Intelligent instant calculation method and device for rock burst pressure, electronic device and storage medium

Abstract

The present disclosure discloses an intelligent instant calculation method and device, electronic device and storage medium of rock burst pressure, which relates to the field of data processing technology. The main technical methods include: obtaining an area to be calculated; the area to be calculated includes an area to be destroyed and an area not to be destroyed; dividing the area to be destroyed into at least two non-continuous solution areas, and integrating the non-to-be-destroyed area into a continuous solution area; respectively inputting the non-continuous solution area and the continuous solution area into a preset rock burst pressure intelligent instant calculation model to obtain the rock burst pressure of the area to be calculated; the preset rock burst pressure intelligent instant calculation model processes the non-continuous solution area and the continuous solution area in a parallel calculation manner. Compared with the related art, the embodiment of the present disclosure uses a preset rock burst pressure intelligent instant calculation model to process the non-continuous solution area and the continuous solution area in a parallel calculation manner, thereby improving the efficiency of rock burst pressure calculation.

Description

Technical Field

[0001] The present disclosure relates to the field of data processing technology, and in particular to an intelligent instant calculation method and device for rock burst, an electronic device and a storage medium. Background Art

[0002] Rock burst refers to the sudden release of rock pressure caused by geological structure, ore body stress changes or mining activities during coal mining. Rock burst is a serious safety issue in coal mining. In order to conduct a more in-depth study of rock burst, it is necessary to calculate the rock burst.

[0003] In the related technologies of intelligent instant calculation of rock burst, the coal mine area is usually decomposed into multiple independent units, and the independent units are calculated one by one. Since the coal mine area is large, the number of independent units obtained by decomposition is large, resulting in low efficiency of rock burst calculation. Summary of the invention

[0004] The present invention provides an intelligent instant calculation method and device for rock burst pressure, an electronic device and a storage medium, the main purpose of which is to solve the problem of low efficiency in rock burst pressure calculation.

[0005] According to a first aspect of the present disclosure, a method for intelligent instant calculation of rock burst is provided, which includes:

[0006] Acquire a region to be calculated; the region to be calculated includes a region to be destroyed and a region not to be destroyed;

[0007] Dividing the to-be-destroyed area into at least two non-continuous solution areas, and integrating the non-to-be-destroyed area into a continuous solution area;

[0008] The discontinuous solution area and the continuous solution area are respectively input into a preset rock burst intelligent calculation model to obtain the rock burst of the area to be calculated; the preset rock burst intelligent calculation model processes the discontinuous solution area and the continuous solution area in a parallel calculation manner.

[0009] Optionally, the training method of the preset rock burst intelligent instant calculation model includes:

[0010] Get training data;

[0011] Based on a deep neural network algorithm, the initial rock burst intelligent calculation model is trained using the training data to obtain the preset rock burst intelligent calculation model.

[0012] Optionally, the acquiring of training data includes:

[0013] Respectively obtaining a simulated rock burst pressure for training and a rock burst pressure for experimental testing for training, and obtaining loads corresponding to the simulated rock burst pressure for training and the rock burst pressure for experimental testing for training;

[0014] The training simulated rock impact pressure, the training experimental test rock impact pressure and the load are combined to obtain the training data.

[0015] Optionally, the deep neural network algorithm is used to train the initial rock burst intelligent calculation model using the training data to obtain the preset rock burst intelligent calculation model, which includes:

[0016] Obtaining model parameters of the initial rock burst intelligent instant calculation model;

[0017] The simulated rock burst pressure for training, the rock burst pressure for training experimental test and the load are studied to obtain a loss function; the loss function includes the model parameters;

[0018] The model parameters are adjusted in a manner that makes the loss function approach the minimum value to obtain the preset rock burst intelligent instant calculation model.

[0019] Optionally, the preset rock burst intelligent instant calculation model processes the discontinuous solution area and the continuous solution area in a parallel calculation manner, including:

[0020] Calling a central processing unit to decompose the non-continuous solution area and the continuous solution area into at least two tasks respectively;

[0021] At least two graphics processor sub-cores are called to respectively perform parallel calculations on the at least two tasks.

[0022] Optionally, the preset rock burst intelligent instant calculation model includes a preset continuous solution neural network and a preset discontinuous solution neural network; wherein,

[0023] The preset continuous solution neural network is used to process the continuous solution area;

[0024] The preset non-continuous solution neural network is used to process the non-continuous solution neural network.

[0025] According to a second aspect of the present disclosure, there is provided an intelligent instant calculation device for rock burst pressure, comprising:

[0026] A first acquisition unit is used to acquire a region to be calculated; the region to be calculated includes a region to be destroyed and a region not to be destroyed;

[0027] A segmentation unit, used for segmenting the to-be-destroyed area into at least two non-continuous solution areas;

[0028] An integration unit, used for integrating the non-to-be-destroyed area into a continuous solution area;

[0029] An input unit is used to input the discontinuous solution area and the continuous solution area into a preset rock burst intelligent calculation model to obtain the rock burst of the area to be calculated; the preset rock burst intelligent calculation model processes the discontinuous solution area and the continuous solution area in a parallel calculation manner.

[0030] Optionally, the device further comprises:

[0031] A second acquisition unit, used to acquire training data;

[0032] The training unit is used to train the initial rock burst intelligent calculation model in seconds using the training data based on a deep neural network algorithm to obtain the preset rock burst intelligent calculation model in seconds.

[0033] Optionally, the second acquiring unit includes:

[0034] A first acquisition module is used to respectively acquire a simulated rock burst pressure for training and a rock burst pressure for experimental testing for training, and to acquire loads corresponding to the simulated rock burst pressure for training and the rock burst pressure for experimental testing for training;

[0035] The combination module is used to combine the training simulated rock impact pressure, the training experimental test rock impact pressure and the load to obtain the training data.

[0036] Optionally, the training unit includes:

[0037] A second acquisition module is used to acquire model parameters of the initial rock burst intelligent instant calculation model;

[0038] A learning module, used for learning the training simulated rock burst pressure, the training experimental test rock burst pressure and the load to obtain a loss function; the loss function includes the model parameters;

[0039] The adjustment module is used to adjust the model parameters in a manner that makes the loss function approach a minimum value, so as to obtain the preset rock burst intelligent instant calculation model.

[0040] Optionally, the device further comprises:

[0041] A decomposition unit, used for calling a central processing unit to decompose the non-continuous solution area and the continuous solution area into at least two tasks respectively;

[0042] The computing unit is used to call at least two graphics processor sub-cores to perform parallel computing on the at least two tasks respectively.

[0043] Optionally, the preset rock burst intelligent instant calculation model includes a preset continuous solution neural network and a preset discontinuous solution neural network; wherein,

[0044] The preset continuous solution neural network is used to process the continuous solution area;

[0045] The preset non-continuous solution neural network is used to process the non-continuous solution neural network.

[0046] According to a third aspect of the present disclosure, there is provided an electronic device, including:

[0047] at least one processor; and

[0048] a memory communicatively connected to the at least one processor; wherein,

[0049] The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method described in the first aspect.

[0050] According to a fourth aspect of the present disclosure, a non-transitory computer-readable storage medium storing computer instructions is provided, wherein the computer instructions are used to enable the computer to execute the method described in the first aspect.

[0051] According to a fifth aspect of the present disclosure, a computer program product is provided, comprising a computer program, wherein when the computer program is executed by a processor, the computer program implements the method as described in the first aspect above.

[0052] The intelligent instant calculation method and device, electronic device and storage medium of rock burst provided by the present disclosure obtain an area to be calculated; the area to be calculated includes an area to be destroyed and an area not to be destroyed; the area to be destroyed is divided into at least two non-continuous solution areas, and the area not to be destroyed is integrated into a continuous solution area; the non-continuous solution area and the continuous solution area are respectively input into a preset rock burst intelligent instant calculation model to obtain the rock burst in the area to be calculated; the preset rock burst intelligent instant calculation model processes the non-continuous solution area and the continuous solution area in a parallel calculation manner. Compared with the related art, the embodiment of the present disclosure improves the efficiency of rock burst calculation by dividing the area to be destroyed in the area to be calculated into at least two non-continuous solution areas, and integrating the non-to-be-destroyed area in the area to be calculated into a continuous solution area, and using the preset rock burst intelligent instant calculation model to process the non-continuous solution area and the continuous solution area in a parallel calculation manner.

[0053] It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present application, nor is it intended to limit the scope of the present application. Other features of the present application will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] The accompanying drawings are used to better understand the present solution and do not constitute a limitation of the present disclosure.

[0055] Figure 1 A schematic diagram of a flow chart of an intelligent instant calculation method for rock burst pressure provided by an embodiment of the present disclosure;

[0056] Figure 2 A flowchart of a training method for a preset rock burst intelligent instant calculation model provided by an embodiment of the present disclosure;

[0057] Figure 3 A technical roadmap for a theoretical framework for rapidly solving continuous-discontinuous mechanical problems provided by the embodiments of the present disclosure;

[0058] Figure 4 A diagram showing a continuous-discontinuous region adaptive conversion provided by an embodiment of the present disclosure;

[0059] Figure 5 A bucket search algorithm and contact detection optimization effect diagram provided by an embodiment of the present disclosure;

[0060] Figure 6 A schematic diagram of least squares reconstruction provided by an embodiment of the present disclosure;

[0061] Figure 7 A diagram showing a stress reconstruction algorithm provided by an embodiment of the present disclosure;

[0062] Figure 8 A technical roadmap for accelerating the solution of rock mass elastic-plastic control equations by deep learning provided in the embodiments of the present disclosure;

[0063] Fig. 9 A process diagram for integrating data into a deep learning network framework provided by an embodiment of the present disclosure;

[0064] Fig.10 A research roadmap for constructing a deep learning framework based on rock mass elastic-plastic control equations provided in the embodiments of the present disclosure;

[0065] Fig.11 A structural diagram of a deep learning network for rock mass elastic-plastic control equations provided in an embodiment of the present disclosure;

[0066] Fig.12 A flowchart of training and optimizing a deep learning model provided in an embodiment of the present disclosure;

[0067] Fig.13 A design method diagram of a finite element hybrid hierarchical parallel computing provided by an embodiment of the present disclosure;

[0068] Fig.14 A flow chart of a sparse row matrix storage principle and parallel direct solution provided by an embodiment of the present disclosure;

[0069] Fig.15 A unit partition diagram provided by an embodiment of the present disclosure;

[0070] Fig.16 A flow chart of a material point mixing parallel optimization provided by an embodiment of the present disclosure;

[0071] Fig.17 A technical roadmap for solving a large model of coal mine rock burst calculation in seconds provided by the embodiment of the present disclosure;

[0072] Fig.18 A schematic diagram of the structure of an intelligent instant calculation device for rock burst pressure provided by an embodiment of the present disclosure;

[0073] Fig.19 A schematic diagram of the structure of another intelligent instant calculation device for rock burst provided by an embodiment of the present disclosure;

[0074] Fig. 20 A schematic block diagram of an exemplary electronic device provided for an embodiment of the present disclosure. DETAILED DESCRIPTION

[0075] The following is a description of exemplary embodiments of the present disclosure in conjunction with the accompanying drawings, including various details of the embodiments of the present disclosure to facilitate understanding, which should be considered as merely exemplary. Therefore, it should be recognized by those of ordinary skill in the art that various changes and modifications may be made to the embodiments described herein without departing from the scope and spirit of the present disclosure. Similarly, for the sake of clarity and conciseness, descriptions of well-known functions and structures are omitted in the following description.

[0076] The following describes the intelligent method and device for calculating rock burst pressure in seconds, electronic equipment, and storage medium according to an embodiment of the present disclosure with reference to the accompanying drawings.

[0077] Figure 1 A flow chart of an intelligent instant calculation method for rock burst pressure provided in an embodiment of the present disclosure.

[0078] like Figure 1 As shown, the method is applied in a server, and the method comprises the following steps:

[0079] Step 101, obtaining a region to be calculated; the region to be calculated includes a region to be destroyed and a region not to be destroyed.

[0080] The area to be calculated is the entire coal mine area where rock burst calculation is required. The area to be calculated includes the influence of various geological conditions, ore body structure and mining activities, and is the basis for analyzing rock burst risk. The area to be destroyed is the area in the area to be calculated where rock destruction or pressure release is required for actual mining operations. It is usually the area where the stress state may reach a critical point under the influence of mining activities. The non-area to be destroyed is the area in the area to be destroyed where rock destruction or pressure release is not performed.

[0081] By conducting geological surveys on actual coal mining areas, information such as the geological structure, rock distribution, lithology, faults and cracks of the coal mines can be collected. Monitoring equipment such as stress gauges, displacement meters, acoustic emission monitoring systems and other equipment can be used to collect information such as the geological structure, rock distribution, lithology, faults and cracks of the coal mines. According to the collected information such as the geological structure, rock distribution, lithology, faults and cracks of the coal mines, numerical simulation software is used to build a three-dimensional model of the coal mine area to obtain the area to be calculated. The numerical simulation software can be Fast Lagrangian Analysi s of Continuum (FLAC), but it should be clear that this statement is not intended to limit the numerical simulation software to FLAC, it can be other software.

[0082] By subdividing the area to be calculated into the area to be damaged and the area not to be damaged, the rock burst analysis can be carried out more specifically. The area to be damaged is often more susceptible to factors such as geological stress and mining activities, resulting in disasters such as rock burst. Therefore, it is very necessary to focus on analyzing and calculating these areas.

[0083] Step 102: divide the to-be-destroyed area into at least two non-continuous solution areas, and integrate the non-to-be-destroyed area into a continuous solution area.

[0084] The area to be destroyed is divided into at least two discontinuous solution areas. There may be obvious discontinuous surfaces such as faults and cracks between the discontinuous solution areas, or the stress state is complex and difficult to describe using the continuous medium algorithm. The discontinuous solution area needs to be calculated more finely separately to improve the calculation accuracy.

[0085] Integrate the non-destructive area into one or more continuous solution areas. The continuous solution areas are physically continuous, or the interactions between them can be simplified to continuous medium algorithms. This can reduce the number of calculation units and improve calculation efficiency.

[0086] The non-continuous solution region is formed by dividing the area to be destroyed into multiple grid units, and then selectively merging or separating these units; or the area to be destroyed is divided based on features such as geological faults or cracks to obtain non-continuous solution regions, so as to ensure that there is a clear separation between the non-continuous solution regions. However, it should be clear that this statement is not intended to limit the method of dividing the area to be destroyed into at least two non-continuous solution regions to only the above two methods, and the area to be destroyed can also be divided into at least two non-continuous solution regions by other methods.

[0087] By accurately identifying the non-to-be-destroyed area from the area to be calculated, this can be achieved through methods such as image segmentation, edge detection, threshold processing, or classification based on specific features. After identifying the non-to-be-destroyed area, an image labeling technique (such as connected component labeling) can be used to assign a unique identifier to each independent non-to-be-destroyed area. Check the spatial relationship between the identifiers to determine which non-to-be-destroyed areas of the identifier can be merged to form a larger continuous solution area. The basis for merging can be the adjacency, similarity (such as grayscale value, texture, etc.) between the regions, or other custom criteria. If two or more non-to-be-destroyed areas are physically adjacent and meet the merging criteria, they can be merged into a larger area. During the merging process, some geometric processing may be required to ensure that the merged area is continuous and has no internal holes. Geometric processing includes operations such as filling small holes, smoothing boundaries, and removing noise. Geometric processing can be achieved through morphological operations (such as dilation, erosion, opening operations, closing operations, etc.). After the merging is completed, the continuous solution area needs to be verified to ensure that the merged continuous solution area meets expectations. This can be done through visual inspection, calculation of area properties (such as area, perimeter, shape factor, etc.). However, it should be clear that this statement is not intended to limit the method of integrating the non-to-be-destroyed area into the continuous solution area to only the above method, and it can also be achieved by other methods.

[0088] Step 103, respectively input the discontinuous solution area and the continuous solution area into a preset rock burst intelligent instant calculation model to obtain the rock burst of the area to be calculated; the preset rock burst intelligent instant calculation model processes the discontinuous solution area and the continuous solution area in a parallel calculation manner.

[0089] The preset rock burst intelligent calculation model is a pre-set model for calculating rock burst. The preset rock burst intelligent calculation model can calculate rock burst through numerical methods (such as finite element method, discrete element method, etc.) based on the input non-continuous solution area and continuous solution area. Rock burst refers to a quantitative indicator that reflects the rock burst risk in the coal mine area, which is calculated by the preset rock burst intelligent calculation model. Rock burst includes but is not limited to stress, displacement, strength failure probability, and impact range.

[0090] The preset rock burst intelligent instant calculation model uses parallel computing to process continuous solution areas and discontinuous solution areas. For continuous solution areas, numerical methods such as finite element method, finite difference method or boundary element method can be used to solve physical quantities such as stress, strain and displacement. For discontinuous solution areas, special methods such as discrete element method, extended finite element method or interface element method can be used to deal with discontinuities and interface effects. Among them, parallel computing allows computing tasks to be distributed to multiple processors or computing nodes for simultaneous execution, thereby significantly reducing the overall computing time.

[0091] Parallel computing is used for continuous and discontinuous solution areas, and different numerical methods are combined for processing, aiming to improve computing efficiency, enhance the adaptability and accuracy of the model, and better simulate complex physical processes under actual geological and mining conditions.

[0092] The intelligent instant calculation method of rock burst provided by the present disclosure obtains an area to be calculated; the area to be calculated includes an area to be destroyed and an area not to be destroyed; the area to be destroyed is divided into at least two non-continuous solution areas, and the area not to be destroyed is integrated into a continuous solution area; the non-continuous solution area and the continuous solution area are respectively input into a preset rock burst intelligent instant calculation model to obtain the rock burst in the area to be calculated; the preset rock burst intelligent instant calculation model processes the non-continuous solution area and the continuous solution area in a parallel calculation manner. Compared with the related art, the embodiment of the present disclosure improves the efficiency of rock burst calculation by dividing the area to be destroyed in the area to be calculated into at least two non-continuous solution areas, and integrating the non-to-be-destroyed area in the area to be calculated into a continuous solution area, and using the preset rock burst intelligent instant calculation model to process the non-continuous solution area and the continuous solution area in a parallel calculation manner.

[0093] In order to improve the accuracy of the preset rock burst intelligent second calculation model in predicting rock burst, it is necessary to train the preset rock burst intelligent second calculation model. The training method of the preset rock burst intelligent second calculation model can be implemented in the following ways, but is not limited to: Figure 2 As shown, Figure 2A flowchart of a training method for a preset rock burst intelligent instant calculation model provided by an embodiment of the present disclosure includes:

[0094] Step 201, obtaining training data.

[0095] The training data is an input data set for the preset rock burst intelligent calculation model. The training data includes but is not limited to the simulated values ​​of rock burst, experimental test values ​​and corresponding loads. The training data is used to teach the preset rock burst intelligent calculation model how to predict rock burst from loads.

[0096] The training data needs to be representative to a certain extent, so that the trained preset rock burst intelligent calculation model can have good generalization ability, that is, it can make accurate predictions for different situations in actual applications.

[0097] Step 202: Based on a deep neural network algorithm, the initial rock burst intelligent calculation model is trained using the training data to obtain the preset rock burst intelligent calculation model.

[0098] The deep neural network algorithm is a complex machine learning algorithm that processes information by simulating the structure and function of the human brain's neural network. The deep neural network contains multiple layers, each layer contains multiple neurons, and extracts data features and makes predictions through nonlinear transformations and weight connections. In the training of the rock burst model, the deep neural network is used to learn the complex mapping relationship from input parameters (such as load, geological conditions, etc.) to output parameters (such as rock burst intensity).

[0099] The initial rock burst intelligent calculation model is the model version at the beginning of deep neural network algorithm training, with preliminary parameter settings and structure. The initial rock burst intelligent calculation model needs to be optimized through the training process so that it can simulate and predict rock burst more accurately.

[0100] The training data is organized into a format suitable for input into a deep neural network, and a deep neural network architecture is constructed, including an input layer, a hidden layer, and an output layer. Appropriate activation functions and loss functions are selected, and the deep neural network is trained using the training data. The network parameters are adjusted to minimize the loss function, and a preset rock burst intelligent instant calculation model is obtained.

[0101] Through a large amount of training data, the deep neural network can learn the complex nonlinear relationship between input features and rock burst pressure. This learning ability enables the model to predict rock burst pressure more accurately, thereby improving the accuracy of the prediction.

[0102] As a refinement of step 201, when executing the acquisition of training data, it can be implemented in the following manner but not limited to, respectively acquiring a training simulated impact ground pressure and a training experimental test impact ground pressure, and acquiring a load corresponding to the training simulated impact ground pressure and the training experimental test impact ground pressure; combining the training simulated impact ground pressure, the training experimental test impact ground pressure and the load to obtain the training data.

[0103] Exemplarily, the simulated rock burst for training is a rock burst-related value obtained by calculating the training coal mine area through numerical simulation methods (such as finite element analysis, discrete element method, etc.). The training rock burst is a value calculated by computer simulation under assumed or known geological and mining conditions, and is used to simulate the occurrence and evolution of rock burst under actual conditions. The experimental test rock burst for training is a rock burst-related value obtained through experimental testing. The experimental test rock burst for training is usually obtained through actual measurement in the laboratory or on-site, reflecting the actual situation of rock burst under specific geological and mining conditions. The experimental test value is an important basis for verifying and calibrating the results of numerical simulation. Load is the external condition or force that causes the rock burst to occur.

[0104] By combining the simulation values, experimental values ​​and load data in the same format, a data matrix, i.e., training data, is formed. However, it should be clear that this statement is not intended to limit the combination method to the above method, and other methods can also be used for combination.

[0105] By combining different types of data, the quality and efficiency of deep neural network training can be improved, enabling it to more accurately predict rock burst and provide strong support for mine safety management and disaster prevention.

[0106] As a refinement of step 202, when executing the deep neural network algorithm, using the training data to train the initial rock burst intelligent second calculation model to obtain the preset rock burst intelligent second calculation model, it can be implemented in the following ways but not limited to: obtaining the model parameters of the initial rock burst intelligent second calculation model; learning the training simulated rock burst, the training experimental test rock burst and the load to obtain a loss function; the loss function includes the model parameters; adjusting the model parameters in a way that makes the loss function approach the minimum value to obtain the preset rock burst intelligent second calculation model. By optimizing the model parameters, it can be ensured that the preset rock burst intelligent second calculation model is better adapted to the actual data, thereby improving the accuracy and reliability of the prediction.

[0107] As a refinement of step 103, when executing the preset rock burst intelligent instant calculation model to process the discontinuous solution area and the continuous solution area in a parallel computing manner, it can be implemented in but not limited to the following ways: calling the central processing unit (CPU) to decompose the discontinuous solution area and the continuous solution area into at least two tasks respectively; calling at least two graphics processing units (GPU) sub-cores to perform parallel computing on the at least two tasks respectively. By decomposing a large problem into multiple small tasks by the central processing unit and assigning them to multiple graphics processor sub-cores for simultaneous processing, the overall computing time can be significantly reduced. Due to its powerful parallel processing capability, the graphics processor can process large amounts of data at the same time, thereby accelerating the computing process. Parallel computing can make more effective use of computing resources and improve overall computing efficiency.

[0108] As a refinement of the above-mentioned embodiment, the preset rock burst intelligent instant calculation model includes a preset continuous solution neural network and a preset discontinuous solution neural network; wherein, the preset continuous solution neural network is used to process the continuous solution area; the preset discontinuous solution neural network is used to process the discontinuous solution neural network. By customizing specialized neural networks for areas with different characteristics, the data features of each area can be more effectively utilized, thereby improving the accuracy of the overall prediction or simulation. In the face of complex and changeable rock burst problems, the use of a combination of multiple neural networks can enhance the robustness of the system. Even if the prediction results of a certain area are affected by interference or errors, the prediction results of other areas can still maintain a certain degree of accuracy.

[0109] In one possible implementation of the present disclosure, in order to facilitate the processing of the continuous solution area and the non-continuous solution area in a parallel computing manner, a better understanding is provided, such as Figure 3 As shown, Figure 3 A technical roadmap for a theoretical framework for quickly solving continuous-discontinuous mechanics problems provided in the embodiments of the present invention selects nonlinear finite element method and material point method as representative numerical methods of continuum medium mechanics and discontinuum mechanics, and analyzes the generation of potential damage areas before rock burst occurs based on the continuum medium mechanics method. In view of the characteristics of discrete and ill-conditioned numerical problems existing in rock burst engineering problems, deformation reinforcement theory is introduced to ensure stable convergence of nonlinear finite element solutions.

[0110] Analyze typical rock burst problems, form test cases that are representative of engineering, use continuum mechanics and discontinuum mechanics methods to carry out numerical simulations on these cases, comprehensively compare the bandwidth occupancy, symmetric / asymmetric matrix storage, and pathological degree of continuum mechanics and discontinuum mechanics methods in different cases, and establish an adaptive conversion evaluation function of continuum mechanics and discontinuum mechanics methods with unbalanced force as the core. In order to better understand the adaptive conversion of continuous-discontinuum regions, such as Figure 4 As shown, Figure 4 A display diagram of a continuous-discontinuous region adaptive conversion is provided in an embodiment of the present disclosure. A continuous-discontinuous region advantage algorithm selection strategy is constructed, and a corresponding program is written to realize the adaptive selection of the continuous-discontinuous region advantage algorithm for the entire process of rock burst from damage generation, destruction expansion to dynamic instability.

[0111] Based on the established adaptive selection strategy of the advantage algorithm, data storage architectures that can adapt to the global search algorithm are established for the patches in the continuous area and the particles in the discontinuous area. Combining classic algorithms such as bucket search, position code, and linear octree, an efficient global search algorithm suitable for the new data storage architecture is constructed. In order to better understand the bucket search algorithm and contact detection optimization, such as Figure 5 As shown, Figure 5 A bucket search algorithm and contact detection optimization effect diagram provided in an embodiment of the present disclosure realizes efficient search, matching and contact detection of "continuous region patch-non-continuous region particle".

[0112] Based on the moving least squares reconstruction method, in order to better understand the principle of least squares reconstruction, such as Figure 6 As shown, Figure 6 The schematic diagram of the least squares reconstruction provided by the embodiment of the present disclosure is a singularity study on the least squares matrix, and a highly robust physical quantity reconstruction algorithm from discrete points to continuous space is established. Based on the hybrid integration scheme, the cause of continuous-discontinuous body cross-grid noise is analyzed, and a method for eliminating continuous-discontinuous cross-grid noise is proposed. A stress reconstruction algorithm with a high-order convergence rate is established. In order to facilitate a better understanding of the stress reconstruction algorithm, as shown in FIG. Figure 7 As shown, Figure 7 A diagram showing a stress reconstruction algorithm provided in an embodiment of the present disclosure, which realizes high-performance exchange of physical quantities at a continuous-discontinuous interface.

[0113] The embodiments of the present disclosure accelerate the solution of rock mass elastic-plastic control equations through deep learning. In order to facilitate a better understanding of deep learning accelerating the solution of rock mass elastic-plastic control equations, as shown in FIG. Figure 8 As shown, Figure 8A technical roadmap for accelerating the solution of rock mass elastic-plastic control equations through deep learning provided in the embodiments of the present disclosure, organizes the theoretical model of rock mass elastic-plasticity, derives the two-dimensional and three-dimensional rock mass elastic-plastic control equations, organizes the research progress of computational mechanics in deep learning, analyzes and masters the international mainstream deep learning technologies and frameworks, explores the frontier theories and directions of deep learning models, and determines the input and output parameters required for the deep learning framework type. A detailed comparison of the numerical simulation and experimental operation of rock burst disasters is made, and a high-fidelity data set of the entire process of rock mass elastic-plastic response is constructed based on the rock mass elastic-plastic data obtained by continuous-discontinuous calculations in traditional numerical methods and laboratory uniaxial, triaxial and other experimental data. A suitable interpolation and sampling algorithm is selected to generate a low-fidelity data set, and the information generated by the load module is embedded in the data set to form spatiotemporal information together, and used as training data and test data for deep learning neural networks. In order to facilitate a better understanding of the integration of applied loads and created data in the deep learning network framework, such as Fig. 9 As shown, Fig. 9 A process diagram for integrating data into a deep learning network framework provided in an embodiment of the present disclosure is provided. A deep learning neural network model development environment integrating physical knowledge is established, a graphics computing acceleration module is introduced, and the neural network infrastructure required in the rock mass elastic-plastic model is reconstructed under the selected deep learning open source framework, specifically including the number of neural hidden layers and the number of neurons in each layer, exploring suitable basic neural network results, and finding the optimal initialization values ​​of the neural network hyperparameters; based on the reconstructed neural network, the rock mass elastic-plastic control equation is embedded, and the loss function is designed for the initial state of the rock mass, the Dirichlet boundary, the Neumann boundary and the elastic-plastic control equations at different stages; the gradient descent algorithm is used to realize the back propagation of the neural network, and the hyperparameters in the neural network are updated in the process of minimizing the loss function; the training and prediction modules are designed, and a visualization post-processing program is designed for the output calculation structure, in order to facilitate the research route of building a deep learning framework based on the rock mass elastic-plastic control equation to continue to twist and better understand, such as Fig.10 As shown, Fig.10 A research roadmap for constructing a deep learning framework based on rock mass elastic-plastic control equations is provided in the embodiments of the present disclosure. In order to facilitate a better understanding of the deep learning network of rock mass elastic-plastic control equations, as shown in FIG. Fig.11 As shown, Fig.11 A structural diagram of a deep learning network for rock mass elastic-plastic control equations provided in an embodiment of the present disclosure.

[0114] Research the technology of accelerating the solution of deep learning neural networks that integrate physical knowledge. First, pre-process the generated high-fidelity data and low-fidelity data, use standardization and normalization methods to solve the data scale and data dimension problems, prevent the model from overfitting, and construct appropriate tensor and scalar matrices as the input of the deep learning model; secondly, compare the performance of the deep learning neural network structure and neural operators on the fitting function and the impact of different activation functions on the training results, and select the appropriate neural network structure and activation function; then, design a step-by-step training and pre-training strategy for the neural network model, save the results of the deep neural network calculation graph after each training, save the time of retraining, and facilitate subsequent model training and adjustment optimization; the next step is to study the gradient descent algorithm optimizer In the specific performance of the model training process, the influence of mainstream optimizers on the convergence speed of the loss function at different learning rates is analyzed, and a learning rate adjustment strategy suitable for this study is designed; at the same time, TensorBoard is used to monitor the deep learning neural network model in real time, to achieve visual management of the data generated during the training process, and to adjust the model in time according to the visualization results; finally, based on the test data set, predictions are made on the trained deep learning model, and the continuous-discontinuous calculation results, laboratory experimental results and results based on the deep learning model of the rock mass elastic-plastic control equation are compared to evaluate the time complexity, space complexity and calculation accuracy of the deep learning model of the rock mass elastic-plastic control equation. In order to better understand the training and optimization of the deep learning model, Fig.12 As shown, Fig.12 A flowchart of the training and optimization of a deep learning model provided in an embodiment of the present disclosure. The trained model can be embedded in the overall computing model.

[0115] On the basis of the traditional regional decomposition method, after reading the model information required for calculation, the model grid needs to be divided into N sub-regions, and the finite element model information file (unit stiffness matrix information, unit mass matrix information, external load vector) and partition information file of each sub-region are generated. It can ensure that the grid volume of each sub-region is basically the same and the grid of each sub-region is continuous in space. Considering the hardware architecture of heterogeneous multi-core supercomputers, the design of finite element hybrid hierarchical parallel computing method is completed. In order to facilitate a better understanding of the design of finite element hybrid hierarchical parallel computing method, as shown in the following figure: Fig.13 As shown, Fig.13A design method diagram of finite element hybrid hierarchical parallel computing provided by the embodiment of the present disclosure. Multiple message passing interface (MPI) processes are started synchronously at the node end, a certain number of MPI processes are responsible for reading the data of a sub-region data file, and one MPI process derives multiple threads, each thread is loaded into the heterogeneous end, and the subtasks are decomposed and distributed to the graphics processor computing core in the heterogeneous end of the node, and multiple graphics processor computing cores are solved in parallel. Distributed storage is introduced to store data in a memory space shared by the heterogeneous group end, and MPI is used to asynchronously execute the same or different codes through multiple processes to achieve parallelism, and synchronization between processes is achieved through roadblocks, communication, etc., to improve the memory access efficiency of data. In order to facilitate a better understanding of finite element hybrid hierarchical parallel computing, in the finite element solution process, the overall steel matrix often has a large number of zero elements, and storing it according to the overall matrix will occupy a huge memory. Therefore, sparse rows (Compressed Sparse Row, CSR) are used to store non-zero elements in the overall stiffness matrix. According to the CSR storage rule, the non-zero elements, the corresponding column index and the index at the beginning of each row are stored in three arrays respectively, so as to greatly reduce the memory usage. Select a suitable parallel direct solver to calculate the linear equations stored in the sparse matrix. In order to better understand the sparse row matrix storage principle and the parallel direct solution process, such as Fig.14 As shown, Fig.14 A flow chart of a sparse row matrix storage principle and parallel direct solution provided in an embodiment of the present disclosure. Then, the load balancing design of the hybrid hierarchical parallel computing method is studied, and the parallel code is analyzed to clarify the system's operating environment, application type, traffic pattern, user distribution and peak processing requirements, and identify key performance indicators such as response time, system throughput, etc. Based on these analyses, appropriate load balancing strategies are selected, such as polling, minimum number of connections, etc., to formulate the best load balancing solution. Based on the above content, a parallel program for finite element solution of large rock burst models is developed, and multiple large models are selected for comparison of parallel and serial calculations to verify the effect of parallel optimization.

[0116] The material point method background grid is divided into regions, and the cells without common nodes are classified into one category. The background grid node data is updated in sequence according to the cell classification. Only one type of cell data is updated at the same time to avoid data competition. Taking two dimensions as an example, Fig.15 As shown, Fig.15 A unit partition diagram provided by the embodiment of the present disclosure. Units with the same number are classified into one category, called a zone, which is divided into 9 zones in total. The distance between units in the same zone is at least two unit side lengths, thereby ensuring that there are no common nodes between units of the same category, and when the background grid node data is updated, there will be no data competition.

[0117] The material point method is solved according to the explicit time integration method. First, the model, parameter and other data are read and initialized on the central processing unit. If the amount of data is large, MPI can be used for parallel optimization. Multiple processes / threads are created to read and process multiple data files respectively, and then the data tasks are split into multiple threads for processing. Each thread passes the data to the graphics processor for the next step of calculation; the graphics processor Fig.15 The particle data of n partitions are distributed to n processing cores, and each processing core creates multiple threads to calculate the shape function and derivative of each particle respectively, and then the calculation results are distributedly stored using the MPI shared information interface; then multiple processes / threads are created through MPI to calculate the node variables according to the particle variables, and the data tasks are also split into multiple threads for processing. Each thread passes the data to the graphics processor, and the loop decomposition method is introduced to parallel optimize the material point interaction and state update to avoid serious data competition in the update stage. The state update mainly includes the following steps: applying boundary conditions and updating node variables, updating the particle position and velocity, and updating the particle variables, and using the MPI shared information interface for distributed storage; the above steps are repeated in the cyclic time step until the end. In order to better understand the material point hybrid parallel optimization process, as shown in Fig.16 As shown, Fig.16 A flow chart of hybrid parallel optimization of material points provided in the embodiment of the present disclosure. Finally, a parallel program for solving material points of a large rock burst model based on the hybrid parallel optimization method is developed, and a large rock burst model solving program is finally developed by combining the finite element parallel program. Finally, a large model is selected for parallel effect verification.

[0118] For different operating systems and service platforms such as Windows, Linux, and Unix, one is to use cross-platform development frameworks, such as React Native, Flutter, Ionic, etc., to develop applications for multiple platforms by using a unified code base; the second is to use Web technologies, such as HTML, CSS, and JavaScript, to build cross-platform applications on different platforms, and use the functions of HTML5 and CSS3 to achieve rich user interfaces and functions. At the same time, JavaScript is used to implement the logic and interaction of the application. Evaluate and select the most suitable cross-platform development framework. When conducting cross-platform development, fully test the compatibility and performance of the software on different platforms, make necessary optimizations and adjustments, and finally formulate a cross-platform deployment strategy for the rock burst calculation software to optimize the performance of the software on different platforms. Finally, carry out relevant applications in the actual working conditions of production mines and newly built mines. In order to better understand the technical route for solving the large model of coal mine rock burst calculation in seconds, such as Fig.17 , Fig.17A technical roadmap for solving a large-scale coal mine rock burst calculation model in seconds provided in an embodiment of the present disclosure.

[0119] For the first time, the research concept of realizing "second-level" calculation for large-scale numerical calculation models of rock burst in coal mines was proposed, which promoted the precise digital simulation technology of rock burst and coal mining in my country into the era of "real-time" precise simulation; a coupled nonlinear finite element material point simulation method was proposed, and deformation reinforcement theory and unbalanced force were integrated to maximize the calculation speed while ensuring the simulation accuracy of the continuous-discontinuous process of disasters; a deep learning solver based on the rock mass elastic-plastic control equation was developed for the first time, and it was proposed to use the continuous-discontinuous rock burst calculation data as the input data of the deep learning model solver; for the coupled nonlinear finite element material point method, a large-scale high-performance computing technology integrating CPU-GPU hybrid parallel scheme was proposed. A CPU-GPU fusion parallel technology for continuous-discontinuous simulation of large-scale numerical models of rock burst was established, and the parallel calculation time was reduced by no less than 40%, and the parallel link acceleration ratio was no less than 5. For large numerical calculation models with more than 10 million units, the calculation time is no more than 100 seconds.

[0120] In summary, the embodiments of the present disclosure can achieve the following effects:

[0121] The disclosed embodiment divides the to-be-destroyed area in the area to be calculated into at least two non-continuous solution areas, and integrates the non-to-be-destroyed area in the area to be calculated into a continuous solution area, and uses a preset rock burst intelligent instant calculation model to process the non-continuous solution area and the continuous solution area in a parallel calculation manner, thereby improving the efficiency of rock burst calculation.

[0122] Corresponding to the above-mentioned intelligent second calculation method of rock burst, the present invention also proposes an intelligent second calculation device of rock burst. Since the device embodiment of the present invention corresponds to the above-mentioned method embodiment, the details not disclosed in the device embodiment can be referred to the above-mentioned method embodiment, and will not be repeated in the present invention.

[0123] Fig.18 A schematic diagram of the structure of an intelligent instant calculation device for rock burst pressure provided by an embodiment of the present disclosure, such as Fig.18 As shown, the device is applied to a server and includes:

[0124] A first acquisition unit 31 is used to acquire a region to be calculated; the region to be calculated includes a region to be destroyed and a region not to be destroyed;

[0125] A segmentation unit 32, used for segmenting the to-be-destroyed area into at least two non-continuous solution areas;

[0126] An integration unit 33, used for integrating the non-to-be-destroyed area into a continuous solution area;

[0127] The input unit 34 is used to input the discontinuous solution area and the continuous solution area into the preset rock burst intelligent instant calculation model respectively to obtain the rock burst of the area to be calculated; the preset rock burst intelligent instant calculation model processes the discontinuous solution area and the continuous solution area in a parallel calculation manner.

[0128] The intelligent instant calculation device for rock burst provided by the present disclosure obtains an area to be calculated; the area to be calculated includes an area to be destroyed and an area not to be destroyed; the area to be destroyed is divided into at least two non-continuous solution areas, and the area not to be destroyed is integrated into a continuous solution area; the non-continuous solution area and the continuous solution area are respectively input into a preset rock burst intelligent instant calculation model to obtain the rock burst in the area to be calculated; the preset rock burst intelligent instant calculation model processes the non-continuous solution area and the continuous solution area in a parallel calculation manner. Compared with the related art, the embodiment of the present disclosure improves the efficiency of rock burst calculation by dividing the area to be destroyed in the area to be calculated into at least two non-continuous solution areas, and integrating the non-to-be-destroyed area in the area to be calculated into a continuous solution area, and using the preset rock burst intelligent instant calculation model to process the non-continuous solution area and the continuous solution area in a parallel calculation manner.

[0129] Furthermore, in a possible implementation of this embodiment, as Fig.19 As shown, the device also includes:

[0130] A second acquisition unit 35, used to acquire training data;

[0131] The training unit 36 ​​is used to train the initial rock burst intelligent calculation model in seconds using the training data based on a deep neural network algorithm to obtain the preset rock burst intelligent calculation model in seconds.

[0132] Furthermore, in a possible implementation of this embodiment, as Fig.19 As shown, the second acquisition unit 35 includes:

[0133] A first acquisition module 351 is used to respectively acquire a training simulated rock burst pressure and a training experimental rock burst pressure, and acquire loads corresponding to the training simulated rock burst pressure and the training experimental rock burst pressure;

[0134] The combination module 352 is used to combine the training simulated rock burst pressure, the training experimental test rock burst pressure and the load to obtain the training data.

[0135] Furthermore, in a possible implementation of this embodiment, as Fig.19 As shown, the training unit 36 ​​includes:

[0136] The second acquisition module 361 is used to acquire the model parameters of the initial rock burst intelligent instant calculation model;

[0137] A learning module 362 is used to learn the training simulated rock burst pressure, the training experimental test rock burst pressure and the load to obtain a loss function; the loss function includes the model parameters;

[0138] The adjustment module 363 is used to adjust the model parameters in a manner that makes the loss function approach the minimum value, so as to obtain the preset rock burst intelligent instant calculation model.

[0139] Furthermore, in a possible implementation of this embodiment, as Fig.19 As shown, the device also includes:

[0140] A decomposition unit 37, used for calling a central processing unit to decompose the non-continuous solution area and the continuous solution area into at least two tasks respectively;

[0141] The computing unit 38 is used to call at least two graphics processor sub-cores to perform parallel computing on the at least two tasks respectively.

[0142] Furthermore, in a possible implementation of this embodiment, the preset rock burst intelligent instant calculation model includes a preset continuous solution neural network and a preset discontinuous solution neural network; wherein,

[0143] The preset continuous solution neural network is used to process the continuous solution area;

[0144] The preset non-continuous solution neural network is used to process the non-continuous solution neural network.

[0145] It should be noted that the above explanation of the method embodiment is also applicable to the device of the embodiment of the present disclosure, and the principle is the same, which is no longer limited in the embodiment of the present disclosure.

[0146] According to an embodiment of the present disclosure, the present disclosure also provides an electronic device, a readable storage medium and a computer program product.

[0147] Fig. 20A schematic block diagram of an example electronic device 400 that can be used to implement an embodiment of the present disclosure is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present disclosure described and / or required herein.

[0148] like Fig. 20 As shown, the device 400 includes a computing unit 401, which can perform various appropriate actions and processes according to a computer program stored in a ROM (Read-Only Memory) 402 or a computer program loaded from a storage unit 408 to a RAM (Random Access Memory) 403. In the RAM 403, various programs and data required for the operation of the device 400 can also be stored. The computing unit 401, the ROM 402, and the RAM 403 are connected to each other via a bus 404. An I / O (Input / Output) interface 405 is also connected to the bus 404.

[0149] A number of components in the device 400 are connected to the I / O interface 405, including: an input unit 406, such as a keyboard, a mouse, etc.; an output unit 407, such as various types of displays, speakers, etc.; a storage unit 408, such as a disk, an optical disk, etc.; and a communication unit 409, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 409 allows the device 400 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0150] The computing unit 401 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the computing unit 401 include, but are not limited to, a CPU (Central Processing Unit), a GPU (Graphic Processing Units), various dedicated AI (Artificial Intelligence) computing chips, various computing units running machine learning model algorithms, a DSP (Digital Signal Processor), and any appropriate processor, controller, microcontroller, etc. The computing unit 401 performs the various methods and processes described above, such as the intelligent second calculation method of rock burst pressure. For example, in some embodiments, the intelligent second calculation method of rock burst pressure may be implemented as a computer software program, which is tangibly contained in a machine-readable medium, such as a storage unit 408. In some embodiments, part or all of the computer program may be loaded and / or installed on the device 400 via the ROM 402 and / or the communication unit 409. When the computer program is loaded into the RAM 403 and executed by the computing unit 401, one or more steps of the method described above may be executed. Alternatively, in other embodiments, the calculation unit 401 may be configured to execute the aforementioned intelligent second calculation method of impact ground pressure by any other appropriate means (for example, by means of firmware).

[0151] Various embodiments of the systems and techniques described above herein may be implemented in digital electronic circuit systems, integrated circuit systems, FPGAs (Field Programmable Gate Arrays), ASICs (Application-Specific Integrated Circuits), ASSPs (Application Specific Standard Products), SOCs (System On Chips), CPLDs (Complex Programmable Logic Devices), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include: being implemented in one or more computer programs that may be executed and / or interpreted on a programmable system including at least one programmable processor that may be a dedicated or general-purpose programmable processor that may receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0152] The program code for implementing the method of the present disclosure may be written in any combination of one or more programming languages. These program codes may be provided to a processor or controller of a general-purpose computer, a special-purpose computer, or other programmable data processing device, so that the program code, when executed by the processor or controller, enables the functions / operations specified in the flow chart and / or block diagram to be implemented. The program code may be executed entirely on the machine, partially on the machine, partially on the machine and partially on a remote machine as a stand-alone software package, or entirely on a remote machine or server.

[0153] In the context of the present disclosure, a machine-readable medium may be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, device, or equipment. A machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or device, or any suitable combination of the foregoing. More specific examples of machine-readable storage media may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a RAM, a ROM, an EPROM (Electrically Programmable Read-Only-Memory) or a flash memory, an optical fiber, a CD-ROM (Compact Dis sc Read-Only Memory), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0154] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device (e.g., a CRT (Cathode-Ray Tube) or LCD (Liquid Crystal Display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user can provide input to the computer. Other types of devices can also be used to provide interaction with the user; for example, the feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including acoustic input, voice input, or tactile input).

[0155] The systems and techniques described herein may be implemented in a computing system that includes backend components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes frontend components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such backend components, middleware components, or frontend components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: LAN (Local Area Network), WAN (Wide Area Network), the Internet, and blockchain networks.

[0156] A computer system may include a client and a server. The client and the server are generally remote from each other and usually interact through a communication network. The relationship between the client and the server is generated by computer programs running on the corresponding computers and having a client-server relationship with each other. The server may be a cloud server, also known as a cloud computing server or cloud host, which is a host product in the cloud computing service system to solve the defects of difficult management and weak business scalability in traditional physical hosts and VPS services ("Virtual Private Server", or "VPS" for short). The server may also be a server of a distributed system, or a server combined with a blockchain.

[0157] It should be noted that artificial intelligence is a discipline that studies how computers can simulate certain human thought processes and intelligent behaviors (such as learning, reasoning, thinking, planning, etc.), and includes both hardware-level and software-level technologies. Artificial intelligence hardware technologies generally include technologies such as sensors, dedicated artificial intelligence chips, cloud computing, distributed storage, and big data processing; artificial intelligence software technologies mainly include computer vision technology, speech recognition technology, natural language processing technology, as well as machine learning / deep learning, big data processing technology, knowledge graph technology, and other major directions.

[0158] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps recorded in this disclosure can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solutions disclosed in this disclosure can be achieved, and this document does not limit this.

[0159] The above specific implementations do not constitute a limitation on the protection scope of the present disclosure. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present disclosure shall be included in the protection scope of the present disclosure.

Claims

1. An intelligent method for calculating rock burst in seconds, characterized in that: include: Acquire a region to be calculated; the region to be calculated includes a region to be destroyed and a region not to be destroyed; Dividing the to-be-destroyed area into at least two non-continuous solution areas, and integrating the non-to-be-destroyed area into a continuous solution area; The discontinuous solution area and the continuous solution area are respectively input into a preset rock burst intelligent second calculation model to obtain the rock burst in the area to be calculated; the preset rock burst intelligent second calculation model processes the discontinuous solution area and the continuous solution area in a parallel calculation manner; The preset rock burst intelligent instant calculation model includes a preset continuous solution neural network and a preset discontinuous solution neural network; The preset continuous solution neural network is used to process the continuous solution area; The preset non-continuous solution neural network is used to process the non-continuous solution neural network.

2. The method according to claim 1, characterized in that The training method of the preset rock burst intelligent instant calculation model includes: Get training data; Based on a deep neural network algorithm, the initial rock burst intelligent calculation model is trained using the training data to obtain the preset rock burst intelligent calculation model.

3. The method according to claim 2, characterized in that The obtaining of training data comprises: Respectively obtaining a simulated rock burst pressure for training and a rock burst pressure for experimental testing for training, and obtaining loads corresponding to the simulated rock burst pressure for training and the rock burst pressure for experimental testing for training; The training simulated rock impact pressure, the training experimental test rock impact pressure and the load are combined to obtain the training data.

4. The method according to claim 3, characterized in that The deep neural network algorithm is based on which the initial rock burst intelligent calculation model is trained using the training data to obtain the preset rock burst intelligent calculation model, which includes: Obtaining model parameters of the initial rock burst intelligent instant calculation model; The simulated rock burst pressure for training, the rock burst pressure for training experimental test and the load are studied to obtain a loss function; the loss function includes the model parameters; The model parameters are adjusted in a manner that makes the loss function approach the minimum value to obtain the preset rock burst intelligent instant calculation model.

5. The method according to claim 1, characterized in that The preset rock burst intelligent instant calculation model processes the discontinuous solution area and the continuous solution area in a parallel calculation manner, including: Calling a central processing unit to decompose the non-continuous solution area and the continuous solution area into at least two tasks respectively; At least two graphics processor sub-cores are called to respectively perform parallel calculations on the at least two tasks.

6. An intelligent instant calculation device for rock burst, characterized in that: include: A first acquisition unit is used to acquire a region to be calculated; the region to be calculated includes a region to be destroyed and a region not to be destroyed; A segmentation unit, used for segmenting the to-be-destroyed area into at least two non-continuous solution areas; An integration unit, used for integrating the non-to-be-destroyed area into a continuous solution area; An input unit is used to input the discontinuous solution area and the continuous solution area into a preset rock burst intelligent second calculation model respectively to obtain the rock burst of the area to be calculated; the preset rock burst intelligent second calculation model processes the discontinuous solution area and the continuous solution area in a parallel calculation manner; The preset rock burst intelligent instant calculation model includes a preset continuous solution neural network and a preset discontinuous solution neural network; The preset continuous solution neural network is used to process the continuous solution area; The preset non-continuous solution neural network is used to process the non-continuous solution neural network.

7. An electronic device, characterized in that: include: at least one processor; as well as a memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method according to any one of claims 1 to 5.

8. A non-transitory computer-readable storage medium storing computer instructions, characterized in that: The computer instructions are used to cause the computer to execute the method according to any one of claims 1-5.

9. A computer program product, characterized in that The invention comprises a computer program which, when executed by a processor, implements the method according to any one of claims 1 to 5.

Images