A numerical simulation method for shear characteristics of cold ice joint in cold region based on digital twinning

Abstract

The present application provides a kind of numerical simulation method of cold region cold ice joint shear characteristic based on digital twinning, it is related to geotechnical engineering technical field, the present application is prepared and three-dimensional scanning cold ice joint specimen, constructs the twinning model of two-way containing macroscopic layer and micro layer, in shear test, the macro mechanics, acoustics, vision and temperature field data of synchronous acquisition, the shear stress and acoustic emission cumulative energy of actual and simulation output are compared time by time to calculate model error, and when error is over threshold value, ice-rock interface adhesion strength and ice fracture energy are inverted and corrected, realize model dynamic updating and parameter optimization, finally output high-precision dynamic cohesion, peak friction angle and equivalent shear stiffness.

Description

Technical Field

[0001] This invention relates to the field of geotechnical engineering technology, specifically to a numerical simulation method for the shear characteristics of joints in cold-region ice based on digital twins. Background Technology

[0002] In cold-region rock engineering, the shear characteristics of glacial joints significantly impact engineering stability. However, existing numerical simulation methods often fail to accurately reflect the complex mechanical behavior of the ice-rock interface and its dynamic evolution under low-temperature-disturbance coupled environments. Traditional models rely heavily on macroscopic mechanical parameters, neglecting the bidirectional coupling mechanism between microscopic damage and macroscopic response, leading to significant discrepancies between simulation results and experimental data. Furthermore, key parameters such as ice fracture behavior and interfacial adhesion strength are difficult to obtain directly from experiments, limiting the model's applicability and prediction accuracy under real-world conditions. Therefore, a dynamic numerical simulation method capable of integrating multi-source monitoring data and achieving bidirectional feedback between macroscopic and microscopic dimensions is urgently needed to improve the accuracy and reliability of glacial joint shear characteristic analysis.

[0003] In the prior art, CN112784403B discloses a numerical simulation method for establishing a discrete element model of jointed rock mass based on point cloud data. First, the point cloud dataset of the jointed rock mass is obtained and subjected to denoising and simplification filtering. Then, the attitude of the structural planes is calculated. The three-dimensional model of the structural planes established by the point cloud data cannot be directly imported into the discrete element software 3DEC. Therefore, Fracman and Rhino software are used. Fracman software has powerful structural plane modeling function, and Rhino software has complete model processing function. The model is processed into a data file that can be read by 3DEC software, and then imported into 3DEC software for numerical analysis to generate the final three-dimensional model of the jointed rock mass.

[0004] The main problem with the above scheme is that it relies on the initial point cloud data to build the geometric model, but in the subsequent numerical analysis, the model itself is fixed and cannot be dynamically updated and corrected according to the real-time mechanical and acoustic data obtained during the experiment or monitoring. The key point of the scheme is to accurately transform the complex geometric structure expressed by the point cloud data into a discrete element model, which solves the problem of digitizing the geometric shape, but it does not delve into the coupling of macro-micro mechanical behavior inside the model, and cannot simulate the micro-damage evolution process such as the initiation and propagation of cracks inside ice and their correlation with acoustic emission energy.

[0005] The information disclosed in the background section is only intended to enhance the understanding of the background of this disclosure, and therefore may include information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0006] The purpose of this invention is to provide a numerical simulation method for the joint shear characteristics of cold-region ice based on digital twins, so as to solve the problems mentioned in the background art.

[0007] To achieve the above objectives, the present invention provides the following technical solution: A numerical simulation method for the joint shear characteristics of cold-region ice based on digital twins, the specific steps of which include: Step 1: Prepare cold-region ice joint specimens and scan to obtain point cloud data of the specimens. Based on the three-dimensional point cloud data of the specimens, as well as the ice-rock interface adhesion strength and ice fracture energy, construct an initial two-way twin model. Step 2: Conduct a frost joint shear test on the specimen, and collect macroscopic mechanical data, acoustic data, visual data and temperature field data at each moment during the test; Step 3: Acquire the time-series data of shear stress and cumulative acoustic emission energy output from the ice joint shear test at each time step, and input the macroscopic mechanical data, acoustic data, visual data and temperature field data collected from the ice joint shear test into the initial two-way twin model. Perform a simulated shear test on the initial two-way twin model to obtain the simulated shear stress and simulated cumulative acoustic emission energy at each acquisition time, so as to calculate the model error at each acquisition time. Step 4: Set an error threshold. When the error of the initial two-way twin model at a certain acquisition time exceeds the error threshold, invert the ice-rock interface adhesion strength and ice fracture energy of the initial two-way twin model at that acquisition time. Based on the inversion results, correct the initial two-way twin model until the model error of the corrected initial two-way twin model at each acquisition time does not exceed the error threshold. Output the corresponding model as the final two-way twin model. Step 5: Output the dynamic cohesion, peak friction angle, and equivalent shear stiffness of the specimen through the final two-way twin model.

[0008] Furthermore, the principle for constructing the initial bidirectional twin model is as follows: The initial bidirectional twin model includes a macroscopic layer and a mesoscopic layer; The macroscopic layer is the Mohr-Coulomb model. The input parameters include macroscopic mechanical data and temperature field data, and the output parameters include shear stress, dynamic cohesion, peak friction angle and equivalent shear stiffness. The input parameters for the mesoscopic layer include acoustic data, visual data, and temperature distribution field, while the output parameter is the cumulative acoustic emission energy. The shear stress calculated in the macroscopic layer is used as a boundary condition input to the mesoscopic layer, and the energy release from acoustic emission caused by crack propagation in the mesoscopic layer is fed back to the macroscopic layer.

[0009] Furthermore, the macroscopic mechanical data includes shear displacement, normal stress, and shear stress, thus obtaining a complete shear stress-displacement curve; the acoustic data includes the three-dimensional coordinates and energy of acoustic emission events; the visual data includes the specimen surface roughness, ice layer thickness, and the location, length, and propagation direction of cracks in the ice layer and ice-rock interface; and the temperature field data represents the temperature distribution field on the specimen surface.

[0010] Furthermore, the principle underlying the calculation of model error is as follows: A cryogenic shear test was conducted on the specimen. Continuous data acquisition was performed at each acquisition time to obtain the actual shear stress and actual cumulative acoustic emission energy of the specimen. The macroscopic mechanical data, acoustic data, visual data, and temperature field data of the specimen during the cryogenic shear test at each acquisition time were input into an initial two-way twin model for simulated shear testing. The simulated shear stress and simulated cumulative acoustic emission energy were obtained at each acquisition time. Based on the actual shear stress, actual cumulative acoustic emission energy, simulated shear stress, and simulated cumulative acoustic emission energy, the model error for each corresponding acquisition time was calculated using the following formula: in, Indicates the time of data collection Model error, Indicates the first Each data collection moment, Indicates the index of the acquisition time, and , Indicates the number of data collection moments. Indicates the time of data collection The actual shear stress, Indicates the time of data collection Simulated shear stress, This represents the shear stress normalization factor, which is the expected peak shear stress from the icy joint shear test. Indicates the time of data collection The actual cumulative acoustic emission energy, Indicates the time of data collection Simulated acoustic emission energy, This represents the normalization factor for the cumulative acoustic emission energy, which is the expected maximum cumulative acoustic emission energy from the icy joint shear test. These represent the weighting coefficients for shear stress and cumulative acoustic emission energy, respectively. ,and .

[0011] Furthermore, the principle underlying the inversion of the ice-rock interface adhesion strength and ice fracture energy of the initial two-way twin model is as follows: When the model error at a certain acquisition time is not greater than the error threshold, the initial bidirectional twin directly uses the current parameters and continues to simulate until the next acquisition time; when the model error at a certain acquisition time is greater than the error threshold, inversion is performed. The adhesion strength of the ice-rock interface and the fracture energy of ice are used as inversion targets; Generate initial vectors for ice-rock interface adhesion strength and ice fracture energy. The initial vector is subjected to two local perturbations to create two copies, A and B, of the bidirectional twin model. In copy A, the initial parameters are modified to... ,in, Represents the initial vector. This represents the initial value of the adhesion strength at the ice-rock interface. This represents the initial value of the ice's fracture energy. This represents a small increment in the adhesion strength at the ice-rock interface, and ; Record copy A from arrive Output and ,in, Indicates the first A collection time exceeding the threshold, Indicates the index of the collection time that exceeds the threshold. This indicates the time interval between adjacent data collection moments. Indicates that copy A is in The simulated shear stress output at that time, Indicates that copy A is in The accumulated energy of the simulated acoustic emission output is used to calculate the sensitivity, based on the following formula: in, This indicates the sensitivity of the simulated shear stress to changes in the adhesion strength at the ice-rock interface. This indicates the sensitivity of the simulated acoustic emission cumulative energy to the change in ice-rock interface adhesion strength. Indicates in Simulated shear stress at that time, Indicates in Accumulated energy from simulated acoustic emission; In copy B, the initial parameters are modified to... ,in, This represents a tiny increase in the fracture energy of ice, and ; Record copy B from arrive Output and ,in, Indicates that copy B is in The simulated shear stress output at that time, Indicates that copy B is in The accumulated energy of the simulated acoustic emission output is used to calculate the sensitivity, based on the following formula: in, This indicates the sensitivity of the simulation of shear stress as a function of the fracture energy of ice. This indicates the sensitivity of the simulated acoustic emission cumulative energy to the change in ice fracture energy; The Jacobian matrix is ​​constructed based on four sensitivities, and the specific matrix is ​​as follows: in, Represents the Jacobian matrix; Define error vector Then, the parameter update amount is calculated using the following formula: in, express The transpose of the matrix, Indicates the parameter update amount. Indicates the damping factor. Represents the identity matrix; The candidate parameter vector is calculated using the following formula: in, Represents the candidate parameter vector. Indicates the relaxation factor. , Represents the transpose of the parameter update. The ice-rock interface adhesion strength and ice fracture energy in the candidate parameter vector are the inverted ice-rock interface adhesion strength and ice fracture energy.

[0012] Furthermore, after each inversion to obtain the ice-rock interface adhesion strength and ice fracture energy, these two parameters are substituted into the mesoscale layer of the initial two-way twin model, replacing the original corresponding parameters; the initial two-way twin model with updated parameters is rerun, and the model error relative to the glacial joint shear test is calculated sequentially at each acquisition time until the model error at all acquisition times is no greater than the preset error threshold. The model obtained at this time is the final two-way twin model.

[0013] Compared with the prior art, the beneficial effects of the present invention are: This invention simultaneously collects macroscopic mechanical data, acoustic data, visual data, and temperature field data, improving the accuracy of constructing digital twin models. It constructs a two-way twin model from both macroscopic and mesoscopic levels. The macroscopic level directly reflects the overall mechanical response of the specimen, while the mesoscopic level reveals microscopic damage processes such as crack propagation and interface debonding within the ice body. This achieves cross-scale correlation from macroscopic to mesoscopic, reflecting the dynamic coupling and two-way feedback between macroscopic and mesoscopic behaviors of ice joints during shearing, ensuring that the simulation results are closer to the actual situation. This invention also achieves real-time error monitoring, comparing data at each acquisition moment to enable immediate detection and quantification of model errors. Through continuous error monitoring, the bidirectional twin model continuously adjusts its parameters throughout the shearing process to approximate the true values ​​as closely as possible, effectively avoiding error accumulation and distortion caused by parameter fixation in traditional simulations. This ensures high synchronization and high fidelity between the virtual model and the physical specimen throughout the entire lifecycle. By using real-time acquired macroscopic mechanical and acoustic data, the parameter values ​​that best enable the model output to closely approximate the actual response are derived, improving the prediction accuracy of the twin model throughout its entire lifecycle. Attached Figure Description

[0014] Figure 1 This is a schematic diagram of the method flow of an embodiment of the present invention; Figure 2 This is a schematic diagram illustrating the variation of shear stress with shear displacement in an embodiment of the present invention; Figure 3 This is a schematic diagram illustrating the change of acoustic emission accumulated energy over time according to an embodiment of the present invention. Detailed Implementation

[0015] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to specific embodiments.

[0016] It should be noted that, unless otherwise defined, the technical or scientific terms used in this invention should have the ordinary meaning understood by one of ordinary skill in the art to which this invention pertains. The terms "first," "second," and similar terms used in this invention do not indicate any order, quantity, or importance, but are merely used to distinguish different components. Terms such as "comprising" or "including" mean that the element or object preceding the word encompasses the elements or objects listed following the word and their equivalents, without excluding other elements or objects. Terms such as "connected" or "linked" are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect. Terms such as "upper," "lower," "left," and "right" are used only to indicate relative positional relationships; when the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0017] Example: Please see Figures 1 to 3 The present invention provides a technical solution: A numerical simulation method for the joint shear characteristics of cold-region ice based on digital twins, the specific steps of which include: Step 1: Prepare cold-region ice joint specimens and scan to obtain point cloud data of the specimens. Based on the three-dimensional point cloud data of the specimens, as well as the ice-rock interface adhesion strength and ice fracture energy, construct an initial two-way twin model. In this embodiment, rock masses with the same properties as actual rock masses in cold regions are selected and cut into rock block specimens with dimensions of 100mm×100mm×50mm. The rock block specimens are polished to ensure that the surface is flat and without significant unevenness. The surface roughness is measured using a surface profilometer. In a temperature environment below -10℃, pure water is uniformly sprayed onto the surface of the rock block specimens, allowing them to freeze naturally at low temperatures to form an ice-rock interface. Multiple freeze-thaw cycles are performed until the ice layer thickness reaches the expected thickness, thus completing the preparation of the cold region ice joint specimen. In a low-temperature environment, a laser 3D scanner is used to scan from multiple perspectives to obtain the 3D point cloud of the cold region ice joint specimen. Based on the 3D point cloud of the specimen, combined with the ice-rock interface adhesion strength and the fracture energy of ice, an initial two-way twin model is constructed.

[0018] Ice-rock interface adhesion strength represents the bond strength between the ice layer and the rock surface. It is affected by factors such as freezing conditions, surface roughness, and temperature, and is used to control the debonding behavior of the ice-rock interface. When the shear stress exceeds this adhesion strength, the ice and rock interface peel off. Ice fracture energy represents the ice body's ability to resist crack propagation, reflecting the ice's toughness or brittleness. It is used to control the initiation and propagation of cracks inside the ice body. The higher the fracture energy, the tougher the ice layer, the more difficult the crack propagation, and the slower the acoustic emission energy accumulation. An initial value is set for the ice-rock interface adhesion strength and the ice fracture energy as the starting point of the initial two-way twin model. The initial value is estimated based on the rock-ice interface adhesion strength and the ice fracture energy in the target cold region. In the target cold region, the actual temperature, humidity, and load conditions are simulated. The adhesion strength is determined by a pull-out test, and the fracture toughness of the ice is determined by a three-point bending test. The fracture energy is calculated by dividing the square of the fracture toughness by the elastic modulus of the ice. The above two values ​​are used as the initial values ​​of the ice-rock interface adhesion strength and the ice fracture energy in the initial two-way twin model. The principle for constructing the initial bidirectional twin model is as follows: The initial bidirectional twin model includes a macroscopic layer and a mesoscopic layer; The macroscopic layer is the Mohr-Coulomb model. The input parameters include macroscopic mechanical data and temperature field data, and the output parameters include shear stress, dynamic cohesion, peak friction angle and equivalent shear stiffness. The input parameters for the mesoscopic layer include acoustic data, visual data, and temperature distribution field, while the output parameter is the cumulative acoustic emission energy. The shear stress calculated in the macroscopic layer is used as a boundary condition input to the mesoscopic layer, and the energy release from acoustic emission caused by crack propagation in the mesoscopic layer is fed back to the macroscopic layer.

[0019] The macroscopic layer describes the strength changes of rock and soil masses under shear stress, characterizing the contributions of material cohesion and internal friction angle to shear strength. Its outputs are dynamic cohesion, peak friction angle, and equivalent shear stiffness. Dynamic cohesion reflects the strength of the bond between ice and rock, essentially stemming from intermolecular forces between ice crystals and mineral particles on the rock surface, the mechanical interlocking formed by ice penetrating into the micropores of rock during freezing, and the phase transition adhesion effect caused by temperature changes. Dynamic cohesion changes in real time with shearing, temperature variations, and damage accumulation; it reflects the bonding quality between the rock mass and ice layer—the higher the dynamic cohesion, the stronger the bond between the rock mass and ice layer; it also reflects temperature sensitivity and damage state. The peak friction angle reflects the sliding friction characteristics of the ice-rock interface. The limiting parameter, which appears when the shear stress reaches its peak, represents the maximum frictional resistance. It originates from the geometric interlock between the rock surface and the ice body, as well as the sliding resistance generated by the deformation of the ice body during shearing. It reflects the influence of surface roughness and ice hardness; the rougher the surface, the larger the friction angle. At low temperatures, the ice body hardens, and the friction angle increases. The equivalent shear stiffness reflects the overall stiffness of the ice-rock system against shear deformation. It originates from the elastic stiffness of the ice body itself, the deformation stiffness of the rock body as a whole, and the coupling stiffness of the contact interface. It reflects the integrity of the specimen and the degree of damage accumulation. The higher the equivalent shear stiffness, the more intact the specimen and the better the coupling. The more damage accumulates, the lower the equivalent shear stiffness. Dynamic cohesion and friction angle together determine the peak strength of the specimen, while stiffness determines the deformation characteristics of the specimen.

[0020] The mesoscale layer is used to reveal the internal damage evolution of the specimen, specifically including crack initiation and propagation within the ice body and debonding at the ice-rock interface, providing a microscopic mechanism explanation for macroscopic phenomena. Through processing acoustic, visual, and temperature field data, the mesoscale layer simulates the initiation and propagation of internal cracks in the ice layer during shearing, as well as the ice... The debonding behavior of the rock interface; the output of the microstructure is the acoustic emission cumulative energy, which reflects the cumulative state of internal damage of the material. The acoustic emission cumulative energy is the sum of the acoustic energy released during the shearing process due to the propagation of internal cracks in the ice body and the debonding of the interface. The higher the acoustic emission cumulative energy, the more obvious the internal crack propagation and the more serious the damage. A sudden increase in energy usually corresponds to the peak strength or the destruction of the specimen.

[0021] The macroscopic layer reflects the overall mechanical response of the specimen, such as shear stress-displacement relationship, strength parameters, and stiffness changes. Based on the Mohr-Coulomb model, it describes the overall stress-strain behavior of the ice-rock system under external loads. The mesoscopic layer reflects the microscopic damage processes such as the initiation and propagation of cracks inside the ice, debonding of the ice-rock interface, and release of acoustic emission energy. Based on acoustic emission, visual, and temperature data, it simulates the damage evolution and its energy release. Based on the macroscopic and mesoscopic layers, an initial two-way twin model is constructed through bidirectional coupling. The macroscopic stress state drives the mesoscopic damage, and the mesoscopic damage feedback affects the macroscopic stiffness and strength, reflecting the dynamic adjustment process of load-damage-stiffness degradation-stress redistribution. Traditional macroscopic models neglect microscopic damage, leading to overly stiff predictions. The two-way coupled model can more realistically reflect the strength softening and stiffness degradation caused by damage accumulation. The acoustic emission energy output from the microscopic layer and the macroscopic shear stress form a dual verification, providing inversion basis for parameters that are difficult to measure directly, such as ice-rock interface adhesion strength and ice fracture energy. By comparing experimental and simulation data in real time, the model can be dynamically updated and predict the mechanical behavior of the next stage, making it suitable for early warning of engineering stability.

[0022] The model is constructed at both macroscopic and mesoscopic levels because shear behavior is governed by both overall mechanical laws and microscopic mechanisms such as ice microfracture and interface adhesion, which cannot be fully described by a single-scale model. Mechanical sensor data and acoustic emission and visual data come from different sources and scales, so layered processing is conducive to efficient data fusion and mechanism mapping. Relying solely on macroscopic data to invert parameters is prone to multiple solutions. Combining mesoscopic acoustic emission energy can enhance the uniqueness of the inversion and achieve more accurate dynamic updates.

[0023] Step 2: Conduct a frost joint shear test on the specimen, and collect macroscopic mechanical data, acoustic data, visual data and temperature field data at each moment during the test; In this embodiment, the macroscopic mechanical data includes shear displacement, normal stress, and shear stress, thus obtaining a complete shear stress-displacement curve; the acoustic data includes the three-dimensional coordinates and energy of acoustic emission events; the visual data includes the surface roughness of the specimen, the thickness of the ice layer, the location, length, and propagation direction of cracks in the ice layer and ice-rock interface; and the temperature field data represents the temperature distribution field on the surface of the specimen.

[0024] The prepared icy joint specimens were placed in a cryogenic-disturbance coupled test chamber to simulate the actual temperature, humidity, and load conditions in cold regions. Various sensors were installed and calibrated, including mechanical sensors for measuring shear displacement, normal stress, and shear stress; acoustic emission sensors for capturing the three-dimensional coordinates and energy of acoustic emission events; a high-resolution camera and infrared thermal imager for recording the surface morphology, crack evolution, and temperature distribution of the specimens; and a temperature sensor array for acquiring the surface and internal temperature fields of the specimens. During the icy joint shear test, macroscopic mechanical data, acoustic data, visual data, and temperature field data were simultaneously acquired at fixed time intervals.

[0025] Step 3: Acquire the time-series data of shear stress and cumulative acoustic emission energy output from the ice joint shear test at each time step, and input the macroscopic mechanical data, acoustic data, visual data and temperature field data collected from the ice joint shear test into the initial two-way twin model. Perform a simulated shear test on the initial two-way twin model to obtain the simulated shear stress and simulated cumulative acoustic emission energy at each acquisition time, so as to calculate the model error at each acquisition time. In this embodiment, the principle underlying the calculation of model error is as follows: A cryogenic shear test was conducted on the specimen. Continuous data acquisition was performed at each acquisition time to obtain the actual shear stress and actual cumulative acoustic emission energy of the specimen. The macroscopic mechanical data, acoustic data, visual data, and temperature field data of the specimen during the cryogenic shear test at each acquisition time were input into an initial two-way twin model for simulated shear testing. The simulated shear stress and simulated cumulative acoustic emission energy were obtained at each acquisition time. Based on the actual shear stress, actual cumulative acoustic emission energy, simulated shear stress, and simulated cumulative acoustic emission energy, the model error for each corresponding acquisition time was calculated using the following formula: in, Indicates the time of data collection Model error, Indicates the first Each data collection moment, Indicates the index of the acquisition time, and , Indicates the number of data collection moments. Indicates the time of data collection The actual shear stress, Indicates the time of data collection Simulated shear stress, This represents the shear stress normalization factor, which is the expected peak shear stress from the icy joint shear test. Indicates the time of data collection The actual cumulative acoustic emission energy, Indicates the time of data collection Simulated acoustic emission energy, This represents the normalization factor for the cumulative acoustic emission energy, which is the expected maximum cumulative acoustic emission energy from the icy joint shear test. These represent the weighting coefficients for shear stress and cumulative acoustic emission energy, respectively. ,and .

[0026] During the icy joint shear test, as the test progresses, the macroscopic mechanical data, acoustic data, visual data, and temperature field data also change. By acquiring the above data at different acquisition times and inputting them into the initial bidirectional twin model, it can be ensured that the data in the simulation test are consistent with the actual test. By dividing by their respective normalization factors, shear stress and acoustic emission energy are converted into dimensionless relative errors. Shear stress is a core indicator of macroscopic mechanical response, and its error significantly affects model accuracy. Therefore, its squared term is calculated to reflect the nonlinear effect of the error, penalizing large errors and making the model more sensitive to shear stress prediction. Acoustic emission cumulative energy reflects the accumulation of mesoscopic damage, and its value is usually a monotonically increasing function. The error accumulates over time, and the linear term reflects the trend of this cumulative deviation. Shear stress is a direct mechanical response and is more critical for engineering stability assessment. Acoustic emission energy is a time-dependent damage indicator, more often used for mesoscopic mechanism verification; therefore, its weighting coefficient is... ,Pick , .

[0027] The principle of calculating model error is to quantify the difference between actual shear test data and the simulated output data of the initial two-way twin model at a specific acquisition time by comparing the two. In the actual icy joint shear test, the following two types of key data are collected in real time: actual shear stress and actual acoustic emission cumulative energy. Macroscopic mechanical data, acoustic data, visual data, and temperature field data at the same time are input into the initial two-way twin model, and the simulated shear test is run to obtain simulated shear stress and simulated acoustic emission cumulative energy. The model error at each acquisition time is calculated and compared with the error threshold. If the error threshold is not exceeded, it means that the current model's prediction value is close to the actual value. If the values ​​are close, no model adjustment is needed. If the error exceeds the error threshold, it indicates a significant gap between the model's predicted values ​​and the actual values, triggering inversion for model updates. Model error reflects the degree of matching between the model output and the actual experiment. The smaller the error, the closer the simulation results of the twin model are to the actual physical experiment, and the higher the model fidelity. A large error indicates that some key parameters in the model, such as the adhesion strength of the ice-rock interface and the fracture energy of ice, deviate from the actual situation, requiring parameter inversion and dynamic updates. A consistently low error indicates that the model has good predictive ability and can be used for shear behavior simulation and engineering stability assessment under subsequent working conditions.

[0028] Table 1 reflects the shear stress changes of the icy joint specimens during shear loading. When the shear displacement range is 0~1.5mm, the shear stress increases approximately linearly with the shear displacement, indicating that the material is in the elastic deformation stage and damage is not yet significant. When the shear displacement range is 1.5mm~2.2mm, the rate of increase of shear stress slows down, the material enters the nonlinear deformation stage, and internal microcracks begin to initiate and propagate. When the peak strength is reached at about 2.2mm, the maximum shear stress is reached, indicating that the overall shear resistance of the specimen has reached its limit. When it exceeds 2.2mm, the shear stress decreases with increasing displacement, indicating that the specimen has undergone macroscopic failure.

[0029] Table 1. Shear stress as a function of shear displacement As shown in Table 2, the damage accumulation over time is reflected by the simulated and actual accumulated energy. Both the simulated and actual values ​​show a monotonically increasing trend, indicating that the model can capture the physical process of damage accumulation. Both also show a gradually accelerating growth rate, reflecting the law of damage acceleration during shearing. The simulated accumulated energy is slightly lower than the actual accumulated energy, indicating that the model underestimates the rate or number of crack propagation and does not fully simulate energy release mechanisms such as interface debonding. Further inversion is needed to correct this.

[0030] Table 2. Cumulative Energy Variation over Time Step 4: Set an error threshold. When the error of the initial two-way twin model at a certain acquisition time exceeds the error threshold, invert the ice-rock interface adhesion strength and ice fracture energy of the initial two-way twin model at that acquisition time. Based on the inversion results, correct the initial two-way twin model until the model error of the corrected initial two-way twin model at each acquisition time does not exceed the error threshold. Output the corresponding model as the final two-way twin model. In this embodiment, the principle underlying the inversion of the ice-rock interface adhesion strength and ice fracture energy of the initial two-way twin model is as follows: When the model error at a certain acquisition time is not greater than the error threshold, the initial bidirectional twin directly uses the current parameters and continues to simulate until the next acquisition time; when the model error at a certain acquisition time is greater than the error threshold, inversion is performed. The adhesion strength of the ice-rock interface and the fracture energy of ice are used as inversion targets; Generate initial vectors for ice-rock interface adhesion strength and ice fracture energy. The initial vector is subjected to two local perturbations to create two copies, A and B, of the bidirectional twin model. In copy A, the initial parameters are modified to... ,in, Represents the initial vector. This represents the initial value of the adhesion strength at the ice-rock interface. This represents the initial value of the ice's fracture energy. This represents a small increment in the adhesion strength at the ice-rock interface, and ; The principle behind selecting ice-rock interface adhesion strength and ice fracture energy as the inversion targets is as follows: In the initial two-dimensional twin model, the role of its mesoscopic layer is to simulate the two key mesoscopic damage processes of crack propagation inside the ice body and debonding at the ice-rock interface. The fracture energy of ice directly determines the ease with which cracks initiate and propagate inside the ice body. The higher the fracture energy, the tougher the ice body and the stronger its ability to resist crack propagation. During the shearing process, a large number of microcracks are generated inside the ice layer. The propagation and convergence of these cracks are the main sources of the reduction in macroscopic strength, stiffness degradation, and acoustic emission energy release of the ice layer. The accumulated acoustic emission energy is essentially the elastic energy released during crack propagation. Therefore, reflecting the fracture energy of ice allows the crack propagation process simulated by the two-dimensional twin model to match the actual process. The adhesion strength of the ice-rock interface determines the firmness of the bond between the ice layer and the rock, reflecting the combined effects of mechanical interlocking force and intermolecular forces generated by ice penetrating into the rock pores during the freezing process. Shear failure usually takes two forms: internal shear failure of the ice body and debonding at the ice-rock interface. The interface adhesion strength controls when the second failure mode occurs and under what stress it occurs, directly affecting the macroscopic shear stress transfer efficiency and peak strength. Therefore, by inverting the ice-rock interface adhesion strength, we can ensure that the shear stress transfer process matches the actual situation.

[0031] Record copy A from arrive Output and ,in, Indicates the first A collection time exceeding the threshold, Indicates the index of the collection time that exceeds the threshold. This indicates the time interval between adjacent data collection moments. Indicates that copy A is in The simulated shear stress output at that time, Indicates that copy A is in The accumulated energy of the simulated acoustic emission output is used to calculate the sensitivity, based on the following formula: in, This indicates the sensitivity of the simulated shear stress to changes in the adhesion strength at the ice-rock interface. This indicates the sensitivity of the simulated acoustic emission cumulative energy to the change in ice-rock interface adhesion strength. Indicates in Simulated shear stress at that time, Indicates in Accumulated energy from simulated acoustic emission; Sensitivity represents how sensitive an output variable, such as shear stress or acoustic emission cumulative energy, is to changes in an input parameter, such as adhesion strength or fracture energy. Two model copies, A and B, are created, respectively, for... and When performing small perturbations and calculating the sensitivity using the finite difference method, while calculating the sensitivity of shear stress to adhesion strength, maintain... Unchanged, will Add a tiny amount Run model copy A to obtain the new shear stress. Then calculate the sensitivity. ; This represents the rate of change of macroscopic shear stress when the ice-rock interface adhesion strength increases by one unit. Essentially, it reflects the contribution of interface adhesion strength to the overall shear stress. A large and positive value indicates that increasing the interfacial adhesion strength will significantly increase the shear stress. A relatively small value indicates that shear stress is not sensitive to interfacial adhesion strength, and the failure mode is more likely to be internal shear failure leading to ice layer fracture. When calculating the sensitivity of acoustic emission energy to interfacial adhesion strength, if the value is negative and has a large absolute value, it means that increasing adhesion strength will reduce the accumulated acoustic emission energy because the interface is less prone to debonding, thus reducing crack propagation. If the value is positive, it means that increasing adhesion strength increases the accumulated acoustic emission energy, possibly due to stress concentration causing earlier internal cracking of the ice. This represents the change in cumulative acoustic emission energy when the interfacial adhesion strength increases by one unit, reflecting the influence of the interfacial adhesion state on the evolution of microscopic damage.

[0032] In copy B, the initial parameters are modified to... ,in, This represents a tiny increase in the fracture energy of ice, and ; Record copy B from arrive Output and ,in, Indicates that copy B is in The simulated shear stress output at that time, Indicates that copy B is in The accumulated energy of the simulated acoustic emission output is used to calculate the sensitivity, based on the following formula: in, This indicates the sensitivity of the simulation of shear stress as a function of the fracture energy of ice. This indicates the sensitivity of the simulated acoustic emission cumulative energy to the change in ice fracture energy; When calculating the sensitivity of shear stress to fracture energy, maintain Unchanged, will Add a small increment, run model copy B, obtain the new shear stress, and calculate the sensitivity. ; The rate of change of shear stress when the fracture energy of ice increases by one unit reflects the contribution of ice toughness to macroscopic strength. A higher value indicates a more significant contribution of ice toughness to shear strength, meaning the ice fractures only after undergoing significant plastic deformation. A lower value indicates that ice toughness has little impact on macroscopic strength, and the ice is more prone to brittle fracture. The sensitivity of acoustic emission energy to the fracture energy of ice represents the change in cumulative acoustic emission energy when the fracture energy increases by one unit. This reflects the influence of ice toughness on the energy release process of crack propagation. When the value is less than zero, a smaller value indicates stronger ice toughness, requiring more energy to drive crack propagation, and releasing less acoustic energy with each crack formation, indicating that acoustic emission mainly originates from ice fracture. A value closer to 0 indicates that changes in fracture energy have little impact on acoustic emission, and the acoustic emission energy mainly originates from interfacial debonding and other mechanisms.

[0033] Based on the different sensitivity values ​​and signs, the dominant mechanism of destruction can be determined. Larger and positive Smaller and negative, Both are close to zero; at this point, the failure mode is that the ice layer peels off completely from the rock surface, with significant interface slippage. The value is relatively small. Approaching zero The value is a positive value of moderate size. When the value is a moderately negative value, multiple cracks appear inside the ice layer, exhibiting a brittle fracture mode; when The value is relatively small. Approaching zero Take a large positive value. A small negative value indicates that the ice layer is undergoing parallel plastic shearing, with few cracks but a large degree of deformation. Other cases are mixed failures where interface debonding and ice fracture occur simultaneously.

[0034] The value of each small increment is taken as 1% of the original parameter. This is to ensure that the approximation is locally linear and to avoid crossing multiple mechanical states, which would lead to loss of local sensitivity. At the same time, it is to prevent the sensitivity calculation from being unstable.

[0035] The Jacobian matrix is ​​constructed based on four sensitivities, and the specific matrix is ​​as follows: in, Represents the Jacobian matrix; Define error vector Then, the parameter update amount is calculated using the following formula: in, express The transpose of the matrix, Indicates the parameter update amount. Indicates the damping factor. Represents the identity matrix; The Jacobian matrix describes the sensitivity of each output variable to each input parameter. Each element in the matrix is ​​a sensitivity, representing the strength of a particular output's response to a particular parameter. By comparing the sign and magnitude of different sensitivities, the dominant failure mode in the shearing process is determined. The error vector reflects the error between the model output and experimental data, aiming to find the parameter update amount. This makes the updated model output closer to the experimental data, i.e. To minimize the variance, take the derivative of the objective function and set the derivative to zero. Solving for To increase numerical stability, a damping term is introduced, resulting in... , It is the damping factor, and The damping factor is used to adjust the update step size.

[0036] The candidate parameter vector is calculated using the following formula: in, Represents the candidate parameter vector. Indicates the relaxation factor. , Represents the transpose of the parameter update. Solving Subsequently, a relaxation factor is introduced during real-time updates to prevent overshoot. This is achieved by adjusting the initial vector using the transpose of the parameter update vector. The reason for using the transpose of the parameter update vector is that, according to calculations, the parameter update vector is a 2×1 column vector, while the initial vector... is a 1×2 dimension row vector. To ensure that the two have the same dimensions, the transpose of the parameter update vector is combined with the initial vector; the calculated candidate parameter vector is... This is the result of the inversion.

[0037] The ice-rock interface adhesion strength and ice fracture energy in the candidate parameter vector are the inverted ice-rock interface adhesion strength and ice fracture energy.

[0038] After each inversion to obtain the ice-rock interface adhesion strength and ice fracture energy, these two parameters are substituted into the mesoscale layer of the initial two-way twin model, replacing the original corresponding parameters. The initial two-way twin model with updated parameters is run again, and the model error relative to the glacial joint shear test is calculated at each acquisition time until the model error at all acquisition times is no greater than the preset error threshold. The model obtained at this time is the final two-way twin model.

[0039] Step 5: Output the dynamic cohesion, peak friction angle, and equivalent shear stiffness of the specimen through the final two-way twin model.

[0040] In this embodiment, the final bidirectional twin model consists of a macroscopic layer and a mesoscopic layer, which interact with each other through a bidirectional coupling mechanism. The macroscopic layer takes in macroscopic mechanical data and temperature field data, and outputs shear stress, dynamic cohesion, peak friction angle, and equivalent shear stiffness; the mesoscopic layer takes in acoustic data, visual data, and temperature distribution field, and outputs acoustic emission accumulated energy. In the model, cohesion is a crucial parameter describing the material's resistance to shear failure. During shearing, the macroscopic layer dynamically adjusts the cohesion value based on real-time mechanical and temperature data, reflecting changes in the adhesion strength at the ice-rock interface. Since the model continuously updates the ice-rock interface adhesion strength during inversion, the cohesion also changes dynamically, no longer being a fixed value. During shearing, when the shear stress reaches its peak, the model automatically calculates the corresponding friction angle. This value is affected by factors such as surface roughness, temperature, and damage state. Because the model considers factors like temperature and damage evolution, the friction angle also dynamically adjusts with the shearing process. The equivalent shear stiffness reflects the overall stiffness of the system against shear deformation and is the stiffness parameter output by the macroscopic layer. Based on the relationship curve between shear stress and shear displacement, the model calculates the stiffness value in real time. As damage accumulates, the stiffness gradually degrades, and the model dynamically reflects this process. The model updates the ice-rock interface adhesion strength and the fracture energy of ice through multiple inversions, resulting in a high degree of agreement between the simulation output and experimental data. The macroscopic layer directly outputs accurate dynamic cohesion, peak friction angle, and equivalent shear stiffness.

[0041] The above formulas are all dimensionless calculations. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0042] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented in software, the above embodiments can be implemented, in whole or in part, as a computer program product. Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution.

[0043] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment, depending on actual needs.

[0044] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application.

Claims

1. A numerical simulation method for the joint shear characteristics of cold-region ice based on digital twins, characterized in that, The specific steps include: Step 1: Prepare cold-region ice joint specimens and scan to obtain point cloud data of the specimens. Based on the three-dimensional point cloud data of the specimens, as well as the ice-rock interface adhesion strength and ice fracture energy, construct an initial two-way twin model. Step 2: Conduct a frost joint shear test on the specimen, and collect macroscopic mechanical data, acoustic data, visual data and temperature field data at each moment during the test; Step 3: Acquire the time-series data of shear stress and cumulative acoustic emission energy output from the ice joint shear test at each time step, and input the macroscopic mechanical data, acoustic data, visual data and temperature field data collected from the ice joint shear test into the initial two-way twin model. Perform a simulated shear test on the initial two-way twin model to obtain the simulated shear stress and simulated cumulative acoustic emission energy at each acquisition time, so as to calculate the model error at each acquisition time. Step 4: Set an error threshold. When the error of the initial two-way twin model at a certain acquisition time exceeds the error threshold, invert the ice-rock interface adhesion strength and ice fracture energy of the initial two-way twin model at that acquisition time. Based on the inversion results, correct the initial two-way twin model until the model error of the corrected initial two-way twin model at each acquisition time does not exceed the error threshold. Output the corresponding model as the final two-way twin model. Step 5: Output the dynamic cohesion, peak friction angle, and equivalent shear stiffness of the specimen through the final two-way twin model.

2. The numerical simulation method for joint shear characteristics of cold-region ice based on digital twins according to claim 1, characterized in that: The principle behind constructing the initial bidirectional twin model in step 1 is as follows: The initial bidirectional twin model includes a macroscopic layer and a mesoscopic layer; The macroscopic layer is the Mohr-Coulomb model. The input parameters include macroscopic mechanical data and temperature field data, and the output parameters include shear stress, dynamic cohesion, peak friction angle and equivalent shear stiffness. The input parameters for the mesoscopic layer include acoustic data, visual data, and temperature distribution field, while the output parameter is the cumulative acoustic emission energy. The shear stress calculated in the macroscopic layer is used as a boundary condition input to the mesoscopic layer, and the energy release from acoustic emission caused by crack propagation in the mesoscopic layer is fed back to the macroscopic layer.

3. The numerical simulation method for joint shear characteristics of cold-region ice based on digital twins according to claim 1, characterized in that: In step 2, the macroscopic mechanical data includes shear displacement, normal stress, and shear stress, thus obtaining a complete shear stress-displacement curve; the acoustic data includes the three-dimensional coordinates and energy of acoustic emission events; the visual data includes the surface roughness of the specimen, the thickness of the ice layer, the location, length, and propagation direction of cracks in the ice layer and ice-rock interface; and the temperature field data represents the temperature distribution field on the surface of the specimen.

4. The numerical simulation method for joint shear characteristics of cold-region ice based on digital twins according to claim 1, characterized in that: The principle underlying the calculation of model error in step 3 is as follows: A cryogenic shear test was conducted on the specimen. Continuous data acquisition was performed at each acquisition time to obtain the actual shear stress and actual cumulative acoustic emission energy of the specimen. The macroscopic mechanical data, acoustic data, visual data, and temperature field data of the specimen during the cryogenic shear test at each acquisition time were input into an initial two-way twin model for simulated shear testing. The simulated shear stress and simulated cumulative acoustic emission energy were obtained at each acquisition time. Based on the actual shear stress, actual cumulative acoustic emission energy, simulated shear stress, and simulated cumulative acoustic emission energy, the model error for each corresponding acquisition time was calculated using the following formula: in, Indicates the time of data collection Model error, Indicates the first Each data collection moment, Indicates the index of the acquisition time, and , Indicates the number of data collection moments. Indicates the time of data collection The actual shear stress, Indicates the time of data collection Simulated shear stress, This represents the shear stress normalization factor, which is the expected peak shear stress from the icy joint shear test. Indicates the time of data collection The actual acoustic emission accumulated energy Indicates the time of data collection Simulated acoustic emission energy, This represents the normalization factor for the cumulative acoustic emission energy, which is the expected maximum cumulative acoustic emission energy from the icy joint shear test. These represent the weighting coefficients for shear stress and cumulative acoustic emission energy, respectively. ,and .

5. The numerical simulation method for joint shear characteristics of cold-region ice based on digital twins according to claim 4, characterized in that: The principle underlying the inversion of the ice-rock interface adhesion strength and ice fracture energy of the initial two-way twin model in step 4 is as follows: When the model error at a certain acquisition time is not greater than the error threshold, the initial bidirectional twin directly uses the current parameters and continues to simulate until the next acquisition time; when the model error at a certain acquisition time is greater than the error threshold, inversion is performed. The adhesion strength at the ice-rock interface and the fracture energy of ice are used as inversion targets; Generate initial vectors for ice-rock interface adhesion strength and ice fracture energy. The initial vector is subjected to two local perturbations to create two copies, A and B, of the bidirectional twin model. In copy A, the initial parameters are modified to... ,in, Represents the initial vector. This represents the initial value of the adhesion strength at the ice-rock interface. This represents the initial value of the ice's fracture energy. This represents a small increment in the adhesion strength at the ice-rock interface, and ; Record copy A from arrive Output and ,in, Indicates the first A collection time exceeding the threshold, Indicates the index of the collection time that exceeds the threshold. This indicates the time interval between adjacent data collection moments. Indicates that copy A is in The simulated shear stress output at that time, Indicates that copy A is in The accumulated energy of the simulated acoustic emission output is used to calculate the sensitivity, based on the following formula: in, This indicates the sensitivity of the simulated shear stress to changes in the adhesion strength at the ice-rock interface. This indicates the sensitivity of the simulated acoustic emission cumulative energy to the change in ice-rock interface adhesion strength. Indicates in Simulated shear stress at that time, Indicates in Accumulated energy from simulated acoustic emission; In copy B, the initial parameters are modified to... ,in, This represents a tiny increase in the fracture energy of ice, and ; Record copy B from arrive Output and ,in, Indicates that copy B is in The simulated shear stress output at that time, Indicates that copy B is in The accumulated energy of the simulated acoustic emission output is used to calculate the sensitivity, based on the following formula: in, This indicates the sensitivity of the simulation of shear stress as a function of the fracture energy of ice. This indicates the sensitivity of the simulated acoustic emission cumulative energy to the change in ice fracture energy; The Jacobian matrix is ​​constructed based on four sensitivities, and the specific matrix is ​​as follows: in, Represents the Jacobian matrix; Define error vector Then, the parameter update amount is calculated using the following formula: in, express The transpose of the matrix, Indicates the parameter update amount. Indicates the damping factor. Represents the identity matrix; The candidate parameter vector is calculated using the following formula: in, Represents the candidate parameter vector. Indicates the relaxation factor. , Represents the transpose of the parameter update. The ice-rock interface adhesion strength and ice fracture energy in the candidate parameter vector are the inverted ice-rock interface adhesion strength and ice fracture energy.

6. The numerical simulation method for joint shear characteristics of cold-region ice based on digital twins according to claim 1, characterized in that: In step 4, after each inversion to obtain the ice-rock interface adhesion strength and ice fracture energy, these two parameters are substituted into the mesoscale layer of the initial two-way twin model, replacing the original corresponding parameters; the initial two-way twin model with updated parameters is rerun, and the model error relative to the glacial joint shear test is calculated sequentially at each acquisition time until the model error at all acquisition times is no greater than the preset error threshold. The model obtained at this time is the final two-way twin model.

Images