Virtual reality system for rock uniaxial compression experiment

Abstract

The invention provides a virtual reality system for a rock uniaxial compression experiment, and the system comprises an offline calculation module which is used for generating a numerical simulation sample set based on a random simulation scheme generation service and an elastic brittle damage constitutive model driven numerical simulation system; the online calculation module is used for constructing and training a numerical simulation process agent model based on the numerical simulation sample set and the U-Net network; and the mixed reality visualization and interaction module is used for applying the trained numerical simulation process agent model to a mixed reality scene to complete a mixed reality experiment. According to the virtual experiment system provided by the invention, different accurate experiment results can be obtained according to different parameters, meanwhile, real experiment operation experience is obtained, the experiment cost can be greatly saved, and the environmental protection property is improved; the experiment risk is reduced, and the safety is improved; the space-time limitation is broken, and the self-learning enthusiasm of students is stimulated.

Description

Technical Field

[0001] The invention relates to the technical field of rock mechanics virtual simulation experiments, and in particular to a rock uniaxial compression experiment virtual reality system. Background Art

[0002] Rock mechanics experiments are an important way for students to learn and master the theoretical knowledge of rock mechanics and cultivate practical ability, and they play an indispensable role in teaching and scientific research. However, rock mechanics experiments usually require the use of large equipment, such as presses, etc. During the experiment, there are phenomena such as rock crushing and splashing. Improper operation can easily cause safety accidents. Conducting experimental training is a necessary means to reduce the occurrence of accidents. However, the traditional manual experimental training method requires a lot of time and manpower, is inefficient, and is difficult to meet the needs of large-scale training, and the training effect is difficult to guarantee. VR virtual simulation experiments can simulate real rock mechanics experiments, allowing students and operators to learn and train in a safe environment. It has the advantages of strong interactivity, intuitive effects, strong targeting, flexibility, convenience, and repeatability. In recent years, based on virtual reality technology, mixed reality technology (MR) has been derived. This technology, with its unique characteristics such as virtual and real combination, tracking registration and intelligent interaction, brings users an immersive experience. Based on this, students can get the same interactive experience as using a real test machine, which not only greatly improves the fun of learning, but also enables students to perform more proficiently and standardized when actually operating the test machine in the future.

[0003] However, most current virtual reality experiments rely on static data resources, that is, fixed experimental results are displayed through preset programs and data, which makes it difficult for students to truly understand the physical and mechanical properties of rocks and has certain limitations. Numerical simulation is an effective tool for truly displaying the rock damage and destruction process, but its calculation time is generally long, especially for complex constitutive models, which is difficult to apply to such fast-paced training. In addition, general numerical simulation software has copyright issues and is difficult to access virtual reality systems. Summary of the invention

[0004] In view of this, the purpose of the present invention is to propose a rock uniaxial compression experiment virtual reality system to solve the technical problem that the existing virtual reality experiment cannot give students the experience of a real experiment.

[0005] The technical means adopted by the present invention are as follows:

[0006] A rock uniaxial compression experiment virtual reality system comprises the following steps:

[0007] Offline calculation module, used to generate numerical simulation sample sets based on random simulation scheme generation service and numerical simulation system driven by elastic-brittle damage constitutive model;

[0008] Online computing module, used to build and train the numerical simulation process proxy model based on the numerical simulation sample set and U-Net network;

[0009] The mixed reality visualization and interaction module is used to apply the trained numerical simulation process agent model to the mixed reality scene to complete the mixed reality experiment.

[0010] Furthermore, the calculation steps of the offline calculation module are as follows:

[0011] S11, forming a random simulation scheme generation service;

[0012] S12. Construct a numerical simulation system driven by elastic-brittle damage constitutive model;

[0013] S13. Convert the random simulation scheme generation service into a Socket client, convert the numerical simulation system driven by the elastic-brittle damage constitutive model into a numerical simulation Socket server, and dynamically transfer the random simulation scheme to the numerical simulation system driven by the elastic-brittle damage constitutive model, so as to obtain the multi-source mechanical response output results matching the random simulation scheme, and integrate all the multi-source mechanical response output results to form a numerical simulation sample set.

[0014] Furthermore, S11 specifically includes the following steps:

[0015] Determine the experimental mechanical parameters of rock, including elastic modulus E, Poisson's ratio μ, density ρ, loading displacement d and heterogeneity coefficient n;

[0016] Determine the range of experimental mechanical parameters;

[0017] The Latin hypercube sampling method is used to sample within the range of experimental mechanical parameters to generate random simulation schemes, which are then coded in Python to form a random simulation scheme generation service.

[0018] Furthermore, S12 specifically includes the following steps:

[0019] The isotropic continuous damage formula is used to characterize the damage evolution law of the material; the isotropic continuous damage formula is as follows:

[0020] ε ij =((1+v)σ ij -υσ kk δ ij ) / E (1)

[0021] E=E0(1-D) (2)

[0022] In the formula, ε ij is the damage elastic strain tensor, σ ijis the Cauchy stress tensor, E and E0 are the elastic moduli of the damaged and undamaged materials respectively, D is the damage variable, υ is the Poisson's ratio, δ ij is the Kronecker symbol;

[0023] Among them, for rock damage and failure, the maximum tensile stress criterion is preferentially used to judge the tensile failure of the rock; when the maximum tensile stress criterion is not met, the Mohr-Coulomb criterion is used to evaluate the shear failure of the rock, and the elasto-brittle damage constitutive model is further embedded in COMSOL to construct a numerical simulation system driven by the elasto-brittle damage constitutive model.

[0024] Furthermore, S13 specifically includes the following steps: based on the Socket communication protocol, converting the random simulation scheme generation service based on Python coding into a Socket client; using Java secondary development method, converting the numerical simulation system driven by the elastic-brittle damage constitutive model based on COMSOL software into a numerical simulation Socket server, and through the Socket communication protocol, dynamically transmitting the random simulation scheme to the numerical simulation system driven by the elastic-brittle damage constitutive model, thereby obtaining a multi-source mechanical response output result matching the random simulation scheme.

[0025] Furthermore, the calculation steps of the online calculation module are as follows:

[0026] Based on the above numerical simulation sample set, the attribute points of the sample set obtained in S1 are converted into a matrix by using a two-dimensional interpolation method;

[0027] With the attribute point matrix of elastic modulus E, Poisson's ratio μ, density ρ, and displacement boundary d as input, and the attribute point matrix of stress σ and damage D as output, the fully convolutional neural network U-Net is selected to construct the numerical simulation process proxy model;

[0028] Training the numerical simulation process proxy model: The loss function in the numerical simulation process proxy model uses SmoothL1Loss, the optimizer selects Adam, the initial value of the learning rate is 0.0002, and the learning rate decay strategy uses StepLR. After every 50 rounds of training, the learning rate is changed to 0.3 times the original for training. During the training process, the neural network needs to calculate the average loss value J and R2 of each training round for each training round. When R2>0.9 and the average loss function value J<0.1, the trained numerical simulation process proxy model is obtained.

[0029] Furthermore, the experimental steps of the mixed reality visualization and interaction module are as follows:

[0030] S31, virtual scene registration and virtual model generation;

[0031] S33, setting virtual experiment parameters;

[0032] S33. Observation of experimental results.

[0033] Furthermore, S31 specifically includes the following steps:

[0034] Create the 3D models required for physical experiments in a one-to-one ratio according to the actual size, including uniaxial compression testing machine, computer, work table and rock model, and import the established 3D models into Unity3D to build Unity virtual reality projects;

[0035] In the Unity virtual reality project, a virtual press is used in combination with the actual physical experimental process to create a visualization scene that matches the uniaxial compression experimental process. After the user sets the basic parameter information in the virtual computer, a virtual rock sample is automatically generated and placed on the virtual press. A button to start the virtual experiment is set in the virtual computer to control the start of the virtual experiment.

[0036] Further, S32 specifically includes the following steps:

[0037] Use Python to build a Socket server, integrate the numerical simulation process proxy model into the Socket server, and transform the Unity virtual reality project into a Socket client;

[0038] Start the virtual experiment. The virtual reality scene calls the numerical simulation process proxy model in the form of a socket, and the parameters passed are the elastic modulus E, Poisson's ratio μ, density ρ, and displacement boundary d, where the boundary load and is the product of the loading speed and the loading time step; set the delay function to ensure that the running time of each loading time step is 3s. After the proxy model is calculated, the mechanical response data such as stress field σ and damage D will be further returned to the virtual reality scene to reduce the elastic modulus. At the same time, by updating the loading time step, the loading displacement boundary conditions are modified, and the updated elastic modulus E, Poisson's ratio μ, density ρ, and displacement boundary d are passed to the numerical simulation process proxy model again in the form of a socket, thereby realizing the restoration of the physical experiment process.

[0039] Furthermore, S33 specifically includes the following steps:

[0040] HoloLens 2 glasses are selected as the mixed reality scene integration device. Based on Unity3D, the above Unity virtual reality project is converted into a mixed reality scene project based on HoloLens2 and deployed to HoloLens 2 glasses. Users can learn about experimental equipment and experimental processes by wearing glasses.

[0041] Compared with the prior art, the present invention has the following advantages:

[0042] At present, simulation systems all rely on a single static resource and cannot correctly reflect the real rock characteristics. Although numerical simulation can truly reflect the rock characteristics by setting parameters, the calculation time is long and the efficiency is too low, which is a major factor that prevents the combination of numerical simulation and virtual experiments. However, the present invention constructs a large data sample set representing parameters and results through offline calculation, while taking into account heterogeneity. On this basis, combined with deep learning and other technologies, it can realize the function of correctly reflecting real rocks in virtual experiments.

[0043] The input and output of numerical simulation are highly nonlinear, and it is difficult to find the relationship between the two, especially after considering the heterogeneity. This is a major factor that restricts the fit between virtual realization and reality. Second, the present invention uses a deep neural network to autonomously learn and capture the intrinsic connection between input and output. It has wide applicability for solving complex classification and regression problems and has extremely high computational efficiency. The advantages of numerical simulation in accurately expressing rock damage and fracture behavior and the advantages of deep learning methods in efficiently solving complex problems are combined to construct an online computing module based on the numerical simulation sample set and the U-Net network's numerical simulation process proxy model, so as to obtain the correct input-output relationship, so that there is no need to rely on a single static resource, making the experiment more random and authentic.

[0044] The present invention can use a mixed reality experiment program developed based on the HoloLens2 device to combine the virtual and the real to achieve a real experiment experience at a lower cost;

[0045] Previous virtual experiments were all virtual reality, which was not very interactive and was not conducive to stimulating interest. At the same time, they were all static resources, making it difficult for students to understand real experimental phenomena. The present invention uses mixed reality technology (MR), which brings users an immersive experience with its unique features such as virtual-real combination, tracking registration and intelligent interaction. By constructing a mixed reality laboratory scene, the visualization and interaction module allows students to gain the same interactive experience as using a real testing machine, which not only greatly enhances the fun of learning, but also allows learners to simulate practical operations and improve their practical skills. At the same time, the online calculation module is used to quickly calculate the corresponding numerical simulation model, which can dynamically and intuitively display the physical and mechanical behavior of rocks. Combining this with virtual reality technology can change the problems of the original static model and improve students' understanding of the physical and mechanical properties of rocks.

[0046] The virtual experiment system provided by the present invention can obtain different accurate experimental results according to different parameters, and at the same time obtain a real experimental operation experience, which can greatly save experimental costs and improve environmental protection; reduce experimental risks and improve safety; break the limitations of time and space and stimulate students' enthusiasm for independent learning. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0048] Figure 1 It is a schematic diagram of the process of the present invention.

[0049] Figure 2 It is a schematic diagram of the uniaxial compression test of the present invention.

[0050] Figure 3 This is the elastic-brittle damage constitutive model diagram of the present invention.

[0051] Figure 4 This is a communication relationship diagram between the random simulation scheme and the dynamic simulation service of the present invention.

[0052] Figure 5 This is a network diagram of the numerical simulation process agent model based on U-Net in the present invention.

[0053] Figure 6 Construct a graph for the virtual reality scene of the present invention.

[0054] Figure 7 Schematic diagram of communication between mixed reality scene and proxy model of the present invention.

[0055] Figure 8 Initialize the mixed reality scene graph for the present invention.

[0056] Fig. 9 This is a rendering of the virtual experiment process of the present invention. DETAILED DESCRIPTION

[0057] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0058] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0059] The present invention provides a rock uniaxial compression experiment virtual reality system, comprising:

[0060] Offline calculation module, used to generate numerical simulation sample sets based on random simulation scheme generation service and numerical simulation system driven by elastic-brittle damage constitutive model;

[0061] Online computing module, used to build and train the numerical simulation process proxy model based on the numerical simulation sample set and U-Net network;

[0062] The mixed reality visualization and interaction module is used to apply the trained numerical simulation process agent model to the mixed reality scene to complete the mixed reality experiment.

[0063] The offline calculation module includes the following:

[0064] At present, simulation systems all rely on a single static resource and cannot correctly reflect the real rock characteristics. Although numerical simulation can truly reflect the rock characteristics by setting parameters, the calculation time is long and the efficiency is too low, which is a major factor that prevents the combination of numerical simulation and virtual experiments. However, the present invention constructs a large data sample set representing parameters and results through offline calculation, while taking into account heterogeneity. On this basis, combined with deep learning and other technologies, it can realize the function of correctly reflecting real rocks in virtual experiments.

[0065] (1) Schematic diagram of uniaxial compression test is shown in the attached figure. Figure 2 As shown in the figure, it is believed that the damage and failure process of rock under uniaxial compression conditions presents elastic brittleness. Its physical and mechanical parameters include elastic modulus E, Poisson's ratio μ, density ρ, and boundary conditions include loading displacement d. By considering the heterogeneity of the physical and mechanical parameters of the rock mass, the heterogeneity coefficient n is introduced. Based on this, the range of experimental mechanical parameters shown in Table 1 is set. On this basis, the Latin hypercube sampling method is used to generate random simulation schemes, and Python coding is further used to form a random simulation scheme generation service.

[0066] Table 1 Experimental parameter range

[0067]

[0068] (2) The isotropic continuous damage formula proposed by Lemaitre and Chaboche is used to characterize the damage evolution law of the material, as shown in Equation 1-2. On this basis, the elastic-brittle damage constitutive model is used to describe the rock damage and failure process, as shown in the attached figure. Figure 3 As shown in the figure, for rock damage, the maximum tensile stress criterion is used first to judge its tensile failure; when it is not satisfied, the Mohr-Coulomb criterion is used to evaluate its shear failure. The above damage constitutive model is further embedded in COMSOL to construct a numerical simulation system driven by the elastic-brittle damage constitutive model.

[0069] ε ij =((1+υ)σ ij -vσ kk δ ij ) / E (1)

[0070] E=E0(1-D) (2)

[0071] Where ε ij is the damage elastic strain tensor, σ ij is the Cauchy stress tensor, E and E0 are the elastic moduli of the damaged and undamaged materials respectively, D is the damage variable, v is the Poisson's ratio, δ ij is the Kronecker symbol.

[0072] (3) Based on the Socket communication protocol, the random simulation scheme generation service based on Python coding is converted into a Socket client; using Java secondary development, the numerical simulation system driven by the elastic-brittle damage constitutive model based on COMSOL software is converted into a numerical simulation Socket server. Through the Socket communication protocol, the random simulation scheme is dynamically transmitted to the numerical simulation system driven by the elastic-brittle damage constitutive model, so as to obtain the multi-source mechanical response output results such as stress and damage field that match the random simulation scheme. The relationship between the two service communication methods and data transmission is shown in the attached figure. Figure 4 shown.

[0073] The online calculation module includes the following:

[0074] The input and output of numerical simulation are highly nonlinear, and it is difficult to find the relationship between the two, especially after considering the heterogeneity. This is a major factor that restricts the fit between virtual realization and reality. In recent years, deep neural networks have attracted much attention for their excellent data processing capabilities, especially in processing images, voices, text and other data. This method can autonomously learn and capture the intrinsic connection between input and output. It has wide applicability for solving complex classification and regression problems and has extremely high computational efficiency. Combining the advantages of numerical simulation in accurately expressing rock damage and fracture behavior with the advantages of deep learning methods in efficiently solving complex problems, an online computing module based on the numerical simulation sample set and the U-Net network is constructed to obtain the correct input-output relationship, so that there is no need to rely on a single static resource, making the experiment more random and realistic.

[0075] (1) Based on the above numerical simulation sample set, the attribute point sample sets obtained in S1 are converted into matrices by using a two-dimensional interpolation method.

[0076] (2) For the numerical simulation sample data such as elastic modulus E and Poisson's ratio μ, they are first standardized. Then, based on the fully convolutional neural network U-Net, the processed elastic modulus E, Poisson's ratio μ, density ρ, displacement boundary d matrix are used as input, and the mechanical response matrix such as stress σ and damage D is used as output to construct a numerical simulation process proxy model. The model network structure is attached. Figure 5 shown.

[0077] The mixed reality visualization and interaction module includes the following:

[0078] Previous virtual experiments were all virtual reality, which was not very interactive and was not conducive to stimulating interest. At the same time, they were all static resources, making it difficult for students to understand the real experimental phenomena. The rapid development of the metaverse has brought about the advancement of mixed reality technology (MR), which brings users an immersive experience with its unique features such as virtual-real combination, tracking registration and intelligent interaction. By constructing a mixed reality laboratory scene, the visualization and interaction module allows students to get the same interactive experience as using a real testing machine, which not only greatly enhances the fun of learning, but also allows learners to simulate practical operations and improve their practical skills. At the same time, the online calculation module is used to quickly calculate the corresponding numerical simulation model, which can dynamically and intuitively display the physical and mechanical behavior of rocks. Combining this with virtual reality technology can change the problems of the original static model and improve students' understanding of the physical and mechanical properties of rocks.

[0079] (1) Use Blender 4.0 to create the three-dimensional models required for physical experiments such as experimental rock specimens, uniaxial compression testing machines, and computers according to the actual size. Then import the established models into Unity 3D to construct the virtual reality scene, as shown in the attached figure. Figure 6 As shown;

[0080] (2) Combined with the actual physical experiment process, a virtual scene is created in the virtual computer to set the experimental parameters such as loading speed and rock type and control the experimental process, as shown in the attached figure. Figure 3 As shown; the virtual press is further combined to produce a visualization scene that matches the uniaxial compression test process; after the user sets the loading speed, basic rock parameters and other information in the virtual computer, a virtual rock sample will be automatically generated and automatically placed on the virtual press, and the elastic modulus E, Poisson's ratio μ, density ρ, stress σ, damage D and other information can be viewed; a button for starting the virtual experiment is further set in the virtual computer to control the start of the virtual experiment.

[0081] (3) Python is used to build a Socket server, and the above-mentioned U-Net proxy model is integrated into it. The above-mentioned scene is further transformed into a Socket client. When the user starts the virtual experiment, the virtual reality scene will call the U-Net proxy model in the form of Socket, and the parameters passed are elastic modulus E, Poisson's ratio μ, density ρ, and displacement boundary d, where the boundary load and is the product of the loading speed and the loading time step. Since the proxy model has a fast calculation speed, in order to better display the rock damage and fracture process, a delay function is set to ensure that the running time of each loading time step is 3s. After the proxy model is calculated, it will return mechanical response data such as stress σ and damage E to the virtual reality scene. The elastic modulus and other parameters are further reduced by combining formula (2). At the same time, the loading displacement boundary conditions are modified by updating the loading time step, and the updated elastic modulus E, Poisson's ratio μ, density ρ, and displacement boundary d are passed to the U-Net proxy model again in the form of Socket, thereby realizing the restoration of the physical experiment process. The iterative update process is attached Figure 7 shown.

[0082] Finally, HoloLens 2 glasses are selected as the mixed reality scene integration device. Based on the MixedReality Toolkit (MRTK) module of Unity3D, the above scene is converted into a mixed reality scene and integrated into HoloLens 2 glasses. Users can learn about the experimental equipment and experimental process by wearing glasses. The specific operation process is as follows:

[0083] (1) Start the mixed reality program in an open space. After environmental perception, it will automatically generate virtual models such as the uniaxial compression testing machine and experimental computer, as shown in the attached figure. Figure 8 As shown;

[0084] (2) The loading rate, elastic modulus, density, Poisson's ratio and mean value parameters are set through a virtual computer. The mixed reality system will automatically generate a rock sample that matches the setting and place it on the press.

[0085] (3) After that, the experiment can be started by clicking the [Start Experiment] button on the virtual computer. At this time, the mixed reality system will continuously interact with the numerical simulation process agent model based on U-net to realize the dynamic solution and visualization of the rock damage and fracture process, such as Fig. 9 shown.

[0086] (4) When wearing Hololens2 glasses, users can observe experimental phenomena at any position, understand the process of rock damage and destruction more intuitively, and significantly improve the quality of experimental teaching.

[0087] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A virtual reality system for rock uniaxial compression experiment, characterized in that: include: Offline calculation module, used to generate numerical simulation sample sets based on random simulation scheme generation service and numerical simulation system driven by elastic-brittle damage constitutive model; Online computing module, used to build and train the numerical simulation process proxy model based on the numerical simulation sample set and U-Net network; The mixed reality visualization and interaction module is used to apply the trained numerical simulation process agent model to the mixed reality scene to complete the mixed reality experiment.

2. The rock uniaxial compression experiment virtual reality system according to claim 1 is characterized in that: The calculation steps of the offline calculation module are as follows: S11, forming a random simulation scheme generation service; S12. Construct a numerical simulation system driven by elastic-brittle damage constitutive model; S13. Convert the random simulation scheme generation service into a Socket client, convert the numerical simulation system driven by the elastic-brittle damage constitutive model into a numerical simulation Socket server, and dynamically transfer the random simulation scheme to the numerical simulation system driven by the elastic-brittle damage constitutive model, so as to obtain the multi-source mechanical response output results matching the random simulation scheme, and integrate all the multi-source mechanical response output results to form a numerical simulation sample set.

3. The rock uniaxial compression experiment virtual reality system according to claim 2 is characterized in that: S11 specifically includes the following steps: Determine the experimental mechanical parameters of rock, including elastic modulus E, Poisson's ratio μ, density ρ, loading displacement d and heterogeneity coefficient n; Determine the range of experimental mechanical parameters; The Latin hypercube sampling method is used to sample within the range of experimental mechanical parameters to generate random simulation schemes, which are then coded in Python to form a random simulation scheme generation service.

4. The rock uniaxial compression experiment virtual reality system according to claim 2 is characterized in that: S12 specifically includes the following steps: The isotropic continuous damage formula is used to characterize the damage evolution law of the material; the isotropic continuous damage formula is as follows: e ij =((1+v)σ ij -vσ kk d ij ) / E (1) E=E0(1-D) (2) In the formula, ε ij is the damage elastic strain tensor, σ ij is the Cauchy stress tensor, E and E0 are the elastic moduli of the damaged and undamaged materials respectively, D is the damage variable, υ is the Poisson's ratio, δ ij is the Kronecker symbol; Among them, for rock damage and failure, the maximum tensile stress criterion is preferentially used to judge the tensile failure of the rock; when the maximum tensile stress criterion is not met, the Mohr-Coulomb criterion is used to evaluate the shear failure of the rock, and the elasto-brittle damage constitutive model is further embedded in COMSOL to construct a numerical simulation system driven by the elasto-brittle damage constitutive model.

5. The rock uniaxial compression experiment virtual reality system according to claim 2 is characterized in that: S13 specifically includes the following steps: based on the Socket communication protocol, converting the random simulation scheme generation service based on Python coding into a Socket client; using Java secondary development method, converting the numerical simulation system driven by the elastic-brittle damage constitutive model based on COMSOL software into a numerical simulation Socket server, and through the Socket communication protocol, dynamically transmitting the random simulation scheme into the numerical simulation system driven by the elastic-brittle damage constitutive model, so as to obtain the multi-source mechanical response output results matching the random simulation scheme.

6. The rock uniaxial compression experiment virtual reality system according to claim 1 is characterized in that: The calculation steps of the online calculation module are as follows: Based on the above numerical simulation sample set, the attribute points of the sample set obtained in S1 are converted into a matrix by using a two-dimensional interpolation method; With the attribute point matrix of elastic modulus E, Poisson's ratio μ, density ρ, and displacement boundary d as input, and the attribute point matrix of stress σ and damage D as output, the fully convolutional neural network U-Net is selected to construct the numerical simulation process proxy model; Training the numerical simulation process proxy model: The loss function in the numerical simulation process proxy model uses SmoothL1Loss, the optimizer selects Adam, the initial value of the learning rate is 0.0002, and the learning rate decay strategy uses StepLR. After every 50 rounds of training, the learning rate is changed to 0.3 times the original for training. During the training process, the neural network needs to calculate the average loss value J and R2 of each training round for each training round. When R2>0.9 and the average loss function value J<0.1, the trained numerical simulation process proxy model is obtained.

7. The rock uniaxial compression experiment virtual reality system according to claim 1 is characterized in that: The experimental steps of the mixed reality visualization and interaction module are as follows: S31, virtual scene registration and virtual model generation; S33, setting virtual experiment parameters; S33. Observation of experimental results.

8. The rock uniaxial compression experiment virtual reality system according to claim 7 is characterized in that: S31 specifically includes the following steps: Create the 3D models required for physical experiments in a one-to-one ratio according to the actual size, including uniaxial compression testing machine, computer, work table and rock model, and import the established 3D models into Unity3D to build Unity virtual reality projects; In the Unity virtual reality project, a virtual press is used in combination with the actual physical experimental process to create a visualization scene that matches the uniaxial compression experimental process. After the user sets the basic parameter information in the virtual computer, a virtual rock sample is automatically generated and placed on the virtual press. A button to start the virtual experiment is set in the virtual computer to control the start of the virtual experiment.

9. The rock uniaxial compression experiment virtual reality system according to claim 7 is characterized in that: S32 specifically includes the following steps: Use Python to build the Socket server, integrate the numerical simulation process proxy model into the Socket server, and transform the Unity virtual reality project into a Socket client; Start the virtual experiment. The virtual reality scene calls the numerical simulation process proxy model in the form of a socket, and the parameters passed are the elastic modulus E, Poisson's ratio μ, density ρ, and displacement boundary d, where the boundary load and is the product of the loading speed and the loading time step; set the delay function to ensure that the running time of each loading time step is 3s. After the proxy model is calculated, the mechanical response data such as stress field σ and damage D will be further returned to the virtual reality scene to reduce the elastic modulus. At the same time, by updating the loading time step, the loading displacement boundary conditions are modified, and the updated elastic modulus E, Poisson's ratio μ, density ρ, and displacement boundary d are passed to the numerical simulation process proxy model again in the form of a socket, thereby realizing the restoration of the physical experiment process.

10. The rock uniaxial compression experiment virtual reality system according to claim 7, characterized in that: S33 specifically includes the following steps: HoloLens 2 glasses are selected as the mixed reality scene integration device. Based on Unity3D, the above Unity virtual reality project is converted into a mixed reality scene project based on HoloLens2 and deployed to HoloLens 2 glasses. Users can learn about experimental equipment and experimental processes by wearing glasses.

Images