CN116245007A - Joint rock REV numerical simulation acquisition method and system based on image recognition - Google Patents

Joint rock REV numerical simulation acquisition method and system based on image recognition Download PDF

Info

Publication number
CN116245007A
CN116245007A CN202211581031.1A CN202211581031A CN116245007A CN 116245007 A CN116245007 A CN 116245007A CN 202211581031 A CN202211581031 A CN 202211581031A CN 116245007 A CN116245007 A CN 116245007A
Authority
CN
China
Prior art keywords
rock mass
model
jointed rock
image
jointed
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN202211581031.1A
Other languages
Chinese (zh)
Inventor
刘聪
孙基伟
周宗青
刘洪亮
李刚
白松松
靳高汉
高天
刘雨函
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Shandong University
Original Assignee
Shandong University
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Shandong University filed Critical Shandong University
Priority to CN202211581031.1A priority Critical patent/CN116245007A/en
Publication of CN116245007A publication Critical patent/CN116245007A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F30/00—Computer-aided design [CAD]
    • G06F30/20—Design optimisation, verification or simulation
    • G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F30/00—Computer-aided design [CAD]
    • G06F30/20—Design optimisation, verification or simulation
    • G06F30/23—Design optimisation, verification or simulation using finite element methods [FEM] or finite difference methods [FDM]
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/40—Extraction of image or video features
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
    • G06V10/77—Processing image or video features in feature spaces; using data integration or data reduction, e.g. principal component analysis [PCA] or independent component analysis [ICA] or self-organising maps [SOM]; Blind source separation
    • G06V10/774—Generating sets of training patterns; Bootstrap methods, e.g. bagging or boosting
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
    • G06V10/82—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using neural networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2111/00—Details relating to CAD techniques
    • G06F2111/10—Numerical modelling
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2119/00—Details relating to the type or aim of the analysis or the optimisation
    • G06F2119/14—Force analysis or force optimisation, e.g. static or dynamic forces
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
    • Y02T90/00—Enabling technologies or technologies with a potential or indirect contribution to GHG emissions mitigation

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Evolutionary Computation (AREA)
  • General Physics & Mathematics (AREA)
  • Artificial Intelligence (AREA)
  • Software Systems (AREA)
  • Multimedia (AREA)
  • General Health & Medical Sciences (AREA)
  • General Engineering & Computer Science (AREA)
  • Computing Systems (AREA)
  • Medical Informatics (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Health & Medical Sciences (AREA)
  • Computer Hardware Design (AREA)
  • Databases & Information Systems (AREA)
  • Geometry (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Biomedical Technology (AREA)
  • Biophysics (AREA)
  • Computational Linguistics (AREA)
  • Data Mining & Analysis (AREA)
  • Molecular Biology (AREA)
  • Mathematical Physics (AREA)
  • Image Analysis (AREA)

Abstract

The invention relates to the technical field of rock mass numerical simulation, and provides an jointed rock mass REV numerical simulation acquisition method and system based on image recognition, which can automatically acquire the occurrence information of a jointed rock mass and ensure the modeling accuracy of the jointed rock mass. The image of the jointed rock mass is identified through the convolutional neural network, the structure surface information acquisition work is simplified, the structure surface acquisition precision and efficiency are improved, a multi-scale synthetic rock mass model is built based on the structure surface information obtained through image identification, and the property of the jointed rock mass is accurately represented. Drawing a model size-mechanical property parameter change curve, determining the size of the jointed rock mass characterization unit body by using a difference ratio, perfecting the determination work of the characterization unit body, and realizing quantification and standardization of the characterization unit body acquisition method. Provides a foundation for simulating a discontinuous medium by adopting a continuous medium mechanical calculation method.

Description

Joint rock REV numerical simulation acquisition method and system based on image recognition
Technical Field
The disclosure relates to the technical field of rock mass numerical simulation, in particular to a jointed rock mass REV numerical simulation acquisition method and system based on image recognition.
Background
The statements in this section merely provide background information related to the present disclosure and may not necessarily constitute prior art.
The tunnel is located in the rock mass environment and has quite large difference with the occurrence environment of the ground building, so that the exploration of the mechanical property and deformation mechanism of the rock mass is very important. The rock mass consists of a discontinuous surface and a complete rock mass, and the mechanical properties of the rock mass are also determined by both. Among them, discontinuous surfaces include joints, fissures, layers, faults, and the like, and the properties of rock mass are extremely complex due to the ever-changing discontinuous surfaces. The jointed rock mass belongs to a discontinuous medium, and the mechanical property of the jointed rock mass changes along with the increase of the size of the rock mass, namely the size effect of the jointed rock mass. The mechanical properties of the jointed rock mass tend to stabilize as the volume of the rock mass increases to a critical value, which is the characteristic unit volume (REV) size of the jointed rock mass. The mechanical properties of the jointed rock mass, which are larger than the characterizing unit body size, remain substantially unchanged. At this time, the jointed rock mass can be regarded as a continuous medium, and the corresponding mechanical parameters can be used in continuous medium mechanical calculation. Therefore, the characterization unit body for determining the jointed rock mass is the basis for simulating the jointed rock mass by adopting a continuous medium mechanical calculation method (finite element method, finite difference method and the like), and has important research significance.
The inventor finds that the prior determination method for representing the unit body mostly adopts a numerical simulation method, the method is not limited by the rock mass environment, and the method can simulate the mechanical test process of the engineering scale rock mass by using a computer and has certain research advantages. An accurate jointed rock mass model needs to be established before simulation, but acquiring joint distribution in a real rock mass is a difficult problem, and currently, rock mass images are captured by a measuring line method and a laser scanning method, and the jointed rock mass is manually identified and the occurrence information is determined, so that the method has certain subjectivity and consumes a great deal of cost and time.
Disclosure of Invention
In order to solve the problems, the invention provides a method and a system for acquiring the REV value simulation of the jointed rock based on image recognition, which can automatically acquire the occurrence information of the jointed rock, ensure the modeling accuracy of the jointed rock and lay a foundation for acquiring the REV of the jointed rock.
In order to achieve the above purpose, the present disclosure adopts the following technical scheme:
one or more embodiments provide an jointed rock REV numerical simulation acquisition method based on image recognition, including the steps of:
identifying the image characteristics of the jointed rock through a trained convolutional neural network by using the acquired jointed rock image to be identified, and obtaining the occurrence parameters of the joints;
establishing a multi-scale jointed rock mass model based on the occurrence parameters by adopting a discrete element method, simulating an indoor test process, and testing to obtain mechanical property parameters of the multi-scale jointed rock mass model;
and drawing a model size-mechanical property parameter curve according to the obtained mechanical property parameter to obtain the REV size and the corresponding mechanical parameter of the jointed rock mass.
One or more embodiments provide an jointed rock mass REV numerical simulation acquisition system based on image recognition, an image acquisition device, and a processor;
the image acquisition device is used for acquiring an image of the jointed rock mass;
the processor is configured to perform the steps of the method described above.
An electronic device comprising a memory and a processor and computer instructions stored on the memory and running on the processor, which when executed by the processor, perform the steps of the method described above.
A computer readable storage medium storing computer instructions which, when executed by a processor, perform the steps of the method described above.
Compared with the prior art, the beneficial effects of the present disclosure are:
in the method, the image of the jointed rock mass is identified through the convolutional neural network, the structure surface information acquisition work is simplified, the structure surface acquisition precision and efficiency are improved, a multi-scale synthetic rock mass model is built based on the structure surface information obtained through image identification, and the property of the jointed rock mass is accurately represented. Drawing a model size-mechanical property parameter change curve, determining the size of the jointed rock mass characterization unit body by using a difference ratio, perfecting the determination work of the characterization unit body, and realizing quantification and standardization of the characterization unit body acquisition method. Provides a foundation for simulating a discontinuous medium by adopting a continuous medium mechanical calculation method.
The advantages of the present disclosure, as well as those of additional aspects, will be described in detail in the following detailed description of embodiments.
Drawings
The accompanying drawings, which are included to provide a further understanding of the disclosure, illustrate and explain the exemplary embodiments of the disclosure and together with the description serve to explain and do not limit the disclosure.
FIG. 1 is a block diagram of an jointed rock mass REV numerical simulation acquisition system of embodiment 2 of the present disclosure;
FIG. 2 is a schematic illustration of a synthetic rock mass model structure of example 1 of the present disclosure;
FIG. 3 is a method flow chart of a jointed rock mass REV numerical simulation acquisition method of embodiment 1 of the present disclosure;
FIG. 4 is a graph of an engineering-scale fracture discrete network model of example 1 of the present disclosure;
FIG. 5 is a graph of a multi-scale discrete fracture network model of embodiment 1 of the present disclosure;
FIG. 6 is a diagram of a multi-scale synthetic rock mass model of example 1 of the present disclosure;
FIG. 7 is a simulated schematic diagram of a mechanical test of example 1 of the present disclosure;
FIG. 8 is a schematic diagram of a characterization unit acquisition method of example 1 of the present disclosure;
fig. 9 is a finite element simulation schematic of a tunnel excavation process of embodiment 1 of the present disclosure.
Detailed Description
The disclosure is further described below with reference to the drawings and examples.
It should be noted that the following detailed description is exemplary and is intended to provide further explanation of the present disclosure. Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this disclosure belongs.
It is noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of exemplary embodiments in accordance with the present disclosure. As used herein, the singular is also intended to include the plural unless the context clearly indicates otherwise, and furthermore, it is to be understood that the terms "comprises" and/or "comprising" when used in this specification are taken to specify the presence of stated features, steps, operations, devices, components, and/or combinations thereof. It should be noted that, without conflict, the various embodiments and features of the embodiments in the present disclosure may be combined with each other. The embodiments will be described in detail below with reference to the accompanying drawings.
Example 1
In the technical scheme disclosed in one or more embodiments, as shown in fig. 3 to 9, a jointed rock REV numerical simulation acquisition method based on image recognition includes the following steps:
step 1, identifying the image characteristics of the jointed rock mass through a trained convolutional neural network by using the acquired jointed rock mass image to be identified, and obtaining the occurrence parameters of the joint;
step 2, establishing a multi-scale jointed rock mass model based on the occurrence parameters by a discrete element method, simulating an indoor test process, and testing to obtain mechanical property parameters of the multi-scale jointed rock mass model;
and 3, drawing a model size-mechanical property parameter curve according to the obtained mechanical property parameter to obtain the REV size and the corresponding mechanical parameter of the jointed rock mass.
In the embodiment, the image of the jointed rock mass is identified through the convolutional neural network, the structure surface information acquisition work is simplified, the structure surface acquisition precision and efficiency are improved, a multi-scale synthetic rock mass model is built based on the structure surface information obtained through image identification, and the property of the jointed rock mass is accurately represented. Drawing a model size-mechanical property parameter change curve, determining the size of the jointed rock mass characterization unit body by using a difference ratio, perfecting the determination work of the characterization unit body, and realizing quantification and standardization of the characterization unit body acquisition method. The method provides a basis for simulating discontinuous media (such as jointed rock mass) by adopting a continuous media mechanical calculation method (such as finite element method, finite difference method and the like).
In the step 1, an image of the jointed rock mass is acquired through a set image sampling device.
Optionally, the image sampling device may include an illumination device, a laser pointer, and a camera device; the camera device may be a camera or a digital camera.
The illuminating device can be placed near the tunnel face to illuminate the surface of the joint rock body to be collected, and a good illumination environment is provided. The laser pointer irradiates the surface of the jointed rock body, and the laser point is used as a digital camera to calibrate the acquisition area, such as a rectangular area of 10 multiplied by 10 m. And acquiring the jointed rock mass image by adopting a digital camera and storing an image file.
The step 1 further includes a step of preprocessing the acquired jointed rock mass image, wherein the preprocessing includes rotation image, graying processing and the like.
Alternatively, the angles of all jointed rock mass images may be made the same by rotating the image angles. By adjusting the image size, the pixels of all images are unified.
The RGB image file is grayed, and the calculation formula can be:
Gray＝(299+587+114+500)/1000
wherein Gray is Gray, R is red, G is green, and B is blue.
In step 1, identifying image features of jointed rock mass through a trained convolutional neural network, including:
step 11, carrying out convolution, pooling and connection operation on the preprocessed image, filtering irrelevant information in the image and extracting to obtain joint image characteristics;
step 12, performing nonlinear combination on the obtained joint image characteristics to obtain integrated high-order information;
step 13: giving classification labels according to the high-order information of the jointed rock mass image to obtain the zone shape information of the jointed rock mass structural plane;
wherein, the jointed rock mass structural plane attitude information may include: information such as inclination angle, inclination, trace length and interval of the jointed rock mass.
Alternatively, the structure of the convolutional neural network may include an input layer, a convolutional layer, a pooling layer, a fully-connected layer, and an output layer. The input layer is used for inputting the preprocessed jointed rock mass image; the convolution layer is used for extracting joint image features; the pooling layer is used for selecting characteristics and filtering irrelevant information; the full-connection layer carries out nonlinear combination on the image characteristics obtained by the convolution layer and the pooling layer to obtain integrated high-order information; the output layer gives classification labels to the high-order information obtained by the full-connection layer, and transmits and outputs label information;
the embodiment identifies the image of the jointed rock mass through the convolutional neural network, acquires the occurrence parameters of the structural surface, and provides basic information for jointed rock mass modeling. Meanwhile, subjectivity and inaccuracy of a survey line method and three-dimensional laser scanning are made up, the structure surface information acquisition work is simplified, and the structure surface acquisition precision and efficiency are improved.
Further, the method also comprises a process of training the convolutional neural network, and specifically comprises the following steps:
s11, acquiring an jointed rock mass image, and preprocessing to obtain a jointed rock mass image set;
the preprocessing step is the same as the preprocessing method in the identification stage, and will not be described here again.
Step S12, constructing a characteristic data set corresponding to the jointed rock mass images and the occurrence labels one by one according to the occurrence label of the structural surface of the images in the image set; alternatively, 70% of the feature data set may be used as the training set and the remaining 30% of the feature data set may be used as the test set.
Wherein the yield marker tag may comprise: trend, dip angle, track length, and pitch.
S13, taking the preprocessed jointed rock mass image as an input sample, taking structural surface occurrence information as an output sample, inputting the structural surface occurrence information into a convolutional neural network for training to obtain a convolutional neural network model, and establishing a nonlinear mathematical relationship between the jointed rock mass image and the structural surface occurrence (tendency, dip angle, trace length and interval);
step S14, using the test set to check the accuracy of the convolutional neural network model, and adopting regression quantization index R 2 Judging the model precision, wherein the model precision does not meet the requirement, and continuously executing the training step; the model precision meets the requirement, and the model parameters are determined to obtain the trained convolutional neural network.
Specifically, in the method for judging the model accuracy, a judgment threshold value may be set to 0.9; if R is 2 If the model is larger than the threshold value, the convolutional neural network model is determined to be accurate, the model is stored, and a new jointed rock mass image is continuously identified; if R is 2 If the parameter is smaller than the threshold value, continuing to adjust the parameter and optimizing the model algorithm;
wherein R is 2 Is the sum of squares of regressionThe larger the number is, the better the image recognition effect is, the more accurate the model is, the more obvious the regression effect is, and the calculation formula is as follows:
Figure BDA0003991159020000071
wherein y is i For the actual value of the i-th test sample,
Figure BDA0003991159020000072
the i-th sample value output for the model, < +.>
Figure BDA0003991159020000073
Is the average of the actual values of the test samples.
In step 2, the indoor test may optionally include a uniaxial compression test, a triaxial compression test, or the like.
Among other mechanical properties parameters, may include uniaxial compressive strength, elastic modulus, poisson's ratio, tensile strength, cohesion and internal friction angle.
In step 2, a multi-scale jointed rock mass model is established based on the occurrence parameters, and the method comprises the following steps:
step 21, constructing a probability distribution model obeyed by the structural surface attitude parameters through statistical analysis, and determining attitude distribution characteristic values;
specifically, the probability distribution model may employ a normal distribution, a negative exponential distribution, a lognormal distribution, or the like. The distribution characteristic values may include: mean and standard deviation.
Step 22, an engineering scale fracture discrete network model is established according to the structural plane occurrence distribution characteristics, as shown in fig. 4, and is used for representing actual joint distribution, and a series of multiscale discrete fracture network models with the same proportion and increasing size are cut outwards from the model center point of the engineering scale fracture discrete network model, as shown in fig. 5.
The synthetic rock mass model includes a discrete fracture network model and a bonded particle model, as shown in fig. 2, wherein the discrete fracture network model is established based on the structural plane probability distribution model and the distribution eigenvalues.
And 23, calibrating mechanical property parameters of the jointed rock mass obtained through an indoor test to obtain microscopic parameters of a bonding particle model, and endowing the bonding particle model to a multi-scale fracture network model to obtain a multi-scale synthetic rock mass model, as shown in fig. 6.
The mechanical property parameters of the jointed rock mass include: uniaxial compressive strength, poisson ratio, elastic modulus, tensile strength, cohesive force and internal friction angle;
specifically, the mesoscopic parameters of the bond particle model include: particle radius, bond stiffness ratio, bond tensile strength, bond shear strength, effective modulus, and the like.
After the multi-scale synthetic rock mass model is constructed, as shown in fig. 7, mechanical test simulation is performed to obtain mechanical property parameters (uniaxial compressive strength, elastic modulus, poisson ratio, tensile strength, cohesive force and internal friction angle) of the multi-scale synthetic rock mass model.
Mechanical test simulations were performed, which may include uniaxial compression, brazilian split, and triaxial compression test simulations.
Optionally, the mechanical property parameters of the multi-scale synthetic rock mass model may include: uniaxial compressive strength, elastic modulus, poisson ratio, tensile strength, cohesive force and internal friction angle.
In the embodiment, a multi-scale synthetic rock mass model is established based on structural plane information obtained through image recognition, so that the properties of the jointed rock mass are accurately represented. The mechanical property parameters of the multi-scale synthetic rock mass model are obtained through experimental simulation, a model size-mechanical property parameter change curve is drawn, the size of the jointed rock mass characterization unit body is determined by using the difference ratio, the determination work of the characterization unit body is perfected, and the quantification and standardization of the characterization unit body acquisition method are realized.
Further, in order to intuitively compare the change of the mechanical property parameters of the rock mass, as shown in fig. 8, a model size-mechanical property parameter curve is drawn, specifically: and drawing a curve by taking the mechanical property parameters of the multi-scale synthetic rock mass model as a vertical axis and the size of the synthetic rock mass model as a horizontal axis to obtain a curve change trend.
The method for obtaining REV size and corresponding mechanical parameters of the jointed rock mass according to the size-mechanical property parameter curve comprises the following specific steps: setting a difference ratio threshold, which may be set to 5% -15%, preferably, may be 10%; calculating the difference ratio of each synthetic rock mass model in the multi-scale synthetic rock mass models, wherein the synthetic rock mass model with the difference ratio smaller than a threshold value (10%) is used as a characterization unit body, and the corresponding mechanical property parameters, namely REV parameters, are used for determining the size of the characterization unit body and the corresponding mechanical property parameters;
wherein the difference ratio d i Reflecting the variation degree of the mechanical property parameters of the adjacent dimension models, d i The smaller the parameter variation degree is, the more stable the mechanical property of the jointed rock mass is, and the difference ratio d is i The calculation formula is as follows:
Figure BDA0003991159020000091
wherein p is i For the property parameter of the ith dimension model, p i-1 Is a property parameter of the i-1 level size model.
In this embodiment, the REV size and the corresponding mechanical parameters of the jointed rock body are obtained by the above method, and further, the method further includes a step of verifying the REV size and the corresponding mechanical parameters of the jointed rock body, and the accuracy is: as shown in fig. 9, the tunnel excavation process is simulated by finite element software, the stratum model in the finite element software adopts the mechanical property parameters corresponding to the characterization unit body, and the accuracy of the obtained characterization unit body is verified by comparing displacement curves obtained by finite element simulation and tunnel construction monitoring measurement. When the curves are consistent, then REV is deemed to be accurate and the simulation is ended. When the curves are inconsistent, REV is determined to be inaccurate, parameters of the synthesized rock mass model are required to be adjusted, and the numerical simulation process can be carried out again.
According to the method, the tunnel excavation process is simulated based on the finite element, the obtained volume of the characterization unit body and the corresponding mechanical property parameters are endowed to the stratum model, and the accuracy of the characterization unit body is verified by comparing the finite element simulation and the tunnel construction monitoring measurement results. The method improves the acquisition method of the characterization unit body, improves the accuracy of REV, improves the experience and verification work of the characterization unit body, and provides a foundation for simulating the jointed rock body by adopting a continuous medium mechanical calculation method according to the obtained REV information of the jointed rock body of the tunnel section.
Example 2
In this embodiment, an image recognition-based jointed rock REV numerical simulation acquisition system is provided, as shown in fig. 1, including: an image acquisition device and a processor configured to perform the steps of the method described in embodiment 1.
In some embodiments, the image acquisition device is used for acquiring and acquiring the jointed rock mass image of the tunnel site, and specifically, the image acquisition device can comprise a digital camera, an illumination device and a laser indicator.
Digital camera: the method comprises the steps of shooting an image of an jointed rock mass and storing an image file;
an illumination device: the illuminating device is used for assisting in collecting images, and can provide a good illuminating environment for the digital camera due to insufficient light in the tunnel;
laser indicator: the laser point is used for assisting in collecting images, and the laser point is used for calibrating a collecting area of the digital camera, so that the images can be conveniently collected.
The processor comprises an image recognition unit, a numerical simulation unit and a characterization unit body acquisition unit.
In some embodiments, the image recognition unit is configured to recognize the acquired jointed rock mass image, and is configured to: identifying the image characteristics of the jointed rock through a trained convolutional neural network by using the acquired jointed rock image to be identified, and obtaining the occurrence parameters of the joints;
the image recognition unit comprises an image preprocessing module, a convolutional neural network module and a structural plane information acquisition module;
further, the image preprocessing module is used for standardizing the acquired image, and graying the acquired RGB image, so that subsequent image processing is facilitated. Simultaneously, the image angle is rotated, so that the angles of all jointed rock mass images are the same. And through adjusting the size of the image, the pixels of all the images are unified, and the consistency of an image set is ensured.
Further, the convolutional neural network module is used for learning and identifying structural plane shape information of the jointed rock mass image. The module consists of an input layer, a convolution layer, a pooling layer, a full connection layer and an output layer. The input layer is used for inputting the preprocessed jointed rock mass image; the convolution layer is used for extracting joint image features; the pooling layer is used for selecting characteristics and filtering irrelevant information; the full-connection layer carries out nonlinear combination on the image characteristics obtained by the convolution layer and the pooling layer to obtain integrated high-order information; the output layer gives classification labels to the high-order information obtained by the full-connection layer, and transmits the label information to the structural surface information acquisition module;
further, the structural plane information acquisition module is used for acquiring structural plane shape parameters obtained through image recognition, including inclination, inclination angle, trace length and interval; and classifying the label information of the output layer to obtain the structural plane shape information on the jointed rock mass image.
In some embodiments, the numerical simulation unit is for building different volumes of jointed rock mass models and testing mechanical parameters thereof, configured for: and establishing a multi-scale jointed rock mass model based on the occurrence parameters by adopting a discrete element method, simulating an indoor test process, and testing to obtain the mechanical property parameters of the multi-scale jointed rock mass model.
The numerical simulation unit comprises a structural plane statistical analysis module, a jointed rock mass modeling module and a mechanical test simulation module;
further, the structural plane statistical analysis module is used for determining probability distribution models (normal distribution, negative exponential distribution, lognormal distribution and the like) and distribution characteristic values (mean and standard deviation) of structural plane morphology compliance. Obtaining a statistical analysis result of the structural surface by drawing a structural surface information-probability density histogram and fitting a curve;
further, the jointed rock mass modeling module is configured to build a multi-scale synthetic rock mass model that characterizes engineering dimensions of the jointed rock mass.
Specifically, an engineering scale synthetic rock mass model is used to characterize the jointed rock mass, the synthetic rock mass model consisting of a discrete fracture network model and a bonded particle model, as shown in fig. 2. And establishing a discrete fracture network model based on the structural plane probability distribution model and the distribution characteristic values. The mechanical property parameters (uniaxial compressive strength, poisson ratio, elastic modulus, tensile strength, cohesive force and internal friction angle) of the jointed rock mass obtained through the indoor test are calibrated to obtain the mesoscopic parameters (particle radius, cohesive stiffness ratio, cohesive tensile strength, cohesive shear strength, effective modulus and the like) of the synthetic rock mass model. A series of models with consistent proportion and increasing volume are selected outwards from the center of the established engineering scale synthetic rock mass model, namely a multi-scale synthetic rock mass model;
further, the mechanical test simulation module: is configured to perform mechanical test simulation (uniaxial compression, brazilian split and triaxial compression test simulation) to obtain mechanical property parameters of the multi-scale synthetic rock mass model.
In some embodiments, the unit volume acquisition unit is characterized by: the method comprises the steps of obtaining the size of a characterization unit body of the jointed rock body and the corresponding mechanical property parameters thereof, and drawing a model size-mechanical property parameter curve according to the obtained mechanical property parameters to obtain the REV size and the corresponding mechanical parameters of the jointed rock body;
the characterization unit body acquisition unit comprises a mechanical property curve drawing module, a quantization index calculation module and a characterization unit body verification module;
further, the mechanical property curve drawing module: the method is configured to draw a change curve by taking the mechanical property parameters of the multi-scale synthetic rock mass model as the vertical axis and the model size as the horizontal axis, so as to obtain the curve change trend.
Further, the quantization index calculation module: is configured for determining a size of the characterization unit cell. Specific: calculating the difference ratio of the multi-scale synthetic rock mass model, taking the absolute value of the difference ratio as less than 10% as a quantization index for determining the characterization unit body, and determining the size of the characterization unit body and the corresponding mechanical property parameters;
further, the characterization unit experience verification module: is configured to verify the accuracy of the acquired characterization unit cell. And simulating a tunnel excavation process through finite element software, wherein the stratum model adopts mechanical property parameters corresponding to the characterization unit body. And verifying the accuracy of the obtained characterization unit body by comparing displacement curves obtained by finite element simulation and tunnel construction monitoring measurement.
Example 3
The present embodiment provides an electronic device comprising a memory and a processor, and computer instructions stored on the memory and running on the processor, which when executed by the processor, perform the steps recited in the method of embodiment 1.
Example 4
The present embodiment provides a computer readable storage medium storing computer instructions that, when executed by a processor, perform the steps of the method of embodiment 1.
The foregoing description of the preferred embodiments of the present disclosure is provided only and not intended to limit the disclosure so that various modifications and changes may be made to the present disclosure by those skilled in the art. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present disclosure should be included in the protection scope of the present disclosure.
While the specific embodiments of the present disclosure have been described above with reference to the drawings, it should be understood that the present disclosure is not limited to the embodiments, and that various modifications and changes can be made by one skilled in the art without inventive effort on the basis of the technical solutions of the present disclosure while remaining within the scope of the present disclosure.

Claims (10)

1. The method for acquiring the REV numerical simulation of the jointed rock mass based on the image recognition is characterized by comprising the following steps of:
identifying the image characteristics of the jointed rock through a trained convolutional neural network by using the acquired jointed rock image to be identified, and obtaining the occurrence parameters of the joints;
establishing a multi-scale jointed rock mass model based on the occurrence parameters by adopting a discrete element method, simulating an indoor test process, and testing to obtain mechanical property parameters of the multi-scale jointed rock mass model;
and drawing a model size-mechanical property parameter curve according to the obtained mechanical property parameter to obtain the REV size and the corresponding mechanical parameter of the jointed rock mass.
2. The method for obtaining the jointed rock REV numerical simulation based on image recognition according to claim 1, wherein the method comprises the following steps: identifying image features of jointed rock mass through a trained convolutional neural network, comprising:
carrying out convolution, pooling and connection operation on the preprocessed image, filtering irrelevant information in the image and extracting to obtain joint image characteristics;
nonlinear combination is carried out on the obtained joint image characteristics to obtain integrated high-order information;
and giving classification labels according to the high-order information of the jointed rock mass image to obtain the zone shape information of the jointed rock mass structural plane.
3. The method for obtaining the jointed rock mass REV numerical simulation based on image recognition according to claim 1, further comprising the process of training a convolutional neural network, specifically comprising the following steps:
acquiring an jointed rock mass image, and preprocessing to obtain a jointed rock mass image set;
constructing a feature data set corresponding to the jointed rock mass images and the occurrence labels one by one according to the occurrence label of the structural surface of the images in the image set;
taking the preprocessed jointed rock mass image as an input sample, taking the structural surface attitude information as an output sample, and inputting the structural surface attitude information into a convolutional neural network for training;
the test set is used for checking the accuracy of the convolutional neural network model, and regression quantization index R is adopted 2 Judging the model precision, wherein the model precision does not meet the requirement, and continuously executing the training step; the model precision meets the requirement, and the model parameters are determined to obtain the trained convolutional neural network.
4. The method for obtaining the model of the jointed rock mass REV based on image recognition according to claim 1, wherein the method for establishing the model of the multi-scale jointed rock mass based on the occurrence parameters comprises the following steps:
constructing a probability distribution model obeyed by the structural plane attitude parameters through statistical analysis, and determining attitude distribution characteristic values;
establishing an engineering scale fracture discrete network model according to the structural surface occurrence distribution characteristics, and cutting and selecting a series of multiscale discrete fracture network models with the same proportion and increasing size from the model center point of the engineering scale fracture discrete network model outwards;
and calibrating mechanical property parameters of the jointed rock mass obtained through an indoor test to obtain mesoscopic parameters of a bonding particle model, and endowing the bonding particle model to a multi-scale fracture network model to obtain the multi-scale synthetic rock mass model.
5. The method for obtaining the jointed rock mass REV numerical simulation based on image recognition according to claim 1, wherein a model size-mechanical property parameter curve is drawn, specifically: and drawing a curve by taking the mechanical property parameters of the multi-scale synthetic rock mass model as a vertical axis and the size of the synthetic rock mass model as a horizontal axis to obtain a curve change trend.
6. The method for obtaining the jointed rock REV numerical simulation based on image recognition according to claim 1, wherein the method comprises the following steps: the method for obtaining REV size and corresponding mechanical parameters of the jointed rock mass according to the size-mechanical property parameter curve comprises the following specific steps:
setting a difference ratio threshold;
and calculating the difference ratio of each synthetic rock mass model in the multi-scale synthetic rock mass models, wherein the synthetic rock mass model with the difference ratio smaller than the threshold value is used as a characterization unit body, and the corresponding mechanical property parameter is the REV parameter.
7. The jointed rock REV numerical simulation acquisition system based on image recognition is characterized in that: an image acquisition device and a processor;
the image acquisition device is used for acquiring an image of the jointed rock mass;
the processor is configured to perform the steps of the method of any of claims 1-6.
8. The image recognition-based jointed rock REV numerical simulation acquisition system of claim 7, wherein: the image acquisition device comprises a digital camera, an illumination device and a laser indicator;
alternatively, the processor comprises an image recognition unit, a numerical simulation unit and a characterization unit body acquisition unit;
the image recognition unit is configured to recognize the image of the jointed rock to be recognized through the trained convolutional neural network, and obtain the occurrence parameters of the joints;
the numerical simulation unit is configured to establish a multi-scale jointed rock mass model based on the production parameters by adopting a discrete element method, simulate an indoor test process and test to obtain mechanical property parameters of the multi-scale jointed rock mass model;
the characterization unit body acquisition unit is configured to draw a model size-mechanical property parameter curve according to the obtained mechanical property parameters and acquire the REV size and the corresponding mechanical parameters of the jointed rock body.
9. An electronic device comprising a memory and a processor and computer instructions stored on the memory and running on the processor, which when executed by the processor, perform the steps of the method of any one of claims 1-6.
10. A computer readable storage medium storing computer instructions which, when executed by a processor, perform the steps of the method of any of claims 1-6.
CN202211581031.1A 2022-12-09 2022-12-09 Joint rock REV numerical simulation acquisition method and system based on image recognition Pending CN116245007A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202211581031.1A CN116245007A (en) 2022-12-09 2022-12-09 Joint rock REV numerical simulation acquisition method and system based on image recognition

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202211581031.1A CN116245007A (en) 2022-12-09 2022-12-09 Joint rock REV numerical simulation acquisition method and system based on image recognition

Publications (1)

Publication Number Publication Date
CN116245007A true CN116245007A (en) 2023-06-09

Family

ID=86630324

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202211581031.1A Pending CN116245007A (en) 2022-12-09 2022-12-09 Joint rock REV numerical simulation acquisition method and system based on image recognition

Country Status (1)

Country Link
CN (1) CN116245007A (en)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116680964A (en) * 2023-08-03 2023-09-01 中国地质大学（北京） Numerical simulation method and device for tunnel crossing active fault water and mud bursting disasters
CN117523299A (en) * 2023-11-21 2024-02-06 江苏财经职业技术学院 Image recognition method, system and storage medium based on computer network

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116680964A (en) * 2023-08-03 2023-09-01 中国地质大学（北京） Numerical simulation method and device for tunnel crossing active fault water and mud bursting disasters
CN116680964B (en) * 2023-08-03 2023-11-07 中国地质大学（北京） Numerical simulation method and device for tunnel crossing active fault water and mud bursting disasters
CN117523299A (en) * 2023-11-21 2024-02-06 江苏财经职业技术学院 Image recognition method, system and storage medium based on computer network

Similar Documents

Publication Publication Date Title
CN116245007A (en) Joint rock REV numerical simulation acquisition method and system based on image recognition
US11634987B2 (en) Safety early warning method and device for full-section tunneling of tunnel featuring dynamic water and weak surrounding rock
CN108489403B (en) Rapid and fine evaluation method for joint attitude of surface mine slope rock mass based on three-dimensional laser scanning
CN102081030B (en) Geotechnical model test system based on macro-micro mechanics as well as elaboration test method
CN113033284A (en) Vehicle real-time overload detection method based on convolutional neural network
CN109490072B (en) Detection system for civil engineering building and detection method thereof
CN104834806A (en) Joint roughness coefficient size effect sampling representativeness evaluation method
CN114894642B (en) Fatigue crack propagation rate testing method and device based on deep learning
CN115310361B (en) Underground coal mine dust concentration prediction method and system based on WGAN-CNN
CN113010835B (en) Dangerous rock collapse early warning method and system based on fracture mechanics
CN108830317B (en) Rapid and fine evaluation method for joint attitude of surface mine slope rock mass based on digital photogrammetry
CN110688955A (en) Building construction target detection method based on YOLO neural network
CN109209505B (en) Construction method of soil-rock mixture tunnel three-dimensional digital information management system
CN115131486B (en) Engineering exploration data acquisition system and method
CN115019294A (en) Pointer instrument reading identification method and system
CN117313506B (en) Tunnel surrounding rock topology safety coefficient and stability calculation method
CN114324580A (en) Intelligent knocking detection method and system for structural defects
CN111779477B (en) Fractal theory-based dynamic evaluation method for complexity of hydraulic fracture
CN116597193A (en) Tunnel intelligent construction-oriented tunnel face unmanned geological logging method and system
CN111462037A (en) Geological exploration drilling quality detection method
CN116559001A (en) Quantitative evaluation method for microcosmic compressibility of tight sandstone reservoir
CN115357964A (en) Complex structural rock mass structure and accurate acquisition method of anisotropic characteristics thereof
CN114492914A (en) Rock mass characterization unit volume value prediction method and system based on machine learning
CN114091347A (en) Electronic component service life prediction method
CN113158558A (en) High-speed railway roadbed continuous compaction analysis method, device and analyzer

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination