Rock damage identification methods and systems, electronic equipment

Abstract

This invention relates to the field of machine learning, providing a method, system, and electronic device for rock damage identification. The method includes: acquiring multi-stage damage data of a rock under test, wherein the multi-stage damage data is acquired through a piezoelectric sensing element disposed on or inside the surface of the rock under test; constructing an original dataset based on the multi-stage damage data; and inputting the original dataset into a pre-constructed convolutional neural network model to obtain the damage identification result of the rock under test. This invention addresses the shortcomings of related technologies, such as the easy loss of key damage details like local spectral deformation and peak shift, enabling real-time, non-destructive, and continuous monitoring of rock damage. It automatically extracts damage features and accurately classifies damage at each cycle stage from initial loading to final failure, providing reliable technical support for early warning of geotechnical engineering disasters.

Description

Technical Field

[0001] This invention relates to the field of machine learning technology, and in particular to a rock damage identification method and system, and electronic equipment. Background Technology

[0002] As the core load-bearing medium in geological structures and various geotechnical engineering projects, rock is susceptible to irreversible cumulative damage during long-term service due to loads such as traffic vibration, earthquakes, and periodic mining disturbances. This damage can lead to major engineering disasters such as surrounding rock collapse, slope instability, and tunnel deformation. Therefore, accurately capturing the damage state of rock at each stage of cyclic loading and achieving real-time dynamic monitoring of the damage evolution process is of great significance for the safety assessment of engineering structures and disaster early warning.

[0003] Mechanical electrical impedance tomography (EMI) technology has been widely used in structural health monitoring due to its advantages such as high operating frequency, sensitivity to localized minor damage, high sensitivity, low energy consumption, and good stability. Its applications have now expanded to aerospace, oil and gas infrastructure, and civil engineering structures. However, traditional EMI data analysis heavily relies on manual feature extraction and interpretation, such as statistical indicators like root mean square deviation (RMSD) and covariance (Cov). These methods compress high-dimensional spectral information into a single scalar, easily losing key damage details such as local spectral deformation and peak shifts. Furthermore, they struggle to handle the nonlinear and non-stationary signal responses caused by damage-stress coupling in heterogeneous materials like rock under complex stress paths. Summary of the Invention

[0004] This invention provides a rock damage identification method, system, and electronic device to address the shortcomings of related technologies, such as the easy loss of key damage details like local spectral deformation and peak shift. It enables real-time, non-destructive, and continuous monitoring of rock damage, automatically extracts damage features, and accurately classifies damage at each cycle stage from initial loading to final failure, providing reliable technical support for early warning of geotechnical engineering disasters.

[0005] This invention provides a method for identifying rock damage, comprising: Multi-stage damage data of the rock under test is acquired by a piezoelectric sensing element disposed on or inside the rock under test. The original dataset is constructed based on the multi-stage damage data; The original dataset is input into a pre-built convolutional neural network model to obtain the damage identification results of the rock under test.

[0006] According to the rock damage identification method provided by the present invention, the method for acquiring multi-stage damage data includes: The process of the rock under test from initial loading to final failure is monitored; Collect the real part data of the admittance of the piezoelectric sensing element within a preset frequency sweep range.

[0007] According to the rock damage identification method provided by the present invention, the acquisition of the real part data of the admittance of the piezoelectric sensing element within a preset frequency sweep range includes: Measure the mechanical impedance signal of the piezoelectric sensing element; Extract the data of the frequency band where the first resonant peak of the mechanical impedance signal is located.

[0008] According to the rock damage identification method provided by the present invention, after constructing the original dataset based on the multi-stage damage data, and before inputting the original dataset into the pre-built convolutional neural network model, the method further includes: The one-dimensional admittance real part data sequence of the original dataset is reconstructed into a two-dimensional matrix form.

[0009] According to the rock damage identification method provided by the present invention, after constructing the original dataset based on the multi-stage damage data, and before reconstructing the one-dimensional admittance real part data sequence of the original dataset into a two-dimensional matrix form, the method further includes: The original dataset is then cleaned and normalized.

[0010] According to the rock damage identification method provided by the present invention, the convolutional neural network model includes a convolutional layer, a pooling layer, a fully connected layer, and an output layer; The convolutional layer includes a first convolutional layer, a second convolutional layer, and a third convolutional layer; The fully connected layer is used to regularize the input feature map; The output layer is used to transform the input feature map into a class probability distribution.

[0011] According to the rock damage identification method provided by the present invention, the first convolutional layer uses 32 3×3 convolutional kernels; The second convolutional layer uses 64 3×3 convolutional kernels; The third convolutional layer uses 128 3×3 convolutional kernels.

[0012] The present invention also provides a rock damage identification system, which applies a rock damage identification method, including: The data acquisition module is used to acquire multi-stage damage data of the rock under test, which is acquired by a piezoelectric sensing element disposed on the surface or inside the rock under test. The data processing module is used to construct the original dataset based on the multi-stage damage data; The damage identification module is used to input the original dataset into a pre-built convolutional neural network model to obtain the damage identification results of the rock to be tested.

[0013] The present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement any of the rock damage identification methods described above.

[0014] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements any of the rock damage identification methods described above.

[0015] The present invention also provides a computer program product, including a computer program that, when executed by a processor, implements any of the rock damage identification methods described above.

[0016] The rock damage identification method provided in this application has at least the following beneficial effects: High sensitivity and intelligence: It combines the advantages of mechatronics technology in being sensitive to local micro-damage with the powerful automatic feature extraction and classification capabilities of convolutional neural networks, and achieves accurate and intelligent identification of rock damage status.

[0017] The system boasts strong robustness: the piezoelectric sensing element is encapsulated in epoxy resin, enhancing its durability and anti-interference capabilities in complex soil and rock environments. LoRa wireless transmission technology ensures reliable data transmission over long distances and in complex terrains.

[0018] Full utilization of data: By reconstructing one-dimensional admittance data into a two-dimensional matrix, the advantages of two-dimensional convolutional neural networks in extracting spatial features are fully utilized, avoiding the information loss problem caused by data compression in traditional statistical indicators.

[0019] High degree of automation: From data acquisition and transmission to damage identification, the entire process can be automated, reducing human intervention and providing reliable technical support for long-term health monitoring and disaster early warning in geotechnical engineering. Attached Figure Description

[0020] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0021] Figure 1 This is one of the flowcharts of the rock damage identification method provided in the embodiments of the present invention; Figure 2 This is a second schematic flowchart of the rock damage identification method provided in the embodiments of the present invention; Figure 3This is one of the structural schematic diagrams of the rock damage identification system provided in the embodiments of the present invention; Figure 4 This is a second schematic diagram of the rock damage identification system provided in this embodiment of the invention; Figure 5 This is a schematic diagram of the physical structure of the electronic device provided in an embodiment of the present invention. Detailed Implementation

[0022] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0023] Figure 1 This is one of the flowcharts of the rock damage identification method provided in the embodiments of the present invention.

[0024] like Figure 1 As shown, this embodiment provides a rock damage identification method, including: Step 101: Obtain multi-stage damage data of the rock to be tested. The multi-stage damage data is obtained by a piezoelectric sensing element set on the surface or inside the rock to be tested. In practice, acquiring multi-stage damage data can measure the mechanical impedance signal of the piezoelectric sensing element; and extract data of the frequency band where the first resonant peak of the mechanical impedance signal is located.

[0025] Step 102: Construct the original dataset based on the multi-stage damage data; In implementation, after step 102 and before step 103, the following may also be included: The original dataset is cleaned and normalized. The one-dimensional admittance real part data sequence of the original dataset is reconstructed into a two-dimensional matrix form.

[0026] Step 103: Input the original dataset into the pre-built convolutional neural network model to obtain the damage identification result of the rock to be tested.

[0027] Figure 2 This is the second flowchart of the rock damage identification method provided in the embodiments of the present invention.

[0028] In a specific embodiment, such as Figure 2 As shown, the rock damage identification method provided in this application includes: In the rock area to be tested, piezoelectric sensing elements are installed on the rock surface or in pre-drilled holes at intervals of 50-100cm, ensuring close contact between the sensing elements and the rock. The power supply module, impedance acquisition module, and wireless transmission module are connected, and system debugging is completed to ensure normal communication and stable data acquisition. The impedance acquisition module collects real admittance data at different loading stages according to preset parameters. The wireless transmission module uploads the collected data to the cloud server in real time, forming the raw dataset. Based on the real admittance data received from the cloud server, it is divided into training, validation, and test sets in a 70%:15%:15% ratio. After standardizing the real admittance data using StandardScaler with zero mean and unit variance, it is reshaped into an 11×11 two-dimensional matrix, constructing a 2D-CNN model with three convolutional layers. The model includes three consecutive convolutional layers, a pooling layer, a fully connected layer, and an output layer. The first convolutional layer uses 32 3×3 convolutional kernels, which, after batch normalization, ReLU activation, and 2×2 max pooling, output a 5×5×32 feature map. The second convolutional layer uses 64 3×3 convolutional kernels, which, after batch normalization, ReLU activation, and 2×2 max pooling, output a 2×2×64 feature map. The third convolutional layer uses 128 3×3 convolutional kernels, which, after batch normalization, ReLU activation, and adaptive average pooling, output a 1×1×128 feature map. The flattened feature maps are then input into two fully connected layers with dropout regularization rates of 0.5 and 0.3, respectively. The output layer is converted into a class probability distribution using the Softmax function. The computational processing system receives the new real part data of admittance from the wireless transmission module in real time. After standardization and two-dimensional matrix reconstruction, the data is input into the trained 2D-CNN model. The model automatically extracts damage features and outputs the damage category, achieving accurate damage identification in each cycle stage of rock failure from initial loading to final destruction.

[0029] The rock damage identification system provided by the present invention is described below. The rock damage identification system described below can be referred to in correspondence with the rock damage identification method described above.

[0030] Figure 3 This is one of the structural schematic diagrams of the rock damage identification system provided in the embodiments of the present invention.

[0031] like Figure 3 As shown, the rock damage identification system provided in this embodiment includes: The data acquisition module 301 is used to acquire multi-stage damage data of the rock to be tested, which is acquired by a piezoelectric sensing element disposed on the surface or inside the rock to be tested. Data processing module 302 is used to construct an original dataset based on the multi-stage damage data; The damage identification module 303 is used to input the original dataset into a pre-built convolutional neural network model to obtain the damage identification result of the rock to be tested.

[0032] Figure 4 This is the second structural schematic diagram of the rock damage identification system provided in the embodiments of the present invention.

[0033] In a specific embodiment, such as Figure 4 As shown, the rock damage identification system may include a power supply module, a piezoelectric sensing element module, an impedance acquisition module, a wireless transmission module, and a computing processing system. The piezoelectric sensing element module includes multiple spaced PZT-5H type piezoelectric ceramic sensing elements, each of which is wrapped with an epoxy resin protective layer. The power supply module outputs a stable DC voltage. The impedance acquisition module first pre-scans the frequency and then focuses on the frequency band where the first resonant peak is located to acquire the real part of the admittance data. The wireless transmission module is LoRa type and the communication frequency is 433MHz. The computing processing system integrates a temperature compensation module, a CNN damage identification model, and a dataset processing algorithm.

[0034] Among them, the piezoelectric sensing element can be attached to the rock surface or buried inside the rock. The impedance analyzer is used to measure the mechanical impedance signal of the piezoelectric sensing element, extract the frequency band data of the first resonance peak, construct a dataset containing multiple damage stages, and use a convolutional neural network model to identify the rock damage state. The real part data of the admittance in the mechanical impedance signal is collected and reconstructed into a two-dimensional matrix. A convolutional neural network model is designed, and the two-dimensional matrix data is input into the convolutional neural network to automatically extract damage features and realize accurate classification of damage in each cycle stage of rock from initial loading to final failure.

[0035] In practice, the piezoelectric ceramic sensing element has dimensions of 10mm×10mm×1mm, the epoxy resin protective layer has a thickness of 2mm, and the spacing between adjacent sensing elements is 50-100cm.

[0036] In practice, the power supply module is a lithium battery converted to a stable voltage by a DC-DC converter, and is equipped with overvoltage and overcurrent protection modules.

[0037] In implementation, the computing system includes a cloud server and a local terminal. The cloud server is equipped with a 64-bit processor and a GPU acceleration unit. The local terminal can obtain data and recognition results by accessing the cloud. The temperature compensation module includes a feature extraction network and a decoding network, which are used to filter the interference of temperature on the admittance signal.

[0038] Furthermore, the application method of the rock damage identification system provided in this embodiment may include: Step 1, System Deployment and Debugging: Deploy multiple piezoelectric sensing elements, each coated with an epoxy resin protective layer, at preset intervals on the surface of the rock to be tested or in internal boreholes, ensuring tight coupling with the rock; connect the power supply module, impedance acquisition module, and wireless transmission module to complete system debugging.

[0039] Step 2, Multi-stage damage data acquisition: The entire cyclic loading process of the rock from initial loading to final failure is monitored through the impedance acquisition module. At each loading stage, the real part data of the admittance of the piezoelectric sensing element within the preset frequency sweep range is acquired. All acquired data is uploaded to the computing and processing system in real time through the wireless transmission module to construct an original dataset containing multiple damage stages.

[0040] Step 3, Data Preprocessing and Dataset Partitioning: Clean and normalize the real impedance or real admittance data in the original dataset; divide the processed data into training set, validation set and test set according to the proportion.

[0041] Step 4, Two-dimensional data reconstruction: The one-dimensional admittance real part data sequence of each sample is reconstructed into a two-dimensional matrix form to adapt to the input requirements of the two-dimensional convolutional neural network.

[0042] Step 5: Build and train a convolutional neural network model: Design a two-dimensional convolutional neural network model containing multiple convolutional layers, pooling layers, and fully connected layers; train the model using the training set, and adjust the model hyperparameters using the validation set until the model converges, thus obtaining a trained damage recognition model.

[0043] Step 6, Online rock damage identification: The real part data of admittance, which is collected in real time and preprocessed and reconstructed in two dimensions, is input into the trained convolutional neural network model. The model automatically outputs the current damage state category of the rock, realizing accurate identification of damage.

[0044] The specific implementation method of the rock damage identification system provided in this embodiment can be implemented with reference to the above embodiment, and will not be repeated here.

[0045] Figure 5 An example is a schematic diagram of the physical structure of an electronic device, such as... Figure 5 As shown, the electronic device may include: a processor 510, a communication interface 520, a memory 530, and a communication bus 540, wherein the processor 510, the communication interface 520, and the memory 530 communicate with each other via the communication bus 540. The processor 510 can call logical instructions in the memory 530 to execute a rock damage identification method, which includes: Multi-stage damage data of the rock under test is acquired by a piezoelectric sensing element disposed on or inside the rock under test. The original dataset is constructed based on the multi-stage damage data; The original dataset is input into a pre-built convolutional neural network model to obtain the damage identification results of the rock under test. Furthermore, the logical instructions in the aforementioned memory 530 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0046] On the other hand, the present invention also provides a computer program product, which includes a computer program that can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer is able to execute the rock damage identification method provided by the above methods, the method comprising: Multi-stage damage data of the rock under test is acquired by a piezoelectric sensing element disposed on or inside the rock under test. The original dataset is constructed based on the multi-stage damage data; The original dataset is input into a pre-built convolutional neural network model to obtain the damage identification results of the rock under test. In another aspect, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, is implemented to perform the rock damage identification method provided by the methods described above, the method comprising: Multi-stage damage data of the rock under test is acquired by a piezoelectric sensing element disposed on or inside the rock under test. The original dataset is constructed based on the multi-stage damage data; The original dataset is input into a pre-built convolutional neural network model to obtain the damage identification results of the rock under test. The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0047] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., including several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods of various embodiments or some parts of embodiments.

[0048] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for identifying rock damage, characterized in that, include: Multi-stage damage data of the rock under test is acquired by a piezoelectric sensing element disposed on or inside the rock under test. The original dataset is constructed based on the multi-stage damage data; The original dataset is input into a pre-built convolutional neural network model to obtain the damage identification results of the rock under test.

2. The rock damage identification method according to claim 1, characterized in that, The method for acquiring multi-stage damage data includes: The process of the rock under test from initial loading to final failure is monitored; Collect the real part data of the admittance of the piezoelectric sensing element within a preset frequency sweep range.

3. The rock damage identification method according to claim 2, characterized in that, The acquisition of the real part of the admittance data of the piezoelectric sensing element within a preset frequency sweep range includes: Measure the mechanical impedance signal of the piezoelectric sensing element; Extract the data of the frequency band where the first resonant peak of the mechanical impedance signal is located.

4. The rock damage identification method according to claim 1, characterized in that, After constructing the original dataset based on the multi-stage damage data, the step of inputting the original dataset into the pre-built convolutional neural network model further includes: The one-dimensional admittance real part data sequence of the original dataset is reconstructed into a two-dimensional matrix form.

5. The rock damage identification method according to claim 4, characterized in that, After constructing the original dataset based on the multi-stage damage data, and before reconstructing the one-dimensional admittance real part data sequence of the original dataset into a two-dimensional matrix form, the method further includes: The original dataset is then cleaned and normalized.

6. The rock damage identification method according to claim 1, characterized in that, The convolutional neural network model includes convolutional layers, pooling layers, fully connected layers, and an output layer; The convolutional layer includes a first convolutional layer, a second convolutional layer, and a third convolutional layer; The fully connected layer is used to regularize the input feature map; The output layer is used to transform the input feature map into a class probability distribution.

7. The rock damage identification method according to claim 6, characterized in that, The first convolutional layer uses 32 3×3 convolutional kernels; The second convolutional layer uses 64 3×3 convolutional kernels; The third convolutional layer uses 128 3×3 convolutional kernels.

8. A rock damage identification system, employing the rock damage identification method according to any one of claims 1-7, characterized in that, include: The data acquisition module is used to acquire multi-stage damage data of the rock under test, which is acquired by a piezoelectric sensing element disposed on the surface or inside the rock under test. The data processing module is used to construct the original dataset based on the multi-stage damage data; The damage identification module is used to input the original dataset into a pre-built convolutional neural network model to obtain the damage identification results of the rock to be tested.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the rock damage identification method as described in any one of claims 1-7.

10. A non-transitory computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by a processor, it implements the rock damage identification method as described in any one of claims 1-7.

Images