impact pressures of 0.3 MPa, 0.4 MPa, 0.5 MPa, 0.6 MPa, and 0.7 MPa, a cross-scale correlation model was established.

[0156] Microscopic damage mapping model ( (All values ​​are greater than 0.85)

[0157]

[0158]

[0159] .

[0160] Macroscopic fracture mapping model:

[0161]

[0162]

[0163] Overall fragmentation mapping model:

[0164]

[0165]

[0166]

[0167] Cross-scale correlation equations:

[0168]

[0169]

[0170] .

[0171] Step 5: Establishing Degradation Mode Criteria

[0172] Based on ROC curve analysis of 15 sets of experimental data, the criterion threshold was determined.

[0173] First threshold (Sensitivity 0.91, Specificity 0.88)

[0174] Second threshold (Sensitivity 0.89, Specificity 0.91)

[0175] Third threshold (Sensitivity 0.93, Specificity 0.85)

[0176] Sixth threshold (Sensitivity 0.87, Specificity 0.92)

[0177] Seventh threshold (Sensitivity 0.94, Specificity 0.89)

[0178] Convergence tolerance (Sensitivity 0.90, Specificity 0.86)

[0179] Energy-time feature vector in this embodiment

[0180] .

[0181] Three criteria are used in sequence:

[0182] Damage-dominant criteria: (Not satisfied) (larger conditions) (satisfy), (Not satisfied) If the conditions are too low (not all conditions are met), it is not considered a damage-dominant type.

[0183] Fracture-dominant criterion: ,exist Within the range (satisfied); Not satisfied ; (Not satisfied), therefore not classified as fracture-dominant type.

[0184] Fractional dominant criterion: (Not satisfied) (smaller conditions) (satisfy), (Satisfied), not all conditions are met, but the confidence level is 0.72.

[0185] Further investigation: Although slightly higher than 150 s threshold, but near Threshold, Significantly greater than 3.0 Significantly greater than 0.60. The overall assessment is that the degradation is dominated by breakage, with a confidence level of 0.85.

[0186] Verification of the discrimination results: After the test, the sample was completely broken into 47 fragments, and the fractal dimension of the fragments was... ( (Conforming to the characteristics of a fracture-dominant type), fracture surface roughness fractal dimension ( Characteristic particle size (Smaller, fully broken down). Microscopic porosity increase Incremental connectivity This indicates that numerous microcracks initiated and connected during the impact process, eventually leading to overall fragmentation. The judgment results are in complete agreement with experimental observations.

[0187] Example 4

[0188] This embodiment uses the Hopkinson bar test of sandstone samples under an impact pressure of 0.3 MPa as an example to illustrate the identification of damage-dominated degradation modes.

[0189] Step 1: Impact Testing and Energy-Time Feature Extraction

[0190] A standard sandstone sample with a diameter of 50 mm and a height of 50 mm was prepared, with an initial porosity of... Initial pore connectivity The same test system as in Example 1 was used, with an impact pressure of 0.3 MPa.

[0191] Stress wave signals were collected, and the total incident energy was calculated to be 15.43 J, the total reflected energy was 6.85 J, the total transmitted energy was 1.76 J, and the total dissipated energy was 6.82 J.

[0192] Based on the energy-time density multidimensional characterization method, the energy-time density spectrum is calculated. Peak time Half height and width .

[0193] Pre-peak accumulation stage ( ): , , Post-peak dissipation phase ( ): , , .calculate , , .

[0194] Constructing energy-time feature vectors

[0195] .

[0196] Step 2: Multi-scale degradation information collection

[0197] (1) Microscale - Nuclear Magnetic Resonance Detection

[0198] The specimen remained intact after the impact, with no obvious macroscopic cracks observed. (Nuclear magnetic resonance) The spectrum shows small pore peaks ( Approximately 1.5 ms, corresponding to an aperture of approximately 15 mm. The amplitude of m) increased significantly compared to before the impact, while the change in the large pore peak was not significant.

[0199] Simulated annealing inversion and maximum sphere algorithm extraction results: porosity after impact , Pore ​​connectivity , ; (Significant positive skewness, numerous microcracks initiate but do not penetrate); .

[0200] (2) Macroscale - Three-dimensional scanning of fracture surface

[0201] The sample did not form a continuous fracture surface; only several discontinuous microcracks appeared on the outer surface. The surface with the largest crack was selected for scanning. 2D-VMD decomposition was performed. Each scale mode.

[0202] Results of analysis at various scales: to They are respectively , , , , The energy proportions were 12.5%, 25.3%, 35.8%, 18.2%, and 8.2%, respectively. (Lower level indicates simple crack surface undulation) (Lower, weak anisotropy).

[0203] (3) Overall scale - fragment morphology statistics

[0204] The sample remained intact and unbroken. A small amount of debris that detached from the surface was counted; all particles were less than 2 mm in diameter, with a total mass of 0.82 g. (Large-diameter fragments predominate, but the amount of debris is extremely small). Not applicable (no effective fragment distribution was formed). and not applicable.

[0205] Step 3: Construction of Multi-Scale Degraded Feature Vectors

[0206] Constructing multi-scale degraded feature vectors:

[0207]

[0208] (Note: The sample was not broken.) , , Unable to calculate, represented as null.

[0209] Step 4: Degradation Mode Criteria Determination

[0210] Will

[0211] ;

[0212] Three criteria are used in sequence:

[0213] Damage-dominant criteria: (satisfy (larger conditions) (satisfy (smaller conditions); (Not satisfied) If two out of the three conditions are met, with a confidence level of 0.78, the propensity to be identified as damage-dominant.

[0214] Fracture-dominant criterion: Exceeding The upper limit of the range is not met; it is not determined to be a fracture-dominant type.

[0215] Fractional dominant criterion: (Not satisfied) (smaller conditions); (Not satisfied) (Under relatively large conditions); therefore, it is not classified as a break-dominant type.

[0216] Overall judgment result: damage-dominated degradation, confidence level 0.82.

[0217] Verification of the discrimination results: The specimen remained intact after the test and no through fracture surface was formed, which is completely consistent with the characteristics of damage-dominated degradation.

[0218] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any person skilled in the art can easily conceive of various variations or substitutions within the technical scope disclosed in the present invention, and these should all be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for identifying rock degradation patterns that considers energy-time effects, characterized in that, Includes the following steps: S1. A split Hopkinson bar impact test is performed on the rock sample to obtain the incident wave signal, reflected wave signal, and transmitted wave signal. Based on the incident wave signal, reflected wave signal, and transmitted wave signal, the dissipated energy time history and energy time density spectrum are obtained. Nuclear magnetic resonance detection, three-dimensional cross-section scanning, and fragment image acquisition are performed on the rock sample after impact to obtain microscopic pore data, fracture surface point cloud data, and fragment morphology data. S2. Identify the peak moment of the energy interaction process based on the energy time density spectrum, and divide the energy interaction process into the pre-peak energy accumulation stage and the post-peak energy dissipation stage using the peak moment as the dividing point. Extract the energy time feature vector including the peak moment, half width at half maximum (WHM), average rate of pre-peak energy injection, rate of post-peak energy decrease, pre-peak and post-peak energy ratio, pre-peak and post-peak dominant frequency ratio, and time-frequency aggregation degree. S3. Based on the micropore data, fracture surface point cloud data and fragment morphology data described in S1, extract the micro-damage characteristic parameters, macro-fracture characteristic parameters and overall fragmentation characteristic parameters respectively, and summarize them to form a multi-scale deterioration feature vector. S4. Using multivariate statistical regression analysis, establish the mapping relationship between the energy time feature vector and the multiscale degradation feature vector to form a cross-scale correlation model; S5. Based on the cross-scale correlation model, establish a degradation mode criterion based on energy time characteristics, and classify rock impact degradation modes into damage-dominated, fracture-dominated, and fragmentation-dominated types, and determine the threshold conditions for each type of criterion.

2. The method for identifying rock degradation patterns considering energy-time effects according to claim 1, characterized in that, The energy time feature vector includes peak time, half width at half maximum (FWHM), average rate of energy injection before peak, rate of energy decrease after peak, energy ratio before and after peak, ratio of dominant frequency before and after peak, and time-frequency clustering. Wherein, the peak moment is the time coordinate corresponding to when the energy-time density spectrum reaches its global maximum value in the time domain; The half-width at half-maximum (WHM) is the duration of time corresponding to the peak value of the energy-time density spectrum; the average rate of energy injection before the peak is used to characterize the speed of energy input during the energy accumulation phase before the peak. The post-peak energy decrease rate is used to characterize the rate of energy decay during the post-peak energy dissipation stage. The pre-peak and post-peak energy ratio is used to characterize the proportional relationship between the energy accumulated before the peak and the energy dissipated after the peak. The pre-peak and post-peak dominant frequency ratio is used to characterize the difference between the dominant frequency in the pre-peak stage and the dominant frequency in the post-peak stage. The time-frequency concentration is used to characterize the degree of energy concentration in the peak region of the energy-time density spectrum relative to the total energy.

3. The method for identifying rock degradation modes considering energy-time effects according to claim 1, characterized in that, The extraction process of the microscopic damage feature parameters includes: The relaxation time spectrum of the impacted rock sample was measured using a nuclear magnetic resonance spectrometer to obtain data on the internal pore distribution of the sample. The three-dimensional digital model of the rock sample was divided into multiple voxel units, and the three-dimensional voxelized pore model was obtained by inverting the relaxation time spectrum using the simulated annealing algorithm. Distance transformation and local maximum search are performed on the three-dimensional voxelized pore model, and the pore volume and throat are extracted using the maximum sphere algorithm to generate the pore network topology. Based on the three-dimensional voxelized pore model and pore network topology, the porosity change, pore connectivity increment, pore volume distribution skewness coefficient, and average coordination number are extracted as the microscopic damage characteristic parameters.

4. The method for identifying rock degradation modes considering energy-time effects according to claim 1, characterized in that, The extraction process of the macroscopic fracture characteristic parameters includes: A three-dimensional laser cross-section scanner was used to acquire three-dimensional point cloud data of the fracture surface of the rock sample after impact, and the three-dimensional point cloud data was interpolated onto a uniform grid to generate a fracture surface height matrix. The fracture surface height matrix is ​​decomposed into multiple two-dimensional intrinsic mode components with different spatial scales using a two-dimensional variational mode decomposition method. The modified Grasselli roughness parameters and anisotropy indices are calculated on each two-dimensional intrinsic modal component, and the multi-scale decomposition weights are determined based on the energy proportion of each scale mode. Based on the roughness parameter sequences and anisotropy index sequences at each scale, the multi-scale fractal dimension of roughness and the mean of the anisotropy index are obtained and used as the macroscopic fracture characteristic parameters.

5. The method for identifying rock degradation modes considering energy-time effects according to claim 1, characterized in that, The extraction process of the overall fragmentation characteristic parameters includes: Collect all fragments produced by the rock sample after the impact test, and supplement and measure the powdery debris; All the fragments were laid flat on the background board, and images of the fragments were captured using a high-resolution industrial camera. The ratio of pixel size to actual size was calculated using a calibration ruler. The fragmented image is subjected to grayscale conversion, filtering, automatic threshold binarization segmentation, and local centroid segmentation to divide adjacent or slightly contiguous fragmented image regions into multiple independent fragmented regions. Based on the independent fragment regions, the major axis dimension, minor axis dimension, major axis-to-minor axis ratio distribution, edge angle, fragment particle size distribution, fragment fractal dimension, characteristic particle size, mean major axis-to-minor axis ratio, and standard deviation of major axis-to-minor axis ratio are extracted as the overall fragmentation characteristic parameters.

6. The method for identifying rock degradation modes considering energy-time effects according to claim 1, characterized in that, The process of establishing the cross-scale correlation model includes: Multiple sets of energy-time characteristics and multi-scale degradation characteristics under different impact conditions were collected as training data. Establish a microscopic damage mapping model between energy time characteristics and porosity change, pore connectivity increment, pore volume distribution skewness coefficient and average coordination number. Establish a macroscopic fracture mapping model between energy time characteristics and roughness multiscale fractal dimension and mean anisotropy index; Establish an overall fragmentation mapping model between energy-time characteristics and fragment fractal dimension, characteristic particle size, mean of major axis to minor axis ratio, and standard deviation of major axis to minor axis ratio; Furthermore, cross-scale correlations were established regarding the transmission of microscopic damage features to macroscopic fracture features, and the transmission of macroscopic fracture features to overall fragmentation features.

7. The method for identifying rock degradation modes considering energy-time effects according to claim 1, characterized in that, The process of establishing the degradation mode criterion includes: Using the degradation mode obtained from experimental observation as the discrimination criterion, the contribution weight analysis of each parameter in the energy-time characteristics is performed to determine the core discrimination parameters and auxiliary discrimination parameters; Threshold conditions for each discriminant parameter were determined by receiver operating characteristic (ROC) curve analysis. When the peak time is late, the average rate of energy injection before the peak is low, and the time-frequency concentration is low, it is identified as damage-dominated. When the peak time, the average rate of energy injection before the peak and the time-frequency concentration are all in the middle range, and the difference between the dominant frequencies before and after the peak is within the preset tolerance range, it is judged to be a fracture-dominated type. When the peak time is early, the average rate of energy injection before the peak is high, the time-frequency concentration is high, and the energy before and after the peak is relatively high, it is identified as a breakup-dominated type.

8. A rock degradation mode discrimination system considering energy time effects, used to implement the rock degradation mode discrimination method considering energy time effects as described in any one of claims 1-7, characterized in that, include: The multi-scale information acquisition subsystem is used to acquire microscopic pore data, fracture surface point cloud data, and fragment morphology data of rock samples after impact, and output multi-scale degradation characteristics. The energy-time density feature extraction subsystem is used to receive the energy-time density spectrum and the dissipated energy time history, perform stage division of the energy action process and energy-time feature extraction, and output the energy-time features. The cross-scale correlation modeling subsystem is used to receive multiple sets of energy-time characteristics and multi-scale degradation characteristics under different impact conditions, and to establish the mapping relationship between energy-time characteristics and multi-scale degradation characteristics. The degradation mode criterion generation subsystem is used to determine the criterion threshold based on the analysis results of the cross-scale correlation modeling subsystem, generate three types of degradation mode criteria: damage-dominated, fracture-dominated, and fragmentation-dominated, and output the rock degradation mode type and confidence score.

9. A rock degradation mode discrimination system considering energy-time effects according to claim 8, characterized in that, The multi-scale information acquisition subsystem includes: The nuclear magnetic resonance detection unit includes a nuclear magnetic resonance analyzer main unit, a sample probe, a pulse sequence controller, a T2 spectrum inversion module, a three-dimensional voxelization inversion module with built-in simulated annealing algorithm, a pore network model construction module with built-in maximum sphere algorithm, and a micro-parameter calculation module, which are used to output porosity change, pore connectivity increment, pore volume distribution skewness coefficient, and average coordination number. The three-dimensional cross-section scanning unit includes a three-dimensional laser cross-section scanner, a point cloud preprocessing module, a multi-scale decomposition module with a built-in two-dimensional variational mode decomposition algorithm, a modified Grasselli roughness calculation module, and a macroscopic parameter calculation module; the macroscopic parameter calculation module outputs the roughness multi-scale fractal dimension and the mean anisotropy index. The fragment image acquisition and processing unit includes a high-resolution industrial camera, a calibration ruler, an image preprocessing module, a local centroid segmentation module, a geometric parameter extraction module based on the second-order central moment, a maximum likelihood estimation module, a fractal dimension calculation module, a feature particle size calculation module, and an overall fragmentation parameter output module; the overall fragmentation parameter output module outputs the fragment fractal dimension, feature particle size, mean of major axis to minor axis ratio, and standard deviation of major axis to minor axis ratio.

10. A rock degradation mode discrimination system considering energy-time effects according to claim 8, characterized in that, The energy-time density feature extraction subsystem includes an energy-time density spectrum input interface, a peak moment identification module, a stage division module, a pre-peak feature extraction module, a post-peak feature extraction module, a comprehensive feature calculation module, and an energy-time feature output module. The cross-scale correlation modeling subsystem includes a training data management module, a microscopic damage mapping modeling module, a macroscopic fracture mapping modeling module, a global fragmentation mapping modeling module, a cross-scale correlation establishment module, and a model reliability evaluation module. The degradation mode criterion generation subsystem includes a feature weight analysis module, a threshold determination module, a damage-dominated criterion module, a fracture-dominated criterion module, a breakage-dominated criterion module, and a comprehensive discrimination module.

Images