Method for measuring deformation field in whole process of damage evolution and crack propagation of dual-optical-path nuclear graphite

Abstract

A dual-optical-path nuclear graphite damage evolution and crack propagation whole-process deformation field measurement method comprises the following steps: constructing a fracture triggering mechanism, setting a low-speed high-resolution camera to continuously collect in a whole period, and capturing a crack image through a high-speed low-resolution camera when a triggering condition is met; a global high-precision deformation field of the surface of the test piece in a full loading period is solved through a digital image correlation method (PINN-DIC) based on a physical information neural network, a displacement field of the surface of the test piece in a high-speed fracture process is solved through the PINN-DIC while a damage feature region is identified, and crack core parameters are extracted; and in combination with strain distribution evolution characteristics and crack initiation positions, high-speed fracture information is mapped to a main coordinate system, and a cross-scale correlation mechanism is established. According to the method, accurate measurement and cross-scale data communication of the nuclear graphite damage evolution-fracture failure complete-period deformation field are realized, the deformation field measurement accuracy of a boundary region and a crack neighborhood is remarkably improved, and a key technical support is provided for analysis of a nuclear graphite fracture mechanism.

Description

Technical Field

[0001] This invention relates to a technology in the field of optical damage fracture testing, specifically a method for measuring the deformation field of the entire process of damage evolution and crack propagation in dual-path graphite. Background Technology

[0002] As a key structural material in advanced nuclear energy systems, nuclear graphite's damage evolution and fracture failure directly impact reactor operational safety. Ensuring long-term stable reactor operation requires precise understanding of the entire lifecycle behavior of nuclear graphite, from damage initiation and accumulation to fracture failure. The analysis of this behavior heavily relies on reliable observation technologies. However, existing observation technologies have significant limitations: they lack full-lifecycle observation capabilities across time scales and struggle to integrate crucial data on progressive damage and instantaneous fracture. This hinders in-depth analysis of the damage-fracture correlation mechanism in nuclear graphite, necessitating an observation system that combines full-lifecycle coverage with high-precision measurement. Summary of the Invention

[0003] This invention addresses the shortcomings of existing technologies, such as the lack of collaborative analysis and correlation of dual time data sources, the inability to simultaneously cover the complete observation requirements of long-term damage evolution and instantaneous crack propagation, the inconsistency of image data benchmarks acquired by different cameras, and the insufficient measurement resolution and spatial resolution of deformation fields in damage evolution and crack propagation. It proposes a dual-optical-path core graphite damage evolution and crack propagation deformation field measurement method, which significantly improves the measurement accuracy of deformation fields in boundary regions and crack neighborhoods.

[0004] This invention is achieved through the following technical solution:

[0005] This invention relates to a method for measuring the deformation field of the entire process of damage evolution and crack propagation in dual-optical-path nuclear graphite, comprising:

[0006] Step 1: Adjust the camera working distance, select the lens focal length, and use the cross laser to assist in calibrating the spatial position of the dual optical paths;

[0007] Step 2: Calibrate using a checkerboard calibration board and solve for the dual-camera coordinate system transformation matrix;

[0008] Step 3: Establish a background grayscale reference, calibrate the trigger threshold, and construct a fracture triggering mechanism for a high-speed, low-resolution camera.

[0009] Step 4: Continuously acquire data throughout the entire cycle using a low-speed, high-resolution camera; when the triggering conditions are met, capture crack images using a high-speed, low-resolution camera.

[0010] Step 5: Solve the high-precision deformation field of the specimen surface throughout the entire loading cycle using the PINN-DIC (Physical Information Neural Network) method. While identifying the damage feature area, solve the displacement field of the specimen surface during the high-speed fracture process using PINN-DIC to extract the crack core parameters.

[0011] Step 6: Combining the strain distribution evolution characteristics with the crack initiation location, map the high-speed fracture information to the main coordinate system and establish a cross-scale correlation mechanism.

[0012] This invention relates to a dual-path graphite damage evolution and crack propagation deformation field measurement system for implementing the above-mentioned method, comprising: a spectroscopic imaging module, a trigger adaptation module, a spatial calibration module, and a deformation analysis module. The spectroscopic imaging module uses a low-speed, high-resolution camera for long-term full-domain deformation observation to completely record the damage evolution and strain concentration process of the specimen from loading to fracture, and uses a high-speed, low-resolution camera for short-term, high-dynamic capture to accurately record the instantaneous dynamic behavior of crack initiation and propagation. The trigger adaptation module monitors the characteristic grayscale abrupt changes and connected domain morphological changes caused by specimen fracture, and uses a post-triggered method to activate the high-speed, low-resolution camera of the spectroscopic imaging module to acquire images. The calibration module completes the spatial deviation calibration of the dual optical paths through spatial coordinate transformation, realizing the spatial matching of the instantaneous fracture area captured by the high-speed low-resolution camera and the long-term deformation field recorded by the low-speed high-resolution camera, laying the foundation for cross-timescale data fusion. The deformation analysis module uses the physical information neural network-driven digital image correlation method (PINN-DIC) to realize the deformation field analysis of the entire process of damage evolution and crack propagation. Based on the long-term deformation field information measured by the low-speed high-resolution camera, it predicts the potential crack propagation area through the spatial distribution characteristics of deformation, while fusing the crack propagation deformation field measured by the high-speed low-resolution camera and associating long-term damage accumulation with instantaneous fracture failure.

[0013] Technical effect

[0014] This invention employs a 7:3 split-ratio dual-optical-path collaborative design, constructing a collaborative observation architecture between a high spatial resolution camera (for long-term, high-precision recording of damage evolution) and a high temporal resolution camera (for short-term, high-frame-rate capture of instantaneous fracture dynamics). It incorporates a post-trigger mechanism of "real-time grayscale monitoring + connected component verification," enabling the high-speed, low-resolution camera to only initiate acquisition when a fracture event occurs, eliminating the need for continuous recording and significantly reducing storage resource consumption. Through spatial coordinate transformation and data fusion technologies, a primary-secondary coordinate system transformation relationship is established, achieving precise alignment of the dual-optical-path observation areas and laying a spatial benchmark for cross-temporal-scale data fusion. The three components work together to form an observation mode of "on-demand division of labor + precise triggering + spatial alignment," ensuring both the integrity of loading full-cycle damage evolution data and high-frame-rate precise capture of instantaneous fracture processes. This achieves cross-scale data continuity from "long-term damage accumulation to instantaneous fracture failure," providing reliable support for the acquisition of deformation information throughout the entire process. In the deformation analysis stage, physical information neural network digital image correlation technology is employed. On the one hand, it accurately processes long-term deformation images captured by low-speed, high-resolution cameras, effectively addressing the technical bottleneck of insufficient accuracy in deformation calculation in boundary regions and achieving high-precision quantification of the global strain field. On the other hand, for instantaneous dynamic crack images acquired by high-speed, low-resolution cameras, it accurately solves for local displacement fields, analyzes core fracture parameters, and avoids the calculation distortion problem in local crack regions caused by traditional methods. Combined with dual-camera spatial alignment technology, spatial consistency matching of the two types of observation data is achieved, significantly improving the reliability and correlation of cross-scale observation data and providing high-precision data support for in-depth analysis of the fracture mechanism of nuclear graphite. Attached Figure Description

[0015] Figure 1 This is a schematic diagram of the system of the present invention;

[0016] Figure 2 This is a flowchart of the present invention;

[0017] Figure 3 This is a schematic diagram of the optical path (including the packaging container) for the dual-camera spectroscopic observation of the specimen.

[0018] Figure 4 This is a diagram illustrating the camera's working distance.

[0019] Figure 5 Schematic diagram of the auxiliary optical path for a laser level;

[0020] Figure 6 A schematic diagram of camera space matching;

[0021] Figure 7 The image shows the x-direction normal strain field contour map of the loading process measured by a low-speed, high-resolution camera in the embodiment.

[0022] Figure 8The image shows the displacement field cloud map of the fracture process measured by a high-speed, low-resolution camera in the embodiment. Detailed Implementation

[0023] like Figure 1 As shown, this embodiment relates to a dual-optical-path graphite damage evolution and crack propagation deformation field measurement system, which includes: a spectroscopic imaging module, a triggering and adaptation module, a spatial calibration module, and a deformation analysis module.

[0024] The aforementioned beam-splitting imaging module includes: a low-speed high-resolution camera, a high-speed low-resolution camera, and a cubic beam-splitting prism, wherein: the cubic beam-splitting prism is positioned directly in front of the nuclear graphite specimen, uniformly splitting the reflected light from the surface of the nuclear graphite specimen into two beams and guiding them to the low-speed high-resolution camera and the high-speed low-resolution camera, respectively.

[0025] The beam-splitting characteristics of the cubic beam-splitting prism are precisely matched with the dual-camera function. Preferably, 30% of the beam is allocated to the low-speed, high-resolution camera, which not only meets the stable lighting requirements for long-term shooting, but also suppresses imaging noise through appropriate light flux, ensuring spatial resolution. Meanwhile, 70% of the beam is allocated to the high-speed, low-resolution camera, providing sufficient light flux for the short-exposure requirements of high-frame-rate shooting, effectively avoiding grayscale signal distortion caused by underexposure, and reducing the risk of triggering misjudgment.

[0026] The trigger adaptation module includes a high-speed low-resolution camera and a trigger board. The trigger board receives real-time grayscale image data transmitted by the high-speed low-resolution camera, performs pixel-by-pixel difference operation and connected component verification according to preset logic, generates a grayscale difference matrix and compares it with the trigger threshold. When the trigger condition is met, the trigger board sends a start acquisition signal to the high-speed low-resolution camera to realize trigger acquisition control.

[0027] The spatial calibration module includes a low-speed, high-resolution camera, a high-speed, low-resolution camera, a cubic beam splitter, and a checkerboard calibration plate. The cubic beam splitter is fixed directly in front of the nuclear graphite specimen. The low-speed, high-resolution camera and the high-speed, low-resolution camera are positioned perpendicular to the optical path of the cubic beam splitter, forming a dual-optical-path observation architecture. The checkerboard calibration plate is placed near the specimen surface and within the shared effective field of view of both cameras. Using the image pixel coordinate system acquired by the low-speed, high-resolution camera as the primary coordinate system and the image pixel coordinate system acquired by the high-speed, low-resolution camera as the secondary coordinate system, corner coordinates are extracted using a corner detection algorithm. The primary-secondary coordinate system transformation matrix is ​​then solved to complete the dual-optical-path spatial deviation calibration, providing a foundation for spatial alignment of data across time scales.

[0028] The deformation analysis module includes a control unit and a digital image correlation unit. The control unit establishes data transmission links with a low-speed high-resolution camera and a high-speed low-resolution camera to synchronously control the acquisition parameters of the two types of cameras. It also stores the acquired full-cycle deformation images and instantaneous crack images according to timestamps and outputs them to the digital image correlation unit. The digital image correlation unit performs full-domain high-precision strain field analysis and damage feature region identification on the full-cycle loading images acquired by the low-speed high-resolution camera to quantify the nuclear graphite damage accumulation process. It also performs local displacement field solution and crack core parameter extraction on the instantaneous crack images acquired by the high-speed low-resolution camera to characterize the failure fracture dynamic behavior.

[0029] like Figure 2 As shown, this embodiment illustrates the method for measuring the deformation field of the entire process of damage evolution and crack propagation in dual-optical-path core graphite based on the aforementioned system, including:

[0030] S1, Build as follows Figure 3 The spectral imaging module shown consists of a cubic beam splitter prism and a low-speed, high-resolution camera placed sequentially in front of the nuclear graphite specimen to be observed, forming the first optical path. At the same time, a high-speed, low-resolution camera is placed perpendicular to the cubic beam splitter prism and the first optical path, forming the second optical path.

[0031] like Figure 4 As shown, the size of the surface target observation area (AOI) of the nuclear graphite specimen to be observed is... and the target surface size of the two cameras satisfy: To obtain the magnification Then, based on the camera's working distance Image distance and lens focal length The relationship, that is ,get .

[0032] like Figure 5 As shown, preferably, when adjusting the spatial position of the two cameras and the cubic beam splitter, a laser level is used in three directions to emit cross lasers to the preset optical path to assist in calibrating the spatial attitude of the cameras and the cubic beam splitter, ensuring the coaxiality and perpendicularity of the two optical paths.

[0033] like Figure 3 As shown, the cubic beam splitter packaging container and light shield are made of black reinforced carbon fiber material. Pre-reserved holes adapted to the lens diameter are opened on the sides facing the two cameras and extend outward to form an optical path channel, which effectively blocks the intrusion of ambient stray light and ensures the stability of imaging signal-to-noise ratio and grayscale signal.

[0034] S2. The spatial coordinates of images acquired by low-speed high-resolution cameras and high-speed low-resolution cameras are unified: such as... Figure 6 As shown, the pixel coordinate system of the low-speed, high-resolution camera image is used as the main coordinate system. The pixel coordinate system of a high-speed, low-resolution camera image is a secondary coordinate system. The checkerboard calibration board was placed within the shared effective field of view of both cameras. The Harris corner detection algorithm was used to extract the corner coordinates from the primary and secondary coordinate system images. Specifically: Where: the coordinates of the corner points in the principal coordinate system are ( The coordinates of the corresponding corner point in the secondary coordinate system are ( ), , , , For quantities related to rotation and scaling, , This represents the translation amount.

[0035] S3. Construct a high-speed, low-resolution camera triggering mechanism, specifically including:

[0036] 3.1 Based on the static image of the graphite before loading, a background grayscale reference matrix is ​​established by averaging multiple images over time. Specifically, images were collected before loading. A background grayscale reference matrix is ​​constructed from a frame of still image using a point-by-point pixel temporal averaging method. ,in: For the first The grayscale matrix of a frame of still image. The corresponding pixel row and column coordinates of the image. In the reference matrix The grayscale value of the pixel at the location.

[0037] 3.2 During the experiment, a real-time grayscale difference matrix was generated through pixel-by-pixel difference operations. ,in: For high-speed low-resolution cameras The image grayscale moments are acquired in real time; in the initial loading stage, the graphite nucleus only experiences slight surface displacement with no significant morphological changes, and the grayscale distribution fluctuates gently. The absolute values ​​are small and all are within the low-level threshold range, so the high-speed, low-resolution camera remains in a ready-to-trigger state.

[0038] 3.3 After the crack initiates and propagates, the grayscale of the specimen surface undergoes a sudden change, and the crack area exhibits a grayscale characteristic close to a pure black background, leading to... Significantly negative, and with an absolute value greatly exceeding the low-level threshold, it forms a characteristic trigger signal characterizing fracture behavior. When continuous... The percentage of pixels exceeding a preset high-level threshold in the frame difference matrix. Effective proportion of monitoring area When the trigger path switches to a stable high level, the high-speed, low-resolution camera immediately starts acquiring data.

[0039] S4. When the grayscale change meets the trigger condition in S3, the high-speed low-resolution camera is activated at a frame rate. The continuous data collection duration meets the following requirements: as well as ,in: To estimate the maximum crack length, To estimate the maximum crack propagation rate, The minimum number of effective image frames required to characterize the entire crack propagation process.

[0040] S5. Deformation field analysis of low-speed, high-resolution images: For the full-cycle deformation images acquired by the low-speed, high-resolution camera in steps S3 and S4, the full-domain deformation field is calculated and damage features are identified using the Physical Information Neural Network-based Digital Image Correlation (PINN-DIC) method; For the instantaneous crack images acquired by the high-speed, low-resolution camera in step S4, the Physical Information Neural Network-driven Digital Image Correlation (PINN-DIC) method is used for refined analysis.

[0041] The aforementioned global deformation field calculation and damage feature identification refers to: inputting the image coordinate domain into a Physical Information Neural Network (PINN) to predict the displacement field, which is then used to predict the deformed image. Compared with actual deformed images The mean squared error is used as the loss function, specifically: Training is achieved by iteratively optimizing parameters through a fully connected network and minimizing the loss function, where: This represents the total number of pixels involved in the calculation. This process addresses the distortion problem in traditional DIC calculations at specimen boundaries; leveraging the automatic differentiation function of a deep learning framework, it calculates the global displacement field output by PINN. Performing differential operations requires no additional numerical difference processing; the result is derived directly. , Directional strain and shear strain By analyzing the spatial distribution and temporal evolution characteristics of the components, the strain concentration area is defined as the damage characteristic area of ​​nuclear graphite, thus clarifying the spatial location and development trend of damage accumulation.

[0042] The predicted deformed image ,in: For reference image, The PINN predicted displacement field is reconstructed using the Dirac interpolation algorithm. The pixel coordinates of the reference image, This is the Dirac function.

[0043] The physical information-based neural network described above employs a fully connected neural network with an input layer containing 2 neurons, a hidden layer containing 3 layers with 50 neurons each, and an output layer containing 2 neurons.

[0044] The refined analysis of PINN-DIC involves inputting the coordinate domain of a high-speed, low-resolution image into a fully connected network to predict the corresponding local displacement field. This fully connected network uses a gray-level consistency assumption as its loss function, and training is achieved through iterative parameter optimization until the loss function converges. This process effectively avoids the information loss problem in the crack region inherent in traditional DIC, accurately extracting core fracture parameters such as crack tip location, propagation trajectory, and local deformation features, providing instantaneous fracture feature data for subsequent correlation mechanism analysis.

[0045] S6. Correlation Mechanism Analysis: Based on the principal-sub-coordinate system transformation matrix established in S2, the crack initiation coordinates and propagation trajectory extracted by PINN-DIC through refined analysis are accurately mapped to the strain field coordinate system for global deformation field measurement and damage feature identification. This achieves spatial alignment of dual-optical-path data, clarifies the correspondence between crack initiation and propagation locations and the early strain concentration areas in low-speed, high-resolution images, and finally establishes the correlation feature of "strain concentration accumulation - crack initiation and propagation," providing experimental evidence for the cross-scale mechanism analysis of "long-term damage evolution - instantaneous fracture failure" in nuclear graphite.

[0046] After specific experiments, 139mm was selected. 10mm An 11mm thick nuclear graphite beam specimen was subjected to a load of 0.005mm / s on a small fatigue testing machine. A high-speed, low-resolution Photron Fast SA 2.1 camera with a 200mm telephoto lens was used, capturing images at a frame rate of 100,000fps with an image resolution of 384. 264 pixels; The selected low-speed, high-resolution camera was a Daheng MER-160-227U3M camera equipped with a 50mm fixed-focus lens. The camera's frame rate was 1fps, and the image resolution was 1440. 1080 pixels; The selected cubic beam splitter is a 50.0mm cubic beam splitter made of 70R / 30T VIS fine-annealed K9 optical glass.

[0047] The aforementioned dual-optical-path method for measuring the deformation field of nuclear graphite damage evolution and crack propagation throughout the entire process was used to capture complete strain field information of the nuclear graphite specimen during loading and deformation, as well as the crack propagation process at specimen failure and fracture. The dual-scale data was processed using the PINN-DIC algorithm, correlating "long-term strain accumulation" and "instantaneous crack propagation" information. This verified the effectiveness of the experimental system in cross-scale deformation field measurement and data correlation. Figure 7 As shown, this is the strain field data of the entire damage evolution process of nuclear graphite obtained by a low-speed, high-resolution camera; Figure 8 The image shows the instantaneous fracture displacement field data of nuclear graphite obtained by a high-speed, low-resolution camera. Based on the strain field evolution law and using the strain field gradient as a reference, it is determined that cracks will appear... Figure 7 At the bottom 25mm position, with Figure 8 The location of the crack at 25mm is consistent with the position shown, indicating that the present invention has successfully achieved deformation field measurement and deformation field data correlation of the entire process of nuclear graphite damage evolution and crack propagation.

[0048] In this embodiment, the AOI area width is 10mm to allow for redundancy in case of deformation and crack propagation, and the actual observation field of view is set to 20mm; the selected low-speed high-resolution camera has a pixel size of 3.45µm and an image resolution of 1440. 1080 pixels, magnification The selected high-speed, low-resolution camera has a pixel size of 20µm and an image resolution of 384. 264 pixels, magnification The working distance of the low-speed, high-resolution camera was measured using a laser rangefinder. mm, working distance of high-speed low-resolution camera For low-speed, high-resolution cameras, a 50mm fixed-focus lens is used, while for high-speed, low-resolution cameras, a 200mm telephoto lens is used.

[0049] like Figure 5 As shown, a laser level was used to assist in adjusting the positions of the high-speed low-resolution camera, the low-speed high-resolution camera, and the cubic beam splitter prism, forming two approximately perpendicular optical paths. Furthermore, a light-shielding tube was constructed using black reinforced carbon fiber material. Figure 3 This forms an optical path channel, reducing the interference of stray light from the environment on the imaging system.

[0050] The pixel coordinate system of the image captured by the low-speed, high-resolution camera is used as the main coordinate system. The pixel coordinate system of images captured by a high-speed, low-resolution camera is a secondary coordinate system. Solve for the mapping parameters to complete the coordinate mapping from the secondary coordinate system to the primary coordinate system. The solution results are shown in Table 1.

[0051] Table 1

[0052] Data collection before loading Frame (taken in this embodiment) A static image is used to generate a real-time grayscale difference matrix through pixel-by-pixel difference operations. When continuous The percentage of pixels exceeding a preset high-level threshold in the frame difference matrix. (In this embodiment, the following is taken) When the trigger path switches to a stable high level, the high-speed, low-resolution camera immediately starts acquiring data.

[0053] When the triggering conditions are met, the high-speed, low-resolution camera immediately starts, operating at a frame rate of... (In this embodiment, the following is taken) Continuous acquisition: Shooting frame rate Shooting duration , In this embodiment, we take... .

[0054] For full-cycle deformation images acquired by a low-speed, high-resolution camera, the PINN-DIC technique is used to obtain the global displacement field. Subsequently, for the instantaneous crack dynamic images acquired by the high-speed, low-resolution camera, a consistent PINN-DIC method was used to obtain the displacement field corresponding to the high-speed image, accurately extracting core fracture parameters such as the crack tip position and propagation trajectory. The evolution results of the displacement field measured by the high-speed, low-resolution camera (…) Figure 8 As can be seen, the crack initiation and propagation occur within the x-coordinate range of [24mm, 25mm], ultimately achieving spatial alignment of the deformation data from the dual optical paths. This is combined with measurements obtained from a low-speed, high-resolution camera. The strain field evolution characteristics reveal that within the x-coordinate range of [25mm, 27mm] in the tensile region at the bottom of the specimen, the strain values ​​exhibit a significant local high value distribution, and a significant strain gradient exists with adjacent regions. As loading time progresses, the strain concentration effect in this region intensifies. Although the strain values ​​in the x-coordinate range of [5mm, 15mm] are also at a high level, the strain difference with the surrounding area is small, and no significant strain concentration characteristic with fracture orientation is formed. This phenomenon clarifies the correspondence between the crack initiation location and the early strain concentration region, confirming the intrinsic correlation between strain concentration and crack initiation and propagation, and providing direct experimental evidence for the cross-scale mechanism analysis of "long-term damage evolution - instantaneous fracture failure" in nuclear graphite.

[0055] The above-described specific implementations can be partially adjusted by those skilled in the art in different ways without departing from the principles and purpose of the present invention. The scope of protection of the present invention is defined by the claims and is not limited to the above-described specific implementations. All implementation schemes within the scope of the claims are bound by the present invention.

Claims

1. A method for measuring the deformation field of the entire process of damage evolution and crack propagation in dual-optical-path nuclear graphite, characterized in that, include: Step 1: Adjust the camera working distance, select the lens focal length, and use the cross laser to assist in calibrating the spatial position of the dual optical paths; Step 2: Calibrate using a checkerboard calibration board and solve for the dual-camera coordinate system transformation matrix; Step 3: Establish a background grayscale baseline, calibrate the trigger threshold, and construct a fracture triggering mechanism; Step 4: Continuously acquire data throughout the entire cycle using a low-speed, high-resolution camera; when the triggering conditions are met, capture crack images using a high-speed, low-resolution camera. Step 5: Solve the high-precision deformation field of the specimen surface throughout the entire loading cycle using the PINN-DIC (Physical Information Neural Network) method. While identifying the damage feature area, solve the displacement field of the specimen surface during the high-speed fracture process using PINN-DIC to extract the crack core parameters. Step 6: Combining the strain distribution evolution characteristics with the crack initiation location, map the high-speed fracture information to the main coordinate system and establish a cross-scale correlation mechanism.

2. The method for measuring the deformation field of the entire process of damage evolution and crack propagation in dual-optical-path nuclear graphite as described in claim 1, is characterized in that, Step 1, namely, building a dual-optical-path imaging module: a cubic beam splitter and a low-speed, high-resolution camera are placed in front of the nuclear graphite specimen to be observed to form the first optical path. At the same time, a high-speed, low-resolution camera is placed in the direction perpendicular to the cubic beam splitter and the first optical path to form the second optical path.

3. The method for measuring the deformation field of the entire process of damage evolution and crack propagation in dual-optical-path nuclear graphite as described in claim 2, is characterized in that... The size of the surface target observation area (AOI) of the nuclear graphite specimen to be observed. and the target surface size of the two cameras satisfy: To obtain the magnification Then, by the camera working distance Image distance and lens focal length The relationship, that is ,get .

4. The method for measuring the deformation field of the entire process of damage evolution and crack propagation in dual-optical-path nuclear graphite as described in claim 2, characterized in that, in When adjusting the spatial position of the two cameras and the cubic beam splitter, a laser level is used in three directions to emit cross lasers to the preset optical path to assist in calibrating the spatial attitude of the cameras and the cubic beam splitter and ensure the coaxiality and perpendicularity of the two optical paths. Select The cubic beam splitter encapsulation container and light shield are made of black reinforced carbon fiber material. Pre-reserved holes adapted to the lens diameter are opened on the sides facing the two cameras and extend outward to form an optical path channel, effectively blocking the intrusion of ambient stray light and ensuring the stability of imaging signal-to-noise ratio and grayscale signal.

5. The method for measuring the deformation field of the entire process of damage evolution and crack propagation in dual-path nuclear graphite according to claim 1, characterized in that, Step 2, as described above, uses the pixel coordinate system of the low-speed, high-resolution camera image as the main coordinate system. The pixel coordinate system of a high-speed, low-resolution camera image is a secondary coordinate system. The checkerboard calibration board (or marker array) is placed within the shared effective field of view of the two cameras. The Harris corner detection algorithm is used to extract the corner coordinates from the primary and secondary coordinate system images. Specifically: Where: the coordinates of the corner points in the principal coordinate system are ( The coordinates of the corresponding corner point in the secondary coordinate system are ( ), , , , For quantities related to rotation and scaling, , This represents the translation amount.

6. The method for measuring the deformation field of the entire process of damage evolution and crack propagation in dual-optical-path nuclear graphite as described in claim 1, characterized in that, Step 3 specifically includes: 3.1 Based on the static image of the graphite before loading, a background grayscale reference matrix is ​​established by averaging multiple images over time. Specifically, images were collected before loading. A background grayscale reference matrix is ​​constructed from a frame of still image using a point-by-point pixel temporal averaging method. ,in: For the first The grayscale matrix of a frame of still image. The corresponding pixel row and column coordinates of the image. In the reference matrix The grayscale value of the pixel at the location; 3.2 During the experiment, a real-time grayscale difference matrix was generated through pixel-by-pixel difference operations. ,in: For low-speed, high-resolution cameras The image grayscale moments are acquired in real time; in the initial loading stage, the graphite nucleus only experiences slight surface displacement with no significant morphological changes, and the grayscale distribution fluctuates gently. The absolute values ​​are small and all are within the low-level threshold range, so the high-speed, low-resolution camera remains in a ready-to-trigger state. 3.3 After the crack initiates and propagates, the grayscale of the specimen surface undergoes a sudden change, and the crack area exhibits a grayscale characteristic close to a pure black background, leading to... Significantly negative, and with an absolute value greatly exceeding the low-level threshold, it forms a characteristic trigger signal characterizing fracture behavior. When continuous The percentage of pixels exceeding a preset high-level threshold in the frame difference matrix. Effective proportion of monitoring area When the trigger path switches to a stable high level, the high-speed, low-resolution camera immediately starts acquiring data.

7. The method for measuring the deformation field of the entire process of damage evolution and crack propagation in dual-path nuclear graphite according to claim 1, characterized in that, In step 4, when the grayscale change meets the trigger condition in S3, the high-speed low-resolution camera is activated at a frame rate. The continuous data collection duration meets the following requirements: as well as ,in: To estimate the maximum crack length, To estimate the maximum crack propagation rate, The minimum number of effective image frames required to characterize the entire crack propagation process.

8. The method for measuring the deformation field of the entire process of damage evolution and crack propagation in dual-optical-path nuclear graphite according to claim 1, characterized in that, The aforementioned global deformation field calculation and damage feature identification refers to: inputting the image coordinate domain into a Physical Information Neural Network (PINN) to predict the displacement field, which is then used to predict the deformed image. Compared with actual deformed images The mean squared error is used as the loss function, specifically: Training is achieved by iteratively optimizing parameters through a fully connected network and minimizing the loss function, where: To account for the total number of pixels involved in the calculation, this process addresses the computational distortion problem at the specimen boundary in traditional DIC; leveraging the automatic differentiation function of a deep learning framework, the global displacement field output by PINN is calculated. Performing differential operations requires no additional numerical difference processing; the result is derived directly. , Directional strain and shear strain By analyzing the spatial distribution and temporal evolution characteristics of the components, the strain concentration area is defined as the damage characteristic area of ​​nuclear graphite, thus clarifying the spatial location and development trend of damage accumulation. The refined analysis of PINN-DIC refers to inputting the coordinate domain of a high-speed, low-resolution image into a fully connected network to predict the local displacement field corresponding to the high-speed, low-resolution image. This fully connected network uses the gray-level consistency assumption as the loss function and achieves training by iteratively optimizing the parameters until the loss function converges. This process effectively avoids the problem of missing information in the crack region in traditional DIC and accurately extracts core fracture parameters such as the crack tip position, propagation trajectory, and local deformation features, providing instantaneous fracture feature data for subsequent correlation mechanism analysis.

9. The method for measuring the deformation field of the entire process of damage evolution and crack propagation in dual-optical-path nuclear graphite according to claim 1, characterized in that, Step 6 refers to: based on the principal-sub-coordinate transformation matrix established in step 2, accurately mapping the crack initiation coordinates and propagation trajectory extracted by refined analysis of PINN-DIC to the strain field coordinate system for global deformation field measurement and damage feature identification, realizing spatial alignment of dual-optical path data, clarifying the correspondence between crack initiation and propagation locations and the early strain concentration areas in low-speed high-resolution images, and finally establishing the correlation feature of "strain concentration accumulation - crack initiation and propagation", providing experimental basis for the cross-scale mechanism analysis of nuclear graphite "long-term damage evolution - instantaneous fracture failure".

10. A dual-optical-path graphite damage evolution and crack propagation deformation field measurement system for implementing the method of any one of claims 1-9, characterized in that, include: The system comprises a spectroscopic imaging module, a trigger adaptation module, a calibration module, and a deformation analysis module. Specifically: the spectroscopic imaging module uses a low-speed, high-resolution camera for long-term, full-domain deformation observation to fully record the damage evolution and strain concentration process of the specimen from loading to fracture; and a high-speed, low-resolution camera for short-term, high-dynamic capture to accurately record the instantaneous dynamic behavior of crack initiation and propagation. The trigger adaptation module monitors characteristic grayscale abrupt changes and connected domain morphological changes caused by specimen fracture, and uses a post-trigger method to activate the high-speed, low-resolution camera of the spectroscopic imaging module to acquire images. The calibration module completes the dual-optical-path spatial offset calibration using a spatial coordinate transformation method. Differential calibration enables spatial matching between the instantaneous fracture region captured by the high-speed low-resolution camera and the long-term deformation field recorded by the low-speed high-resolution camera, laying the foundation for cross-timescale data fusion. The deformation analysis module uses the Physical Information Neural Network-driven Digital Image Correlation (PINN-DIC) method to analyze the deformation field throughout the damage evolution and crack propagation process. Based on the long-term deformation field information measured by the low-speed high-resolution camera, it predicts the potential crack propagation area through the spatial distribution characteristics of deformation, while fusing the crack propagation deformation field measured by the high-speed low-resolution camera and associating it with long-term damage accumulation – instantaneous fracture failure. The aforementioned beam-splitting imaging module includes: a low-speed high-resolution camera, a high-speed low-resolution camera, and a cubic beam-splitting prism, wherein: the cubic beam-splitting prism is positioned directly in front of the nuclear graphite specimen, and uniformly splits the reflected light from the surface of the nuclear graphite specimen into two beams, which are then guided to the low-speed high-resolution camera and the high-speed low-resolution camera, respectively. The deformation analysis module includes a control unit and a digital image correlation unit. The control unit establishes data transmission links with a low-speed high-resolution camera and a high-speed low-resolution camera to synchronously control the acquisition parameters of the two types of cameras. It also stores the acquired full-cycle deformation images and instantaneous crack images according to timestamps and outputs them to the digital image correlation unit. The digital image correlation unit performs full-domain high-precision strain field analysis and damage feature region identification on the full-cycle loading images acquired by the low-speed high-resolution camera to quantify the nuclear graphite damage accumulation process. It also performs local displacement field solution and crack core parameter extraction on the instantaneous crack images acquired by the high-speed low-resolution camera to characterize the failure fracture dynamic behavior.

Images