An AI-assisted mechanical nephogram generation method for mechanoluminescent materials

Abstract

The application relates to the field of intelligent sensing and computer vision, in particular to an AI-assisted mechanical cloud chart generation method of mechanoluminescent materials, which aims to provide an AI-assisted mechanical cloud chart generation method of mechanoluminescent materials. The main problem to be solved is that the afterglow effect in the mechanoluminescent image limits the accurate description of the mechanical state of the component. The beneficial effects are that the AI-assisted high-precision mechanical cloud chart generation model provided by the application can realize more accurate description of the mechanical state of the component by using the mechanoluminescent sensing data, and the average RMSE error is as low as 3.8368%.

Description

Technical Field

[0001] This invention relates to the field of intelligent sensing and computer vision technology, specifically to a method for generating mechanical cloud maps of AI-assisted mechanoluminescent materials. Background Technology

[0002] Methocynical materials, as a special type of piezoelectric luminescent material, exhibit force-induced luminescence through their piezoelectric or triboelectric effects, thus finding wide application in the field of intelligent sensing. These materials are particularly suitable for monitoring the mechanical behavior of structural components and detecting environmental changes, making them ideal for the development of various stress and strain sensors. However, a major problem with methocynical materials is the afterglow effect of luminescence. This characteristic makes it difficult for sensors to accurately describe the mechanical condition of components under complex loading conditions, thereby limiting their widespread application in practical engineering.

[0003] With advancements in deep learning and computer vision technologies, AI (Artificial Intelligence)-assisted image generation methods have been widely applied in various fields, including image generation, style transfer, and image restoration and reconstruction. These methods demonstrate significant advantages in image analysis and error correction, addressing the challenges faced by mechanoluminescence sensing technology and thus improving sensor performance and application scope. However, the accurate description of the mechanical state of components is limited by the afterglow effect in mechanoluminescence images, necessitating the development of an AI-assisted method to convert experimental mechanoluminescence images into precise mechanical contour maps. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides an AI-assisted method for generating mechanical cloud maps of mechanoluminescent materials. The main problem it solves is the issue of how the afterglow of light in mechanoluminescent images affects the accurate description of the mechanical state of components.

[0005] To achieve the above objectives, the following technical solution is provided:

[0006] A method for generating mechanical cloud maps of AI-assisted mechanoluminescent materials, characterized by comprising the following steps:

[0007] Step 1: Fabricate a mechanoluminescence sensing component and capture mechanoluminescence video of the component during the entire loading process;

[0008] Step 2: Use the finite element analysis method to perform a finite element simulation of the entire loading process of the component, and export the mechanical cloud map video of the entire loading process of the component;

[0009] Step 3: Extract frames from the luminescence video obtained in Step 1 and the mechanical cloud map video obtained in Step 2 according to the time step increment; and perform preprocessing such as image cropping, scaling and spatiotemporal registration on the obtained frames;

[0010] Step 4: Use the preprocessed luminescence images and cloud maps obtained in Step 3 to establish a dedicated image dataset of the mechanical state of components based on mechanoluminescent devices; and divide it into training set, validation set and test set according to a certain ratio;

[0011] Step 5: Establish the network architecture of the model, which is based on the VGG16-UNet image generation model;

[0012] Step 6: Set the loss function, optimizer, and training hyperparameters for model training;

[0013] Step 7: Input the training set images into the model for iterative training; evaluate the model performance using the validation set after each training round; when the training and validation losses converge in unison, the final weighted model is obtained.

[0014] Step 8: Load the final weight model into the network; input the luminescence images from the test set into the model to generate the corresponding mechanical cloud map; and further perform error analysis on the generated cloud map.

[0015] Furthermore, in step one, the fabrication of the mechanoluminescence sensing component includes using adhesive and nesting techniques to fix the sensor to the surface of the component under test, requiring good mechanical and optical contact between the sensor and the component surface to ensure accurate transmission and detection of mechanoluminescence signals; the mechanoluminescence data acquisition includes using a high-speed camera to capture the mechanoluminescence signals emitted by the sensor, and the entire process from the start of loading to the end of unloading of the component should be video recorded, and the acquired data includes luminescence intensity, duration, and spectral characteristics.

[0016] Furthermore, in step two, the finite element simulation of the entire process of the component under load includes: establishing the finite element model of the component, setting boundary conditions and loading conditions, mesh generation, and finite element analysis. The simulation results are analyzed in detail to verify the consistency between the data displayed in the cloud map and the experimental and theoretical analysis. If necessary, the model parameters are adjusted according to the analysis results and the simulation is repeated to improve the accuracy and reliability of the simulation. The cloud map video export requires the use of a finite element visualization tool to export the dynamic cloud map as a video format.

[0017] Furthermore, in step three, the video frame extraction includes frame extraction of force luminescence video and stress cloud video. The two types of video frames are extracted synchronously according to the set time step increment to ensure accurate time correspondence between the two for subsequent comparison and analysis. The image preprocessing includes image cropping, image scaling, and spatiotemporal registration to ensure that the processed force luminescence image and stress cloud have uniform resolution and spatiotemporal alignment accuracy.

[0018] Furthermore, in step four, the image dataset establishment integrates the preprocessed luminescent images and mechanical cloud maps from step three, with each pair of images at each time step serving as a data sample, ensuring that each luminescent image and its corresponding mechanical cloud map correspond correctly in the time series; subsequently, the dataset is divided into a training set, a validation set, and a test set in a 7:2:1 ratio.

[0019] Furthermore, in step five, the VGG16-UNet image generation model consists of a backbone feature extraction part, a feature fusion part, and a feature decoding part.

[0020] Furthermore, the backbone feature extraction part adopts VGG16; the feature decoding part adopts the UNet decoder part; and the feature fusion part directly connects the feature maps of different layers of VGG16 to the corresponding upsampling layers of UNet.

[0021] Furthermore, in step six, the model training loss function adopts Huber Loss and SSIM Loss; the optimizer is Adam (β1 = 0.90; β2 = 0.999); the training hyperparameters include: total number of training epochs; training batch parameters; regularization coefficient; initial learning rate; learning rate descent strategy; learning rate descent factor; number of learning rate step decay epochs; validation frequency; weight decay coefficient; gradient threshold.

[0022] Furthermore, in step seven, the consistent convergence of training and validation losses indicates that the training loss and validation loss have a consistent decreasing trend as the training process progresses and approach a stable minimum value.

[0023] Furthermore, in step eight, the cloud map generation error metric is RMSE.

[0024] The beneficial effects of the present invention are as follows: the AI-assisted high-precision mechanical cloud map generation model provided by the present invention can use mechanoluminescence sensing data to achieve a more accurate description of the mechanical state of the component, with an average RMSE error as low as 3.8368%. Attached Figure Description

[0025] Figure 1 This is a flowchart of the present invention;

[0026] Figure 2 These are four types of components of the sensing film arranged in the embodiments of the present invention;

[0027] Figure 3 These are the finite element models of the four types of components in this embodiment of the invention;

[0028] Figure 4 These are the preprocessed luminescence images and corresponding stress cloud maps in the embodiments of the present invention;

[0029] Figure 5 This serves as the training, verification, and test set in the embodiments of the present invention;

[0030] Figure 6 This is a diagram of the VGG16-UNet network model architecture in an embodiment of the present invention;

[0031] Figure 7 This is a graph showing the relationship between training loss and the number of training rounds in an embodiment of the present invention;

[0032] Figure 8 These are five luminescent images and the generated stress cloud map in an embodiment of the present invention. Detailed Implementation

[0033] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0034] A method for generating mechanical cloud maps of AI-assisted mechanoluminescent materials, such as Figure 1 The flowchart shown includes the following steps:

[0035] Step 1: For the sensor component placement, a thin-film bonding method is used to obtain the sensing components, including cylindrical epoxy resin components, defect-free stainless steel components, defective stainless steel components, and stainless steel beams, as shown below. Figure 2 As shown in (a), (b), (c) and (d).

[0036] Step Two: For the finite element simulation of the entire loading process of the four types of components mentioned in Step One, the finite element analysis software Abaqus was used for simulation experiments and conclusion analysis. The constructed uniaxial compression model of the cylindrical epoxy resin component, the uniaxial tension model of the defect-free stainless steel component, the uniaxial tension model of the defective stainless steel component, and the three-point bending model of the stainless steel beam are respectively as follows: Figure 3 As shown in (a), (b), (c) and (d).

[0037] Step 3: For the frame extraction and image preprocessing of the luminescence video and cloud image video obtained in Steps 1 and 2, relevant code was written using Python packages such as OpenCV-Python, Pillow, and os to implement data preprocessing. The luminescence images obtained after processing for each experimental condition are shown below. Figure 4 As shown in (a1), (b1), (c1), and (d1), the resulting stress cloud diagrams are as follows: Figure 4As shown in (a2), (b2), (c2) and (d2).

[0038] Step 4: Using the preprocessed luminescence image and stress cloud map obtained in Step 3, establish a dedicated image dataset, and divide it into training, validation, and test sets according to a certain ratio, as shown below. Figure 5 As shown in (a), (b) and (c).

[0039] Step 5: For the model network architecture construction, a VGG16-UNet-based network architecture is adopted. VGG16 serves as the backbone for feature extraction, and the UNet decoder is the feature decoding part. The specific model architecture is as follows: Figure 6 As shown.

[0040] Step 6: Adjust and set the training loss function, optimization algorithm, and training hyperparameters of the model. The specific parameters are shown in Table 1.

[0041] Table 1 Model Training Parameter Configuration

[0042]

[0043] Step 7: Iteratively train the model using the training set, and validate the model using the validation set according to the model validation frequency, until the training loss and validation loss converge, obtaining the final weighted model; the changes in training and validation losses with the number of training rounds are as follows: Figure 7 As shown.

[0044] Step 8: Using the luminescence image obtained from the experiment and the VGG16-UNet with the final weight model loaded, the corresponding stress cloud map can be generated.

[0045] like Figure 8 As shown, stress cloud maps are generated for four randomly selected luminescence images. Among them, Figure 8 (a) is a luminescent image; Figure 8 (b) is the corresponding actual stress contour plot; Figure 8 (c) shows the generated stress contour plot.

[0046] It can be observed that the generated results are extremely accurate, achieving high-precision stress cloud map generation based on mechanoluminescence sensing images.

[0047] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the invention can be implemented in other specific forms without departing from its spirit or essential characteristics. Therefore, the embodiments should be considered in all respects as exemplary and non-limiting, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be included within the present invention. No reference numerals in the claims should be construed as limiting the scope of the claims.

Claims

1. A method for generating mechanical cloud maps of AI-assisted mechanoluminescent materials, characterized in that, Includes the following steps: Step 1: Fabricate a mechanoluminescence sensing component and capture mechanoluminescence video of the component during the entire loading process; Step 2: Use the finite element analysis method to perform a finite element simulation of the entire loading process of the component, and export the mechanical cloud map video of the entire loading process of the component; Step 3: Extract frames from the luminescence video obtained in Step 1 and the mechanical cloud map video obtained in Step 2 according to the time step increment; and perform image cropping, scaling and spatiotemporal registration preprocessing on the obtained frames. Step 4: Use the preprocessed luminescence images and cloud maps obtained in Step 3 to establish a dedicated image dataset of the mechanical state of components based on mechanoluminescent devices; and divide it into training set, validation set and test set according to a certain ratio; Step 5: Establish the network architecture of the model, which is based on the VGG16-UNet image generation model; Step 6: Set the loss function, optimizer, and training hyperparameters for model training; Step 7: Input the training set images into the model for iterative training; evaluate the model performance using the validation set after each training round; when the training and validation losses converge in unison, the final weighted model is obtained. Step 8: Load the final weight model into the network; The model is then fed with luminescence images from the test set to generate corresponding mechanical cloud maps; further error analysis is performed on the generated cloud maps.

2. The method for generating mechanical cloud maps of AI-assisted mechanoluminescent materials according to claim 1, characterized in that: In step one, the fabrication of the mechanoluminescence sensing component includes using adhesive and nesting techniques to fix the sensor to the surface of the component under test. Good mechanical and optical contact is required between the sensor and the component surface to ensure accurate transmission and detection of mechanoluminescence signals. The mechanoluminescence data acquisition includes using a high-speed camera to capture the mechanoluminescence signals emitted by the sensor. The entire process from loading to unloading of the component should be video recorded. The acquired data includes luminescence intensity, duration, and spectral characteristics.

3. The method for generating mechanical cloud maps of AI-assisted mechanoluminescent materials according to claim 1, characterized in that: In step two, the finite element simulation of the entire process of the component under load includes: establishing the finite element model of the component, setting boundary conditions and loading conditions, mesh generation, and finite element analysis. The simulation results are analyzed in detail to verify the consistency between the data displayed in the cloud map and the experimental and theoretical analysis. If necessary, the model parameters are adjusted according to the analysis results and the simulation is repeated to improve the accuracy and reliability of the simulation. The cloud map video export requires the use of a finite element visualization tool to export the dynamic cloud map as a video format.

4. The method for generating mechanical cloud maps of AI-assisted mechanoluminescent materials according to claim 1, characterized in that: In step three, the video frame extraction includes frame extraction of force luminescence video and stress cloud image video. The two types of video frames are extracted synchronously according to the set time step increment to ensure that the time correspondence between the two is accurate for subsequent comparison and analysis. The image preprocessing includes image cropping, image scaling and spatiotemporal registration to ensure that the processed force luminescence image and stress cloud image have uniform resolution and spatiotemporal alignment accuracy.

5. The method for generating mechanical cloud maps of AI-assisted mechanoluminescent materials according to claim 1, characterized in that: In step four, the image dataset establishment integrates the preprocessed luminescent images and mechanical cloud maps from step three. Each pair of images at each time step is used as a data sample to ensure that each luminescent image and its corresponding mechanical cloud map correspond correctly in the time series. Subsequently, the dataset is divided into training set, validation set, and test set in a ratio of 7:2:

1.

6. The method for generating mechanical cloud maps of AI-assisted mechanoluminescent materials according to claim 1, characterized in that: In step five, the VGG16-UNet image generation model consists of a backbone feature extraction part, a feature fusion part, and a feature decoding part.

7. The method for generating mechanical cloud maps of AI-assisted mechanoluminescent materials according to claim 6, characterized in that: The backbone feature extraction part uses VGG16; the feature decoding part uses the UNet decoder part; the feature fusion part directly connects the feature maps of different layers of VGG16 to the corresponding upsampling layers of UNet.

8. The method for generating mechanical cloud maps of AI-assisted mechanoluminescent materials according to claim 1, characterized in that: In step six, the model training loss function adopts Huber Loss and SSIM Loss; the optimizer is Adam, where β1=0.90 and β2=0.999; the training hyperparameters include: total number of training epochs; training batch parameters; regularization coefficient; initial learning rate; learning rate descent strategy; learning rate descent factor; number of learning rate decay epochs; validation frequency; weight decay coefficient; and gradient threshold.

9. The method for generating mechanical cloud maps of AI-assisted mechanoluminescent materials according to claim 1, characterized in that: In step seven, the consistent convergence of training and validation losses indicates that the training and validation losses have a consistent decreasing trend as the training progresses and approach a stable minimum.

10. The method for generating mechanical cloud maps of AI-assisted mechanoluminescent materials according to claim 1, characterized in that: In step eight, the cloud map generation error metric is RMSE.

Images