Method for testing mechanical properties of rocks in low-temperature freeze-thaw environment based on microstructure evolution

Abstract

The invention discloses a method for testing mechanical properties of rocks in a low-temperature freeze-thaw environment based on microstructure evolution, and relates to the technical field of rock mechanics and freeze-thaw tests.The method comprises the steps that a rock internal structure image sequence is obtained through controllable freeze-thaw cycles, and a digital model is established through three-dimensional reconstruction. And sequentially extracting pore structure and mineral component information, and respectively carrying out evolution trajectory tracking and migration quantification processing to generate a pore structure evolution graph and a component migration evolution graph. And fusing the two types of maps into a structure evolution feature vector for driving mechanical loading and synchronously acquiring mechanical response signals, and finally constructing a correlation mapping table of mechanical properties and microscopic evolution. According to the method, dynamic and quantitative synchronous tracking of two microcosmic evolution behaviors of rock pore expansion and mineral migration in the freezing and thawing process is realized, and a real-time corresponding relation between the microcosmic evolution behaviors and macroscopic mechanical property change is established.

Description

Technical Field

[0001] This invention belongs to the field of rock mechanics and freeze-thaw testing technology, specifically a method for testing the mechanical properties of rocks in low-temperature freeze-thaw environments based on microstructure evolution. Background Technology

[0002] In cold-region engineering and underground space development, the degradation of the mechanical properties of rocks after freeze-thaw cycles is a critical issue. Current testing methods mainly rely on macroscopic mechanical tests, which measure macroscopic parameters such as uniaxial compressive strength and elastic modulus of the rock after a set freeze-thaw cycle to assess damage. These methods can only obtain the final mechanical state after freeze-thaw action, and their testing process is completely disconnected from the real-time changes in the internal microstructure of the rock.

[0003] The shortcomings of existing technologies lie in the lack of dynamic monitoring and quantitative description of microscopic evolution processes. Although microscopic observation or CT scanning techniques have been applied, they are usually only used to compare static structural images before and after freeze-thaw cycles, and cannot fully track the dynamic evolution trajectory of pores and microcracks during the cycle, such as the entire process of initiation, expansion, and convergence. Existing methods are also insufficient to quantify the migration behavior of mineral components within rocks driven by freeze-thaw cycles. Water freezing and expansion, and thawing and seepage can trigger cement dissolution, particle boundary slippage, and secondary mineral precipitation, but there is a lack of effective dynamic quantitative characterization methods for the spatial redistribution processes and rates of these microscopic components.

[0004] A testing method is needed that can simultaneously track the dynamic evolution of the internal pore structure of rocks and quantitatively analyze the migration of mineral components during freeze-thaw cycles, and directly correlate these two microscopic dynamic evolution information with real-time macroscopic mechanical responses to reveal the essential connection between the microscopic mechanism and macroscopic manifestation of freeze-thaw damage. Summary of the Invention

[0005] This invention aims to solve at least one of the technical problems existing in the prior art; Therefore, this invention proposes a method for testing the mechanical properties of rocks in low-temperature freeze-thaw environments based on microstructure evolution, including: Rock samples were prepared and placed in a controlled low-temperature freeze-thaw simulation environment. Periodic freeze-thaw cycle simulations were initiated, and the internal structure image sequence of the rock samples was obtained after processing. Perform three-dimensional reconstruction processing on the internal structure image sequence of the rock sample to generate a digital three-dimensional structural model of the rock sample; Based on the digital three-dimensional structural model, the set of geometric parameters of the pore structure and the distribution map of mineral components inside the rock sample are extracted; Evolution trajectory tracking processing is performed on the set of geometric parameters of the pore structure to generate an evolution map of the pore structure; Perform component migration quantization processing on the mineral component distribution map to generate a component migration evolution map; The pore structure evolution map and the component migration evolution map are input into the mechanical correlation model to generate a structural evolution feature vector; Based on the structural evolution feature vector, the driving mechanical loading device applies a load of a preset mode to the rock sample, and simultaneously collects the mechanical response signal sequence of the rock sample; Based on the mechanical response signal sequence of the rock samples and the corresponding structural evolution feature vector, a rock mechanical property evolution correlation mapping table is constructed.

[0006] Furthermore, the rock sample is prepared and placed in a controlled low-temperature freeze-thaw simulation environment. Periodic freeze-thaw cycle simulations are initiated, and after processing, an image sequence of the internal structure of the rock sample is obtained, including: Prepare rock samples that meet the preset geometric specifications and initial state; The rock sample was placed in a controlled low-temperature freeze-thaw simulation environment, and a periodic freeze-thaw cycle simulation was initiated. At each preset freeze-thaw cycle node, the freeze-thaw simulation process is interrupted, and non-destructive scanning imaging is performed on the rock sample to obtain an image sequence of the internal structure of the rock sample, specifically including: Control the temperature of the low-temperature freeze-thaw simulation environment to return to the preset reference temperature; Rock samples were taken from a low-temperature freeze-thaw simulation environment and placed on the sample stage of a non-destructive scanning device; The X-ray source and detector array of the non-destructive scanning device are activated and rotated to scan the rock sample. Acquired X-ray projection images of rock samples at different rotation angles; Perform fault reconstruction operations on all penetrating ray projection images to generate a set of internal structural fault slice images of rock samples at different preset freeze-thaw cycle nodes; The set of internal structure tomographic images is arranged axially to form an internal structure image sequence of the rock sample.

[0007] Further, a three-dimensional reconstruction process is performed on the internal structure image sequence of the rock sample to generate a digital three-dimensional structural model of the rock sample, including: Gray-level normalization and noise filtering were performed on each fault slice image in the internal structure image sequence of the rock sample. Pixel alignment and interpolation calculations are performed on the processed tomographic slice images to construct the initial three-dimensional voxel space data; In three-dimensional voxel space data, a preset threshold segmentation algorithm is applied to distinguish the spatial regions of rock matrix, pore space and different mineral phases; Based on the distinguished rock matrix, pore space, and spatial regions of different mineral phases, a digital three-dimensional structural model characterizing the internal material distribution of the rock sample is constructed.

[0008] Furthermore, based on the aforementioned digital three-dimensional structural model, the set of geometric parameters of the pore structure and the distribution map of mineral components inside the rock sample are extracted, including: In the pore space represented by the digital three-dimensional structural model, all connected pore networks are identified; For each connected pore network, calculate its volume, surface area, equivalent diameter, shape factor, tortuosity, and connectivity parameters, and combine them to form a set of geometric parameters for the pore structure. In the different mineral phase spatial regions represented by the digital three-dimensional structural model, the spatial distribution density and adjacency relationship of each mineral phase are statistically analyzed based on the three-dimensional spatial coordinates of each mineral phase. The spatial distribution density and adjacency relationship of each mineral phase are encoded in the form of a map to generate a mineral component distribution map.

[0009] Further, evolution trajectory tracing processing is performed on the set of geometric parameters of the pore structure to generate a pore structure evolution map, including: Extract the changes in geometric parameters of pores at the same spatial location during the freeze-thaw cycle from the set of geometric parameters of the pore structure corresponding to different freeze-thaw cycle nodes; For each pore, calculate its volume change rate, shape evolution vector, and connectivity transition path between adjacent freeze-thaw cycle nodes; By integrating the changes in geometric parameters, volume change rate, shape evolution vector, and connectivity transition path of all pores, a dynamic relationship diagram of the evolution of pore structure with the increase of freeze-thaw cycles is drawn, namely, the pore structure evolution diagram.

[0010] Further, the mineral component distribution map is subjected to component migration quantification processing to generate a component migration evolution map, including: By comparing the mineral composition distribution maps corresponding to different freeze-thaw cycle nodes, the spatial displacement of specific mineral particles can be identified. Calculate the displacement vector of the spatial displacement of the specific mineral particle, and statistically analyze the distribution characteristics of the displacement vectors of mineral particles of the same type. On the mineral component distribution map, mineral particles with significant displacement and their displacement trajectories are marked. By superimposing and integrating the mineral particle displacements and evolution trajectories observed in all freeze-thaw cycle nodes, a map reflecting the migration patterns and evolution trends of mineral components under freeze-thaw action is generated, namely, the component migration and evolution map.

[0011] Furthermore, the pore structure evolution map and the component migration evolution map are input into the mechanical correlation model to generate a structural evolution feature vector, including: Extract the pore volume change rate sequence, pore morphology evolution rate sequence, and pore network connectivity change sequence from the pore structure evolution map; Statistical features of mineral displacement, mineral boundary change sequences, and microcrack propagation trajectories were extracted from the component migration and evolution map. The pore volume change rate sequence, pore morphology evolution rate sequence, pore network connectivity change sequence, mineral displacement statistical characteristics, mineral boundary change sequence, and microcrack propagation trajectory description are normalized and vectorized and concatenated to form the original feature set. The original feature set is input into the feature dimensionality reduction and fusion layer of the pre-trained mechanical correlation model, which compresses and fuses the high-dimensional original feature set into a low-dimensional structural evolution feature vector that is highly correlated with the mechanical response.

[0012] Furthermore, based on the structural evolution feature vector, a mechanical loading device is driven to apply a load of a preset pattern to the rock sample, and the mechanical response signal sequence of the rock sample is acquired simultaneously, including: The structural evolution feature vector generated by the current freeze-thaw cycle node is analyzed. The structural evolution feature vector includes information on pore structure stability assessment and mineral component bonding strength assessment. Based on the pore structure stability assessment and mineral component bonding strength assessment information, a matching load application mode is selected from a preset loading mode library. The selected load application mode is converted into control commands for the mechanical loading device, which then applies a combination of axial stress, confining pressure, and shear force to the rock sample corresponding to the current structural state. During the application of load, stress, strain, acoustic emission signals and ultrasonic speed signals are simultaneously collected by a sensor array arranged on the rock sample, forming a mechanical response signal sequence of the rock sample.

[0013] Furthermore, based on the mechanical response signal sequence of the rock sample and the corresponding structural evolution feature vector, a rock mechanical property evolution correlation mapping table is constructed, including: The mechanical response signal sequence of the rock sample obtained at each freeze-thaw cycle node is processed to extract mechanical parameters, including peak strength, elastic modulus, Poisson's ratio and stress-strain curve characteristic parameters. The mechanical parameters extracted from each freeze-thaw cycle node are paired with the structural evolution feature vector corresponding to the freeze-thaw cycle node to form a "mechanical parameter-structural evolution feature vector" data pair. Arrange the "mechanical parameter-structural evolution feature vector" data pairs of all freeze-thaw cycle nodes in ascending order of freeze-thaw cycle number; Based on the arranged data pairs, an association mapping structure is constructed. The association mapping structure uses the number of freeze-thaw cycles as an index to record the correspondence between mechanical parameters and structural evolution feature vectors, forming a rock mechanical property evolution association mapping table.

[0014] Furthermore, an association mapping structure is constructed, which uses the number of freeze-thaw cycles as an index to record the correspondence between mechanical parameters and structural evolution feature vectors, including: Create a multidimensional data table, using the number of freeze-thaw cycles as the main index column of the multidimensional data table; In the multidimensional data table, for each freeze-thaw cycle number index row, establish a mechanical parameter column group and a structural evolution feature vector column group; Fill the extracted peak strength, elastic modulus, Poisson's ratio, and stress-strain curve characteristic parameters into the corresponding sub-columns of the mechanical parameter column group in the corresponding row; Fill each dimension component of the structural evolution feature vector into the corresponding sub-column of the structural evolution feature vector column group of the corresponding row; The mapping relationship between the corresponding set of mechanical parameters and the structural evolution feature vector can be directly obtained by querying the main index column of the multidimensional data table based on the number of freeze-thaw cycles.

[0015] Compared with the prior art, the beneficial effects of the present invention are: By performing evolution trajectory tracking processing of pore structure geometric parameters, and based on time-series images of the internal structure, image registration and feature tracking algorithms are applied to continuously capture and quantify the periodic changes in size, shape, and spatial location of individual or group pores. This generates a pore structure evolution map, directly revealing the continuous dynamic process of freeze-thaw damage from micropore development to macroscopic crack formation. It achieves intuitive visualization and quantitative description of damage accumulation paths and patterns, surpassing the limitations of static analysis that only compares initial and final states.

[0016] By performing component migration quantification on the mineral component distribution map, digital image processing technology was used to identify and segment mineral phases. Pixel or voxel changes in specific mineral regions were tracked through sequential image registration, and their displacement vectors and abundance change rates were calculated. This generated a component migration evolution map, accurately depicting the evolution of dissolution, migration, and redeposition of different mineral particles or cements in three-dimensional space during freeze-thaw cycles. A direct spatial correlation was established between the dynamic redistribution of micro-components and the degradation of macroscopic material properties.

[0017] The structural evolution feature vector generated by fusing the two types of dynamic evolution maps is used as the input parameter for the driving mechanical loading device. This vector can reflect the current damage state and component distribution characteristics inside the sample in real time, allowing the applied loading mode to be spatially correlated with internal structural weaknesses. The synchronously acquired mechanical response signal therefore contains intrinsic mechanical information under specific microstructural states, and the constructed correlation mapping table is essentially a quantitative database of the relationship between the dynamic evolution state of rock microstructure and its instantaneous mechanical properties. Attached Figure Description

[0018] Figure 1 This is a flowchart illustrating the steps of the rock mechanical property testing method based on microstructure evolution in a low-temperature freeze-thaw environment according to the present invention. Figure 2 A flowchart for generating a digital three-dimensional structural model. Detailed Implementation

[0019] The technical solution of the present invention will be clearly and completely described below with reference to the embodiments. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0020] See Figure 1Rock samples meeting experimental requirements were prepared and placed in a controlled low-temperature freeze-thaw simulation environment. A periodic freeze-thaw cycle simulation program with pre-defined temperature range, cycle period, and rate was initiated. After undergoing the pre-defined freeze-thaw cycle, internal structural image sequences of the rock samples at multiple key nodes were acquired using non-destructive imaging technology. Three-dimensional reconstruction processing was performed on the acquired internal structural image sequences of the rock samples to generate a digital three-dimensional structural model that accurately reflects the spatial distribution of materials within the sample. Based on the generated digital three-dimensional structural model, the set of pore structure geometric parameters and the mineral component distribution map of the rock samples were extracted. Evolution trajectory tracking processing across freeze-thaw cycle nodes was performed on the extracted pore structure geometric parameter set to generate a pore structure evolution map revealing the dynamic changes of the pore system. Quantitative analysis of component migration was performed on the mineral component distribution map to generate a component migration evolution map showing the migration and redistribution characteristics of mineral particles. The pore structure evolution map and the component migration evolution map were input into a pre-constructed mechanical correlation model, which outputs a structural evolution feature vector that comprehensively characterizes the current microstructural state. Based on this structural evolution feature vector, the system analyzes and drives a mechanical loading device to apply a load of a preset mode that matches the current state of the rock sample. During this process, the mechanical response signal sequence generated by the rock sample is acquired simultaneously. Finally, based on the acquired mechanical response signal sequence of the rock sample and its corresponding structural evolution feature vector, a systematic rock mechanical property evolution correlation mapping table is constructed. This table establishes a quantitative correlation between freeze-thaw cycles, microstructural evolution, and macroscopic mechanical properties.

[0021] See Figure 2 In one embodiment of the invention, an example scenario using saturated water-absorbing granite as a sample is illustrated. The sample is processed into a standard cylinder with a diameter of 50 mm and a height of 100 mm. In a specific implementation, the rock sample, meeting the preset geometric specifications and initial saturation state, is placed in a high-precision environmental chamber capable of achieving precise temperature cycling from -20°C to +20°C to simulate periodic freeze-thaw cycles. The periodic freeze-thaw cycle simulation program is started, with a single complete freeze-thaw cycle set to 12 hours, including maintaining a low temperature of -20°C for 4 hours and a high temperature of +20°C for 4 hours, with both heating and cooling rates at 10°C per hour. In some embodiments, the preset freeze-thaw cycle nodes are 0, 10, 25, 50, and 100 cycles. When the freeze-thaw simulation process reaches these specified nodes, the system automatically interrupts the freeze-thaw simulation process.

[0022] In practice, the first step after interrupting the freeze-thaw simulation is to control the temperature of the cryogenic freeze-thaw simulation environment, allowing it to slowly return to a preset reference temperature +20°C at a controlled rate and maintain this temperature for at least 2 hours. The aim is to minimize the instantaneous impact of any additional thermal stress caused by sudden temperature changes on the microstructure of the rock sample. In practice, the rock sample is removed from the cryogenic freeze-thaw simulation environment, transferred, and precisely placed on the sample stage of a microfocus X-ray computed tomography (XCT) nondestructive testing (NDT) device. Laser positioning and mechanical leveling ensure that the geometric center of the rock sample coincides with the rotation axis of the NDT device. The X-ray source and detector array of the NDT device are then activated, with the X-ray source voltage set to 160 kV and the current to 100 μA. The sample stage is then rotated 360 degrees at an angular resolution of 0.5 degrees per step, and transmitted X-ray projection images of the rock sample at different rotation angles are acquired.

[0023] Optionally, after all transmitted X-ray projection images have been acquired, dark-field and bright-field corrections are performed on the projection images to eliminate detector noise and inhomogeneities. Then, a fault reconstruction operation based on a filtered back-projection algorithm is performed on all corrected transmitted X-ray projection images to generate a series of parallel internal structure fault slice images of the rock sample at corresponding preset freeze-thaw cycle nodes. The slice thickness is 20 micrometers, and these slices together constitute an internal structure fault slice image set. In some embodiments, to ensure that the image quality meets the requirements of quantitative analysis, an image quality assessment parameter can be introduced during the reconstruction process, such as calculating the local signal-to-noise ratio of the reconstructed slices to guide the optimization of the reconstruction algorithm parameters.

[0024] In practice, all internal structure fault slice images generated at the same freeze-thaw cycle node are precisely arranged and numbered according to their spatial position along the sample axis. This means that for different freeze-thaw cycle nodes, the entire process from temperature regression to fault reconstruction is repeated, ultimately resulting in a set of multiple sets of internal structure fault slice images indexed by the number of freeze-thaw cycles. Optionally, these time-series-arranged sets of internal structure fault slice images are managed holistically, forming a sequence of internal structure images of the rock sample evolving over time during a complete freeze-thaw experiment. This sequence visually records the dynamic changes in the internal porosity and mineral structure of the rock from 0 to 100 freeze-thaw cycles. Data comparison shows that by comparing fault slices from 0 and 50 cycles, it can be observed that the initially uniformly distributed micropores have partially expanded and interconnected after 50 freeze-thaw cycles, forming a distinct microcrack network.

[0025] In one embodiment of the present invention, each tomographic slice image in the internal structure image sequence of the rock sample is preprocessed. The preprocessing includes grayscale normalization and noise filtering. Grayscale normalization linearly transforms the pixel grayscale values ​​of all slices to a standard range of 0-255 to eliminate differences in scanning conditions. Noise filtering uses a median filter with a 3x3 pixel window to remove salt-and-pepper noise from the image. In a specific implementation, a set of tomographic slice images from the same freeze-thaw cycle node is pixel-level aligned to correct for possible minor displacements of the sample during scanning. Then, interpolation calculation is performed in the direction perpendicular to the slice plane (i.e., the sample axis). The interpolation calculation uses a trilinear interpolation algorithm to construct an isotropic initial three-dimensional voxel space data with a voxel size of 20 μm x 20 μm x 20 μm from the originally discrete slice data.

[0026] In some embodiments, a preset threshold segmentation algorithm is applied to distinguish different material phases in three-dimensional voxel space data. The threshold segmentation algorithm automatically calculates and determines the segmentation threshold using the maximum inter-class variance method. This algorithm traverses the grayscale histogram to find the optimal threshold that maximizes the inter-class variance between the foreground and background. It can be understood that through the threshold segmentation algorithm, each voxel in the three-dimensional voxel space data is classified and labeled as a spatial region of rock matrix, pore space, or a specific mineral phase. In a specific implementation, based on the distinguished spatial regions of rock matrix, pore space, and different mineral phases, a three-dimensional visualization software library is used to perform surface rendering or directly construct a digital three-dimensional structural model that can accurately characterize the spatial distribution and geometric morphology of the internal material of the rock sample by using a surface rendering or volume drawing method on a set of voxels with the same label. Data comparison shows that the digital three-dimensional structural model with 0 iterations displays isolated pore distribution, while the model with 50 iterations clearly shows a network structure formed by interconnected pores. The total number of pore phases in the model increases by approximately 110%, which provides a direct three-dimensional data basis for quantitative parameter extraction.

[0027] In practical implementation, parameters are extracted based on a digital 3D structural model. Within the pore space represented by the digital 3D structural model, a 3D 26-neighborhood connectivity analysis algorithm is used to identify all connected pore networks, and each connected pore network is assigned an independent label. Optionally, the geometric parameters of each connected pore network are calculated. The volume of the pore network is obtained by counting the total number of voxels it contains and multiplying by the volume of a single voxel. The surface area of ​​the pore network is obtained by detecting the voxel surfaces at the interface between the pores and the rock matrix and calculating the total area. The equivalent diameter of the pore network is defined as the diameter of a sphere with the same volume as the pore network. The shape factor of the pore network is used to describe the degree of deviation of the pore network shape from the sphere. The tortuosity of the pore network is obtained by calculating the ratio of the length of the most probable fluid path within the pore network to the straight-line distance between its two ends. The connectivity parameters of the pore network are calculated by analyzing the number of nodes and branches in the pore network skeleton. All these parameters are combined to form a set of geometric parameters describing the pore structure of the sample at a specific freeze-thaw stage.

[0028] In some embodiments, within different mineral phase spatial regions represented by the digital three-dimensional structural model, the spatial distribution density and adjacency relationships of each mineral phase are statistically analyzed based on their three-dimensional spatial coordinates. The spatial distribution density is calculated by dividing the digital three-dimensional structural model into multiple cubic sub-regions and calculating the volume percentage of a specific mineral phase voxel within each sub-region. The adjacency relationship is analyzed by examining the proportion of contact interface areas between different mineral phases or between a mineral phase and pores. In specific implementations, the spatial distribution density of each mineral phase is encoded in the form of a two-dimensional contour map or a three-dimensional cloud map, and the main adjacency relationships are marked on the map with boundaries of different colors, generating a comprehensive mineral component distribution map. Data comparison shows that after 100 freeze-thaw cycles, the spatial distribution density map of mica minerals exhibits locally high-density areas, while the proportion of the contact boundary between quartz and feldspar on the map increases from the initial 15% to 22%, indicating relative displacement or boundary deterioration between mineral particles.

[0029] In one embodiment of the present invention, evolution trajectory tracking processing is performed on the set of pore structure geometric parameters. From the set of pore structure geometric parameters corresponding to different freeze-thaw cycle nodes (e.g., 0, 10, 25, 50, 100 cycles), the geometric parameter changes of pores at the same spatial location during the complete freeze-thaw cycle are extracted by matching 3D image coordinates with pore identifiers. In some embodiments, for each successfully tracked pore, its volume change rate between adjacent freeze-thaw cycle nodes is calculated. The volume change rate is calculated as (current node pore volume - previous node pore volume) / previous node pore volume. Optionally, a shape evolution vector is calculated for each pore. The shape evolution vector is used to quantify the direction and extent of pore morphology evolution in three-dimensional space. The connectivity transition path records the process of a pore changing from an isolated state to a connected state between adjacent nodes, or from belonging to one connected cluster to another connected cluster.

[0030] In practical implementation, the changes in geometric parameters, volume change rate, shape evolution vector, and connectivity transition paths of all tracked pores are integrated into a unified 3D visualization environment with the number of freeze-thaw cycles as the time axis. This creates a dynamic relationship diagram of pore structure evolution as the number of freeze-thaw cycles increases, i.e., a pore structure evolution map. Data comparison shows that the pore structure evolution map clearly displays that from 0 to 50 cycles, the volume change rate of pores in the central region of the sample is generally higher than that in the edge region. In practical implementation, component migration quantification is performed on the mineral component distribution map. By comparing the mineral component distribution maps corresponding to different freeze-thaw cycle nodes, the displacement of specific mineral particles in 3D space is identified through feature point matching or digital image correlation algorithms. The displacement vector of the identified specific mineral particles is calculated, and the displacement vector contains both magnitude and direction information. The magnitude and direction distribution of displacement vectors of the same type of mineral particles (such as all tracked feldspar particles) are statistically analyzed to obtain the statistical distribution characteristics of the displacement vectors, such as the average displacement and displacement direction rose diagram. In some embodiments, on the mineral component distribution map, mineral particles whose displacement exceeds a preset threshold and their displacement trajectory from the previous node to the current node are marked with arrowed line segments or highlighted trajectory lines.

[0031] Optionally, the displacement and evolution trajectories of mineral particles observed at all freeze-thaw cycle nodes (0, 10, 25, 50, and 100 cycles) are superimposed and integrated into a three-dimensional or two-dimensional projection map in chronological order, generating a map reflecting the migration patterns and spatial evolution trends of mineral components under freeze-thaw cycles, i.e., a component migration evolution map. It can be understood that the component migration evolution map not only shows the final position of the particles but also characterizes the overall migration pattern through the density and direction of the trajectory lines. Data comparison shows that after 100 freeze-thaw cycles, more than 60% of the tracked mica particles underwent detectable displacement, with an average displacement of 12.5 micrometers. The displacement direction showed a clear dominant orientation perpendicular to the bedding plane, while during cycles 0 to 25, the displacement of mineral particles was very small and the direction was chaotic.

[0032] In one embodiment of the present invention, pore volume change rate sequence, pore morphology evolution rate sequence, and pore network connectivity change sequence are extracted from the pore structure evolution map. The pore volume change rate sequence is an array recording the percentage change in volume of each tracked pore between adjacent freeze-thaw cycle nodes in chronological order. The pore morphology evolution rate sequence is an array recording the change in the magnitude of the shape evolution vector of each pore over time. The pore network connectivity change sequence is an array recording the change in the average coordination number or the number of connected clusters of the pore network with the number of freeze-thaw cycles. Mineral displacement statistical features, mineral boundary change sequences, and microcrack propagation trajectory descriptions are extracted from the component migration evolution map. The mineral displacement statistical features include the average value, standard deviation, and principal direction of the displacement vectors of mineral particles of the same type. The mineral boundary change sequence is an array recording the change in the ratio of the contact interface area between different mineral phases or between minerals and pores with the number of freeze-thaw cycles. The microcrack propagation trajectory description is characterized by quantifying the changes in microcrack length, width, and number of branches. In practice, the pore volume change rate sequence, pore morphology evolution rate sequence, pore network connectivity change sequence, mineral displacement statistical characteristics, mineral boundary change sequence, and microcrack propagation trajectory description are normalized and vectorized. Normalization uses a min-max scaling method to scale each feature to the [0,1] interval. Vectorization concatenates all normalized features in a fixed order to form a high-dimensional vector, creating the original feature set. Data comparison shows that for the same sample, after 25 and 50 cycles, the mean of the pore volume change rate sequence in the original feature set increases from 0.02 to 0.08, and the average displacement in the mineral displacement statistical characteristics increases from 5 micrometers to 15 micrometers, indicating that microstructural damage intensifies with freeze-thaw cycles.

[0033] In some embodiments, referring to Table 1, a partial example of the original feature set, shows some feature values ​​extracted at a specific freeze-thaw cycle node.

[0034] Table 1: Examples of partial features in the original feature set after 50 freeze-thaw cycles

[0035] In practice, the original feature set is input into the feature reduction and fusion layer of a pre-trained mechanical correlation model. This model is a neural network model pre-trained using a large amount of rock sample data. The feature reduction and fusion layer compresses the high-dimensional original feature set into a low-dimensional structural evolution feature vector through linear transformation. This step can be understood as fusing the original feature set of tens or hundreds of dimensions into a low-dimensional vector with more explicit physical meaning, such as a 5-dimensional structural evolution feature vector, whose components may comprehensively represent the degree of porosity development, the degree of weakening of mineral bonding, etc.

[0036] In some embodiments, the structural evolution feature vector generated by the current freeze-thaw cycle node is parsed. This feature vector includes information on pore structure stability assessment and mineral component bonding strength assessment. The pore structure stability assessment is reflected by the value of a specific component of the vector; a lower value indicates a less stable pore structure. The mineral component bonding strength assessment is reflected by the value of another component of the vector; a lower value indicates weaker intermineral bonding. Based on the pore structure stability assessment and mineral component bonding strength assessment information, a matching load application mode is selected from a preset loading mode library. This library stores various load application modes, each defining parameters such as axial stress loading rate, confining pressure magnitude, and shear force introduction method. For example, when the structural evolution feature vector shows a low pore structure stability assessment value and a medium mineral component bonding strength assessment value, the matching load application mode in the loading mode library might be "low confining pressure constant displacement rate compression." Optionally, the selected load application mode is converted into control instructions for a mechanical loading device. These instructions include specific parameters such as target stress value, loading rate, and duration, driving the mechanical loading device to apply a combination of axial stress, confining pressure, and shear force corresponding to the current structural state to the rock sample. During the loading process, stress, strain, acoustic emission signals, and ultrasonic velocity signals are simultaneously acquired by a sensor array deployed on the rock sample. The sensor array includes strain gauges attached to the sample surface, force sensors mounted on a pressure plate, acoustic emission probes coupled to the sides of the sample, and ultrasonic emission receivers. These signals are synchronously recorded and time-aligned to form a sequence of mechanical response signals of the rock sample under that loading condition. Data comparisons show that for samples with lower pore structure stability assessment values, the acoustic emission signal sequence acquired during loading indicates that dense acoustic emission events begin to appear at lower stress levels. Conversely, for samples with higher pore structure stability assessment values, acoustic emission events are mainly concentrated near the peak stress stage.

[0037] In one embodiment of the present invention, continuing with the example scenario of saturated water-absorbing granite as the sample, and based on the mechanical response signal sequence of the rock sample synchronously collected at each freeze-thaw cycle node and the generated structural evolution feature vector, the mechanical response signal sequence of the rock sample acquired at each freeze-thaw cycle node is processed to extract mechanical parameters. The processing includes filtering and calibrating the synchronously collected stress and strain signals, plotting a complete stress-strain curve, directly reading the peak stress as the peak strength from the stress-strain curve, calculating the slope of the linear elastic segment as the elastic modulus, and calculating Poisson's ratio through the ratio of transverse strain to axial strain. The characteristic parameters of the stress-strain curve are extracted by mathematical analysis of the curve, for example, by fitting the nonlinear part of the rising segment of the stress-strain curve to quantify its hardening or softening characteristics. A parameter used to describe the shape of the curve can be expressed by the formula:

[0038] in: Indicates the morphological parameters of the stress-strain curve. This represents the strain corresponding to the change from zero strain to peak stress. The area under the stress-strain curve within the interval, Indicates peak intensity. The axial strain represents the peak strength. Data comparison shows that after 50 freeze-thaw cycles, the peak strength of the sample was 90 MPa and the elastic modulus was 25 GPa, while after 100 cycles, the peak strength decreased to 70 MPa and the elastic modulus decreased to 18 GPa. Simultaneously, the curve morphology parameters... It also decreased from 0.65 to 0.58.

[0039] In some embodiments, the mechanical parameters extracted from each freeze-thaw cycle node, including peak strength, elastic modulus, Poisson's ratio, and stress-strain curve characteristic parameters, are paired with the corresponding structural evolution characteristic vector of that freeze-thaw cycle node to form a "mechanical parameter-structural evolution characteristic vector" data pair. All "mechanical parameter-structural evolution characteristic vector" data pairs for all freeze-thaw cycle nodes are arranged in ascending order of freeze-thaw cycle count. Based on these arranged data pairs, an association mapping structure is constructed. This association mapping structure uses the freeze-thaw cycle count as an index and records the correspondence between mechanical parameters and structural evolution characteristic vectors in tabular form, forming a rock mechanical property evolution association mapping table. Optionally, a multidimensional data table is established as a specific implementation of the association mapping structure. The freeze-thaw cycle count is used as the main index column of the multidimensional data table. In the multidimensional data table, for each freeze-thaw cycle count index row, a mechanical parameter column group and a structural evolution characteristic vector column group are established.

[0040] The above embodiments are only used to illustrate the technical methods of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical methods of the present invention without departing from the spirit and scope of the technical methods of the present invention.

Claims

1. A method for testing the mechanical properties of rocks in low-temperature freeze-thaw environments based on microstructure evolution, characterized in that, The method includes: Rock samples were prepared and placed in a controlled low-temperature freeze-thaw simulation environment. Periodic freeze-thaw cycle simulations were initiated, and the internal structure image sequence of the rock samples was obtained after processing. Perform three-dimensional reconstruction processing on the internal structure image sequence of the rock sample to generate a digital three-dimensional structural model of the rock sample; Based on the digital three-dimensional structural model, the set of geometric parameters of the pore structure and the distribution map of mineral components inside the rock sample are extracted; Evolution trajectory tracking processing is performed on the set of geometric parameters of the pore structure to generate an evolution map of the pore structure; Perform component migration quantization processing on the mineral component distribution map to generate a component migration evolution map; The pore structure evolution map and the component migration evolution map are input into the mechanical correlation model to generate a structural evolution feature vector; Based on the structural evolution feature vector, the driving mechanical loading device applies a load of a preset mode to the rock sample, and simultaneously collects the mechanical response signal sequence of the rock sample; Based on the mechanical response signal sequence of the rock samples and the corresponding structural evolution feature vector, a rock mechanical property evolution correlation mapping table is constructed.

2. The method for testing the mechanical properties of rocks in low-temperature freeze-thaw environments based on microstructure evolution according to claim 1, characterized in that, The prepared rock sample was placed in a controlled low-temperature freeze-thaw simulation environment, and a periodic freeze-thaw cycle simulation was initiated. After processing, an image sequence of the internal structure of the rock sample was obtained, including: Prepare rock samples that meet the preset geometric specifications and initial state; The rock sample was placed in a controlled low-temperature freeze-thaw simulation environment, and a periodic freeze-thaw cycle simulation was initiated. At each preset freeze-thaw cycle node, the freeze-thaw simulation process is interrupted, and non-destructive scanning imaging is performed on the rock sample to obtain an image sequence of the internal structure of the rock sample, specifically including: Control the temperature of the low-temperature freeze-thaw simulation environment to return to the preset reference temperature; Rock samples were taken from a low-temperature freeze-thaw simulation environment and placed on the sample stage of a non-destructive scanning device; The X-ray source and detector array of the non-destructive scanning device are activated and rotated to scan the rock sample. Acquired X-ray projection images of rock samples at different rotation angles; Perform fault reconstruction operations on all penetrating ray projection images to generate a set of internal structural fault slice images of rock samples at different preset freeze-thaw cycle nodes; The set of internal structure tomographic images is arranged axially to form an internal structure image sequence of the rock sample.

3. The method for testing the mechanical properties of rocks in low-temperature freeze-thaw environments based on microstructure evolution according to claim 1, characterized in that, Perform three-dimensional reconstruction processing on the internal structure image sequence of the rock sample to generate a digital three-dimensional structural model of the rock sample, including: Gray-level normalization and noise filtering were performed on each fault slice image in the internal structure image sequence of the rock sample. Pixel alignment and interpolation calculations are performed on the processed tomographic slice images to construct the initial three-dimensional voxel space data; In three-dimensional voxel space data, a preset threshold segmentation algorithm is applied to distinguish the spatial regions of rock matrix, pore space and different mineral phases; Based on the distinguished rock matrix, pore space, and spatial regions of different mineral phases, a digital three-dimensional structural model characterizing the internal material distribution of the rock sample is constructed.

4. The method for testing the mechanical properties of rocks in low-temperature freeze-thaw environments based on microstructure evolution according to claim 1, characterized in that, Based on the aforementioned digital three-dimensional structural model, the set of geometric parameters of the pore structure and the distribution map of mineral components inside the rock sample are extracted, including: In the pore space represented by the digital three-dimensional structural model, all connected pore networks are identified; For each connected pore network, calculate its volume, surface area, equivalent diameter, shape factor, tortuosity, and connectivity parameters, and combine them to form a set of geometric parameters for the pore structure. In the different mineral phase spatial regions represented by the digital three-dimensional structural model, the spatial distribution density and adjacency relationship of each mineral phase are statistically analyzed based on the three-dimensional spatial coordinates of each mineral phase. The spatial distribution density and adjacency relationship of each mineral phase are encoded in the form of a map to generate a mineral component distribution map.

5. The method for testing the mechanical properties of rocks in low-temperature freeze-thaw environments based on microstructure evolution according to claim 1, characterized in that, Evolution trajectory tracing is performed on the set of geometric parameters of the pore structure to generate a pore structure evolution map, including: Extract the changes in geometric parameters of pores at the same spatial location during the freeze-thaw cycle from the set of geometric parameters of the pore structure corresponding to different freeze-thaw cycle nodes; For each pore, calculate its volume change rate, shape evolution vector, and connectivity transition path between adjacent freeze-thaw cycle nodes; By integrating the changes in geometric parameters, volume change rate, shape evolution vector, and connectivity transition path of all pores, a dynamic relationship diagram of the evolution of pore structure with the increase of freeze-thaw cycles is drawn, namely, the pore structure evolution diagram.

6. The method for testing the mechanical properties of rocks in low-temperature freeze-thaw environments based on microstructure evolution according to claim 1, characterized in that, Perform component migration quantization processing on the mineral component distribution map to generate a component migration evolution map, including: By comparing the mineral composition distribution maps corresponding to different freeze-thaw cycle nodes, the spatial displacement of specific mineral particles can be identified. Calculate the displacement vector of the spatial displacement of the specific mineral particle, and statistically analyze the distribution characteristics of the displacement vectors of mineral particles of the same type. On the mineral component distribution map, mineral particles with significant displacement and their displacement trajectories are marked. By superimposing and integrating the mineral particle displacements and evolution trajectories observed in all freeze-thaw cycle nodes, a map reflecting the migration patterns and evolution trends of mineral components under freeze-thaw action is generated, namely, the component migration and evolution map.

7. The method for testing the mechanical properties of rocks in low-temperature freeze-thaw environments based on microstructure evolution according to claim 1, characterized in that, The pore structure evolution map and the component migration evolution map are input into the mechanical correlation model to generate a structural evolution feature vector, including: Extract the pore volume change rate sequence, pore morphology evolution rate sequence, and pore network connectivity change sequence from the pore structure evolution map; Statistical features of mineral displacement, mineral boundary change sequences, and microcrack propagation trajectories were extracted from the component migration and evolution map. The pore volume change rate sequence, pore morphology evolution rate sequence, pore network connectivity change sequence, mineral displacement statistical characteristics, mineral boundary change sequence, and microcrack propagation trajectory description are normalized and vectorized and concatenated to form the original feature set. The original feature set is input into the feature dimensionality reduction and fusion layer of the pre-trained mechanical correlation model, which compresses and fuses the high-dimensional original feature set into a low-dimensional structural evolution feature vector that is highly correlated with the mechanical response.

8. The method for testing the mechanical properties of rocks in low-temperature freeze-thaw environments based on microstructure evolution according to claim 1, characterized in that, Based on the structural evolution feature vector, a driving mechanical loading device applies a load of a preset pattern to the rock sample, and simultaneously acquires the mechanical response signal sequence of the rock sample, including: The structural evolution feature vector generated by the current freeze-thaw cycle node is analyzed. The structural evolution feature vector includes information on pore structure stability assessment and mineral component bonding strength assessment. Based on the pore structure stability assessment and mineral component bonding strength assessment information, a matching load application mode is selected from a preset loading mode library. The selected load application mode is converted into control commands for the mechanical loading device, which then applies a combination of axial stress, confining pressure, and shear force to the rock sample corresponding to the current structural state. During the application of load, stress, strain, acoustic emission signals and ultrasonic speed signals are simultaneously collected by a sensor array arranged on the rock sample, forming a mechanical response signal sequence of the rock sample.

9. The method for testing the mechanical properties of rocks in low-temperature freeze-thaw environments based on microstructure evolution according to claim 1, characterized in that, Based on the mechanical response signal sequence of the rock samples and the corresponding structural evolution feature vectors, a rock mechanical property evolution correlation mapping table is constructed, including: The mechanical response signal sequence of the rock sample obtained at each freeze-thaw cycle node is processed to extract mechanical parameters, including peak strength, elastic modulus, Poisson's ratio and stress-strain curve characteristic parameters. The mechanical parameters extracted from each freeze-thaw cycle node are paired with the structural evolution feature vector corresponding to the freeze-thaw cycle node to form a "mechanical parameter-structural evolution feature vector" data pair. Arrange the "mechanical parameter-structural evolution feature vector" data pairs of all freeze-thaw cycle nodes in ascending order of freeze-thaw cycle number; Based on the arranged data pairs, an association mapping structure is constructed. The association mapping structure uses the number of freeze-thaw cycles as an index to record the correspondence between mechanical parameters and structural evolution feature vectors, forming a rock mechanical property evolution association mapping table.

10. The method for testing the mechanical properties of rocks in low-temperature freeze-thaw environments based on microstructure evolution according to claim 9, characterized in that, A correlation mapping structure is constructed, which uses the number of freeze-thaw cycles as an index to record the correspondence between mechanical parameters and structural evolution feature vectors, including: Create a multidimensional data table, using the number of freeze-thaw cycles as the main index column of the multidimensional data table; In the multidimensional data table, for each freeze-thaw cycle number index row, establish a mechanical parameter column group and a structural evolution feature vector column group; Fill the extracted peak strength, elastic modulus, Poisson's ratio, and stress-strain curve characteristic parameters into the corresponding sub-columns of the mechanical parameter column group in the corresponding row; Fill each dimension component of the structural evolution feature vector into the corresponding sub-column of the structural evolution feature vector column group of the corresponding row; The mapping relationship between the corresponding set of mechanical parameters and the structural evolution feature vector can be directly obtained by querying the main index column of the multidimensional data table based on the number of freeze-thaw cycles.

Images