Prediction method and device for ultimate bearing capacity of composite material wallboard assembly

Abstract

The invention discloses a method for predicting the ultimate bearing capacity of a composite wallboard assembly, and belongs to the technical field of material prediction. The method comprises the steps that a plurality of stress pictures of the composite material wallboard assembly are collected through collection equipment, a picture data set is obtained, each stress picture corresponds to a time step, each stress picture in the picture data set is preprocessed, and a target picture set is obtained; inputting the target picture set into a trained external force prediction model to obtain an external force prediction value corresponding to each picture in the target picture set; obtaining an external force-time change curve based on the external force predicted value corresponding to each picture; and training an ultimate bearing capacity prediction model based on the external force-time change curve, and carrying out iterative prediction through the ultimate bearing capacity prediction model to obtain a predicted value of the ultimate bearing capacity of the composite material wallboard assembly. According to the method, the accuracy and reliability of predicting the ultimate bearing capacity of the composite material wallboard assembly are improved.

Description

Technical Field

[0001] This application belongs to the field of materials prediction technology, and in particular relates to a method and apparatus for predicting the ultimate bearing capacity of composite material wall panel components. Background Technology

[0002] Composite materials, as a product of multidisciplinary integration, are prepared by combining two or more different materials through physical compositing or chemical bonding. Their aim is to integrate the advantages of each component material and compensate for the shortcomings of individual materials, thereby meeting the stringent requirements of modern industry for high-performance materials. Through microscopic structural design and complementary properties, composite materials can simultaneously possess multiple excellent properties such as high strength, high stiffness, excellent corrosion resistance, and thermal stability, becoming a key foundational material driving the development of high-end technologies. With the continuous expansion of composite material applications across various industries, issues such as ensuring the structural safety of composite materials, optimizing material design schemes, and extending material service life are receiving increasing attention. Among these, in-depth research into the real-time stress state and ultimate bearing capacity of composite materials has become central to solving these key problems and is of great significance for promoting the sustainable development of the composite materials industry.

[0003] Existing methods for predicting the ultimate bearing capacity of composite panel components primarily rely on classical laminate theory combined with linear or nonlinear finite element analysis. These methods often depend on direct sensor measurements, requiring the deployment of numerous sensors on the panel, which is complex and may affect the original structural performance. Models based on the macroscopic homogeneity assumption struggle to accurately characterize the unique progressive damage behavior of real composite materials (such as fiber fracture, matrix cracking, delamination, and local buckling) and their complex interactions, leading to prediction errors in failure initiation, propagation paths, and final failure modes. Furthermore, high-precision three-dimensional explicit damage models are computationally extremely expensive, failing to meet the rapid iteration requirements of engineering design. While methods based on cohesion models or continuous damage mechanics can simulate delamination or matrix damage, their ability to characterize complex failure modes (such as the coupling of delamination and fiber micro-buckling) is limited, and model parameters are highly dependent on cumbersome experimental calibration. Traditional semi-empirical design methods based on safety margins are overly conservative and fail to fully exploit the material's potential. In conclusion, existing methods for predicting the ultimate bearing capacity of composite panel components suffer from low prediction accuracy, high computational cost, and operational complexity. Summary of the Invention

[0004] This application aims to address at least one of the technical problems existing in the prior art. To this end, this application proposes a method and apparatus for predicting the ultimate bearing capacity of composite material wall panel assemblies, which improves the accuracy and reliability of predicting the ultimate bearing capacity of composite material wall panel assemblies.

[0005] In a first aspect, this application provides a method for predicting the ultimate bearing capacity of a composite material wall panel assembly, the method comprising: Multiple stress images of the composite material wall panel assembly are acquired by the acquisition device to obtain an image dataset. Each stress image corresponds to a time step. Each stress image in the image dataset is preprocessed to obtain a target image set. The target image set is input into the trained external force prediction model to obtain the external force prediction value corresponding to each image in the target image set; Based on the predicted external force value corresponding to each image, the external force-time variation curve is obtained; The ultimate bearing capacity prediction model is trained based on the external force-time variation curve, and the predicted value of the ultimate bearing capacity of the composite material wall panel assembly is obtained by iterative prediction through the ultimate bearing capacity prediction model.

[0006] According to one embodiment of this application, the preprocessing of each force-affected image in the image dataset to obtain a target image set includes: A mask is created based on the part of the material panel that deforms most significantly under external force in each stress image, and the image parts outside the mask are all set to black. The black portion of each stress image is cropped to obtain the cropped stress image; By smoothing the boundaries of each stress image after cropping using the dilation operation, the edges of each stress image are made smoother and the images are more continuous, resulting in the target image set.

[0007] According to one embodiment of this application, the step of inputting the target image set into a trained external force prediction model to obtain the external force prediction value corresponding to each image in the target image set includes: The target image set is input into the CNN module to obtain the local features of each image in the target image set; The target image set is input into the Transformer module to obtain the global features of each image in the target image set; Based on the gating mechanism, the local and global features of each image are fused to obtain the fused features corresponding to each image; The fusion features corresponding to each image are input into a fully connected network to obtain the predicted external force value for each image in the target image set. The external force prediction model is a CNN-Transformer model, which includes a CNN module and a Transformer module.

[0008] According to one embodiment of this application, the step of iteratively predicting the ultimate bearing capacity of the composite material wall panel assembly using the ultimate bearing capacity prediction model includes: The external force data for 10 unit time intervals is obtained as an input sequence. The input sequence is then input into the ultimate bearing capacity prediction model to obtain the predicted value of the external force for the next unit time interval. Add the predicted value of the external force for the next unit time to the end of the original input sequence, and delete the external force data for the first unit time at the beginning of the original input sequence to obtain the updated input sequence. The updated input sequence is fed into the ultimate bearing capacity prediction model for iterative prediction to obtain the predicted value of the ultimate bearing capacity of the composite panel assembly.

[0009] According to one embodiment of this application, the step of inputting the updated input sequence into the ultimate bearing capacity prediction model for iterative prediction to obtain the predicted value of the ultimate bearing capacity of the composite material wall panel assembly includes: The updated input sequence is fed into the ultimate bearing capacity prediction model to obtain the predicted value of the external force per unit time. Calculate the predicted value of the external force in the next unit of time and the increase in the current value of the external force; When the increase is less than the preset threshold, the predicted value of the external force in the next unit time will be used as the predicted value of the ultimate bearing capacity of the composite material wall panel component. When the increase is greater than or equal to the preset threshold, the predicted value of the external force in the next unit time is added to the end of the original input sequence, the external force data in the first unit time at the beginning of the original input sequence is deleted, and the updated input sequence is obtained. Then, the process is repeated to input the updated input sequence into the ultimate bearing capacity prediction model to obtain the predicted value of the external force in the next unit time.

[0010] According to one embodiment of this application, the training process of the external force prediction model includes: Construct a pre-defined external force prediction model; Multiple stress images of the composite material wall panel assembly were acquired and preprocessed to obtain the first dataset; Perform a sliding window operation on the first dataset to obtain a sliding window dataset; Based on the sliding window dataset, the preset external force prediction model is trained using the mean squared error loss function to obtain the trained external force prediction model.

[0011] According to one embodiment of this application, training the ultimate bearing capacity prediction model based on the external force-time variation curve includes: A preset ultimate bearing capacity prediction model is constructed, wherein the preset ultimate bearing capacity prediction model is an LSTM prediction model; Multiple data points are obtained based on the external force-time variation curve. The multiple data points are then cleaned to obtain cleaned data points. The 3σ principle is used to identify and remove the cleaned data points, resulting in the removed data points. Linear interpolation is used to fill in the removed data points to obtain a second dataset; Based on the second dataset, the preset ultimate bearing capacity prediction model is trained using the mean squared error loss function to obtain the trained ultimate bearing capacity prediction model.

[0012] Secondly, this application provides a device for predicting the ultimate bearing capacity of a composite material wall panel assembly, the device comprising: The acquisition module is used to acquire multiple stress images of composite material wall panel components through an acquisition device to obtain an image dataset. Each stress image corresponds to a time step. Each stress image in the image dataset is preprocessed to obtain a target image set. The first processing module is used to input the target image set into the trained external force prediction model to obtain the external force prediction value corresponding to each image in the target image set; The second processing module is used to obtain the external force-time variation curve based on the predicted external force value corresponding to each image; The prediction module is used to train the ultimate bearing capacity prediction model based on the external force-time variation curve, and to perform iterative prediction through the ultimate bearing capacity prediction model to obtain the predicted value of the ultimate bearing capacity of the composite material wall panel assembly.

[0013] Thirdly, this application 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 computer program to implement the method for predicting the ultimate bearing capacity of a composite material wall panel assembly as described in the first aspect above.

[0014] Fourthly, this application provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method for predicting the ultimate bearing capacity of a composite material wall panel assembly as described in the first aspect above.

[0015] Fifthly, this application provides a chip including a processor and a communication interface, the communication interface being coupled to the processor, the processor being used to run programs or instructions to implement the method for predicting the ultimate bearing capacity of composite material wall panel components as described in the first aspect.

[0016] In a sixth aspect, this application provides a computer program product, including a computer program that, when executed by a processor, implements the method for predicting the ultimate bearing capacity of a composite material wall panel assembly as described in the first aspect above.

[0017] Additional aspects and advantages of this application will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of this application.

[0018] The present invention provides a method for predicting the ultimate bearing capacity of composite material wall panel components, which has the following advantages over the prior art: (1) This invention acquires multiple stress images of composite material wall panel components through acquisition equipment, which can quickly obtain the overall deformation information of the wall panel, reduce interference to the structure, improve the efficiency and convenience of data acquisition, and realize non-contact acquisition. Through the external force prediction model, the local and global features of the tensile images of the composite material wall panel components are extracted and analyzed in depth to obtain the external force prediction value corresponding to each image in the target image set. Then, it is used as the input of the ultimate bearing capacity prediction model, and the ultimate bearing capacity is obtained through iterative prediction. This realizes the efficient conversion from image data to mechanical performance indicators and improves the accuracy and reliability of the ultimate bearing capacity prediction of composite material wall panel components.

[0019] (2) This invention inputs the updated input sequence into the ultimate bearing capacity prediction model and calculates the increase of the predicted value of the external force in the next unit time and the increase of the external force value at the current time. By judging whether to update the input sequence based on the increase, the input of the ultimate bearing capacity prediction model can be dynamically adjusted, which effectively improves the adaptability and accuracy of the ultimate bearing capacity prediction model in the ultimate bearing capacity prediction process.

[0020] (3) By acquiring external force data within a certain time period and inputting it into the ultimate bearing capacity prediction model, this invention can predict the change of external force in the composite material wall panel assembly in the next time unit. By continuously updating the input sequence and performing iterative prediction, the predicted value of bearing capacity can be updated in real time, resulting in a more accurate predicted value of ultimate bearing capacity. This has better applicability, can more comprehensively mine data features, and improves the accuracy of the prediction of the ultimate bearing capacity of the composite material wall panel assembly. Attached Figure Description

[0021] The above and / or additional aspects and advantages of this application will become apparent and readily understood from the description of the embodiments taken in conjunction with the following drawings, in which: Figure 1 This is one of the flowcharts illustrating the method for predicting the ultimate bearing capacity of composite material wall panel components provided in the embodiments of this application; Figure 2 This is a basic structural diagram of the wall panel assembly provided in the embodiments of this application; Figure 3 This is a schematic diagram of mask creation provided in an embodiment of this application; Figure 4 This is a second schematic flowchart of the method for predicting the ultimate bearing capacity of composite material wall panel components provided in the embodiments of this application; Figure 5 This is a schematic diagram of the structure of the predictive device for the ultimate bearing capacity of composite material wall panel components provided in the embodiments of this application; Figure 6 This is a schematic diagram of the structure of the electronic device provided in the embodiments of this application. Detailed Implementation

[0022] The technical solutions of the embodiments of this application will be clearly described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application are within the scope of protection of this application.

[0023] The terms "first," "second," etc., used in the specification and claims of this application are used to distinguish similar objects and not to describe a specific order or sequence. It should be understood that such use of data can be interchanged where appropriate so that embodiments of this application can be implemented in orders other than those illustrated or described herein, and the objects distinguished by "first," "second," etc., are generally of the same class and the number of objects is not limited; for example, a first object can be one or more. Furthermore, in the specification and claims, "and / or" indicates at least one of the connected objects, and the character " / " generally indicates that the preceding and following objects are in an "or" relationship.

[0024] The following description, in conjunction with the accompanying drawings, details the method for predicting the ultimate bearing capacity of composite material wall panel components, the device for predicting the ultimate bearing capacity of composite material wall panel components, the electronic equipment, and the readable storage medium provided in this application, through specific embodiments and application scenarios.

[0025] The method for predicting the ultimate bearing capacity of composite material wall panel components can be applied to the terminal, specifically by the hardware or software in the terminal.

[0026] The terminal includes, but is not limited to, portable communication devices such as mobile phones or tablets with touch-sensitive surfaces (e.g., touchscreen displays and / or touchpads). It should also be understood that, in some embodiments, the terminal may not be a portable communication device, but rather a desktop computer with touch-sensitive surfaces (e.g., touchscreen displays and / or touchpads).

[0027] The following embodiments describe a terminal including a display and a touch-sensitive surface. However, it should be understood that the terminal may include one or more other physical user interface devices such as a physical keyboard, mouse, and joystick.

[0028] The method for predicting the ultimate bearing capacity of composite material wall panel components provided in this application embodiment can be executed by an electronic device or a functional module or entity in an electronic device that can realize the method for predicting the ultimate bearing capacity of composite material wall panel components. The electronic devices mentioned in this application embodiment include, but are not limited to, mobile phones, tablets, computers, cameras, and wearable devices. The following uses an electronic device as the execution subject to illustrate the method for predicting the ultimate bearing capacity of composite material wall panel components provided in this application embodiment.

[0029] In cutting-edge fields such as aerospace, automotive, and new energy, the application of composite materials is becoming increasingly widespread due to the ever-increasing demands for lightweight, high-strength, high-temperature resistance, and corrosion resistance. In aerospace, carbon fiber reinforced composites, with their excellent specific strength and specific modulus, are widely used in the manufacture of aircraft fuselages and key structural components, effectively reducing aircraft weight and significantly improving fuel efficiency and load capacity. In response to the development needs of energy conservation, emission reduction, and improved fuel economy, the automotive manufacturing industry actively applies composite materials to body structures, seating systems, interior components, and engine components, significantly enhancing vehicle safety and collision resistance while reducing vehicle weight. In wind power generation, glass fiber or carbon fiber composites are commonly used to manufacture large-sized wind turbine blades, effectively improving their strength and durability while reducing blade weight, thereby increasing wind energy conversion efficiency.

[0030] Figure 1 This is one of the flowcharts illustrating the method for predicting the ultimate bearing capacity of composite material wall panel components provided in this application embodiment, such as... Figure 1 As shown, the method for predicting the ultimate bearing capacity of the composite material wall panel assembly includes steps 110, 120, 130, and 140.

[0031] Step 110: Collect multiple stress images of the composite material wall panel assembly using the acquisition device to obtain an image dataset. Each stress image corresponds to a time step. Preprocess each stress image in the image dataset to obtain a target image set. In some embodiments, preprocessing each stress image in the image dataset to obtain a target image set includes: A mask is created based on the part of the material panel that deforms most significantly under external force in each stress image, and the image parts outside the mask are all set to black. The black portion of each stress image is cropped to obtain the cropped stress image; By smoothing the boundaries of each stress image after cropping using the dilation operation, the edges of each stress image are made smoother and the images are more continuous, resulting in the target image set.

[0032] For example, the acquisition device is an industrial camera, which captures multiple images of the composite material wall panel assembly under stress to obtain an image dataset, with each stress image corresponding to a time step.

[0033] Figure 2 This is a basic structural diagram of the wall panel assembly provided in the embodiments of this application, such as... Figure 2 As shown, the material is a polymer-based composite material. The composite panel assembly has a length L of 135 mm, a width w of 36 mm, a thickness h of 3 mm, and a hole diameter of... The length is 6mm, the end distance e is 18mm, s is the length of the pad, and d is the diameter of the fastener.

[0034] Figure 3 This is a schematic diagram of mask creation provided in an embodiment of this application, as shown below. Figure 3 As shown, the deformation of the middle bolt portion of the composite panel assembly is most pronounced with increasing tension during the tensile process. A masking method was used to extract the portions of the stress image where deformation significantly changes with external force. The images captured by the industrial camera were converted into grayscale matrices. The blue portion of the panel assembly was then extracted using a masking method. Specifically, a mask was set, with values ​​in the grayscale matrix falling within the blue range set to 1, and all other values ​​set to 0. Multiplying the mask by the original grayscale matrix yielded an image containing only the blue area, with the rest rendered as black.

[0035] It should be noted that the mask is essentially a binary image (with only two pixel values, 0 and 1). A specific algorithm is used to identify areas in the composite panel component image where the deformation changes significantly with external force, and areas with insignificant deformation are set to 0, while other areas are set to 1.

[0036] The mask is multiplied by the original image taken by the industrial camera. Pixels with a value of 1 in the mask retain the corresponding pixel values ​​in the original image, while pixels with a value of 0 are set to 0. This process extracts the parts where deformation changes significantly with external force.

[0037] Furthermore, the black portions of each stress image are cropped to reduce image size and remove useless information. Morphological dilation operations are then used to smooth the edges and eliminate noise in the cropped stress images. Specifically, a disk-shaped structuring element with a radius of 2 pixels is used. During the pixel traversal, the center of the structuring element is aligned with each pixel sequentially. When a pixel with a value of 1 exists within the structuring element's coverage area, the center pixel is forcibly updated to 1. Taking the jagged noise at the edges of the blue area in the image as an example, these noise points are typically discrete pixels. Through dilation, the tiny gaps at the edges are effectively filled, smoothing outwards the originally uneven edges and making the blue area outline closer to the actual physical boundary. Simultaneously, isolated black noise points in the image, since most of their surrounding pixel values ​​are 0, are covered by the surrounding normal pixel values ​​during dilation, ultimately achieving effective noise elimination. This provides a clearer and more accurate data foundation for subsequent image analysis, making the edges of each stress image smoother and the images more continuous, resulting in the target image set.

[0038] In this embodiment, by creating a mask and cropping the part of the material panel that deforms most significantly under external force in each stress image, the image size is reduced, irrelevant information is removed, and computational efficiency is improved. The dilation operation can effectively improve the image quality, resulting in a smoother and more continuous set of target images, thus optimizing the quality of the input data and improving the design efficiency and safety of the panel assembly.

[0039] Step 120: Input the target image set into the trained external force prediction model to obtain the external force prediction value corresponding to each image in the target image set; In some embodiments, inputting the target image set into a trained external force prediction model to obtain the predicted external force value for each image in the target image set includes: The target image set is input into the CNN module to obtain the local features of each image in the target image set; The target image set is input into the Transformer module to obtain the global features of each image in the target image set; Based on the gating mechanism, the local and global features of each image are fused to obtain the fused features corresponding to each image; The fusion features corresponding to each image are input into a fully connected network to obtain the predicted external force value for each image in the target image set. The external force prediction model is a CNN-Transformer model, which includes a CNN module and a Transformer module.

[0040] Understandably, the target image set comprises multiple images across multiple time steps, each containing numerous features reflecting detailed information. These features provide a detailed description of specific locations and regions, reflecting the microscopic changes in the material under stress. These features are local features, typically possessing high resolution and accuracy. In the image sequence across multiple time steps, the overall shape change trend of the panel assembly, the relative positional relationships between different parts, and the overall deformation pattern during stress are all global features.

[0041] For example, minute deformations at bolted connections and textural changes in areas of localized stress concentration in the material are local features. However, as external forces continue to act, changes in the bolt's position cannot be directly derived from a single local feature; these are considered global features.

[0042] In some embodiments, the external force prediction model is a CNN-Transformer model, which includes a CNN module and a Transformer module. The CNN module, leveraging the local connectivity and weight sharing characteristics of convolution operations, slides convolution kernels of different sizes across the image to extract detailed local features such as surface cracks and stress concentration areas from the composite panel component image. The Transformer module, using self-attention and multi-head attention mechanisms, breaks the limitations of local neighborhoods and calculates the correlation weights of features at different locations in the image, achieving long-range dependency modeling of global features such as the overall deformation trend of the panel and the collaborative relationship between components. Finally, by concatenating local and global features and dynamically adjusting their weights using a gating mechanism, the two features complement each other. After nonlinear transformation through multiple fully connected layers, the predicted external force value for each image in the target image set is obtained.

[0043] The process of inputting the target image set into the trained external force prediction model to obtain the predicted external force value for each image in the target image set includes the following steps: (1) Input the target image set into the CNN module and use CNN to extract local features of the wall panel component image.

[0044] (2) Input the target image set into the Transformer module and use Transformer to extract the global features of the wall panel component image.

[0045] (3) By using a gating mechanism to fuse local features and global features, local detailed information and global structural information can be effectively integrated.

[0046] The formula for the gating mechanism is shown below:

[0047]

[0048]

[0049] in, These represent the fused features, the local features extracted by the CNN, and the global features extracted by the Transformer, respectively. To concatenate local and global features, W is the weight matrix and b is the bias term. For the Sigmoid function, For gating mechanism, This is a splicing feature.

[0050] (4) The fused features are passed through a fully connected network to obtain the predicted external force value for each image.

[0051] In this embodiment, by acquiring a target image set and inputting it into a trained CNN-Transformer model, the CNN-Transformer can extract spatial features from the images, capture local deformation details and global structural information of the wall panel, and optimize and fuse local and global features through a gating mechanism, thereby improving the accuracy and safety of predicting the ultimate bearing capacity of composite material wall panel components.

[0052] Step 130: Based on the predicted external force value corresponding to each image, obtain the external force-time variation curve; Step 140: Train the ultimate bearing capacity prediction model based on the external force-time variation curve, and perform iterative prediction through the ultimate bearing capacity prediction model to obtain the predicted value of the ultimate bearing capacity of the composite material wall panel assembly.

[0053] For example, using the curve of external force changing over time as a training set, a prediction model for ultimate bearing capacity is constructed. Based on the external force data of the first ten time units, the model predicts the external force for the eleventh time unit. Then, the input is updated according to the value predicted by the ultimate bearing capacity prediction model. Iteration stops when the increase at the next time step is less than 10%, thus obtaining the ultimate bearing capacity. The specific steps include the following: (1) Using the external force data obtained from the external force prediction model as input (external force for 10 consecutive units of time), the external force for the 11th unit of time is obtained; (2) Remove the data from the first unit time in the input data, add the predicted external force to the end of the input data to obtain a new set of inputs, and predict the external force for the 12th unit time again through the ultimate bearing capacity prediction model.

[0054] (3) Repeat (2) continuously until the external force no longer increases or the increase is less than 10% in the next unit time, and then stop the prediction to obtain the predicted value of the ultimate bearing capacity of the composite material wall panel assembly.

[0055] According to the method for predicting the ultimate bearing capacity of composite material wall panel components provided in this application, multiple stress images of the composite material wall panel components are acquired by an acquisition device, which can quickly obtain the overall deformation information of the wall panel, reduce interference with the structure, improve data acquisition efficiency and convenience, and realize non-contact acquisition. By using an external force prediction model to perform local and global feature depth extraction and analysis on the tensile images of the composite material wall panel components, the predicted external force value corresponding to each image in the target image set is obtained. This value is then used as input to the ultimate bearing capacity prediction model, and the ultimate bearing capacity is obtained through iterative prediction. This achieves efficient conversion from image data to mechanical performance indicators, and improves the accuracy and reliability of the prediction of the ultimate bearing capacity of composite material wall panel components.

[0056] In some embodiments, the step of iteratively predicting the ultimate bearing capacity of the composite material wall panel assembly using the ultimate bearing capacity prediction model includes: The external force data for 10 unit time intervals is obtained as an input sequence. The input sequence is then input into the ultimate bearing capacity prediction model to obtain the predicted value of the external force for the next unit time interval. Add the predicted value of the external force for the next unit time to the end of the original input sequence, and delete the external force data for the first unit time at the beginning of the original input sequence to obtain the updated input sequence. The updated input sequence is fed into the ultimate bearing capacity prediction model for iterative prediction to obtain the predicted value of the ultimate bearing capacity of the composite panel assembly.

[0057] In some embodiments, the step of inputting the updated input sequence into the ultimate bearing capacity prediction model for iterative prediction to obtain the predicted value of the ultimate bearing capacity of the composite material wall panel assembly includes: The updated input sequence is fed into the ultimate bearing capacity prediction model to obtain the predicted value of the external force per unit time. Calculate the predicted value of the external force in the next unit of time and the increase in the current value of the external force; When the increase is less than the preset threshold, the predicted value of the external force in the next unit time will be used as the predicted value of the ultimate bearing capacity of the composite material wall panel component. When the increase is greater than or equal to the preset threshold, the predicted value of the external force in the next unit time is added to the end of the original input sequence, the external force data in the first unit time at the beginning of the original input sequence is deleted, and the updated input sequence is obtained. Then, the process is repeated to input the updated input sequence into the ultimate bearing capacity prediction model to obtain the predicted value of the external force in the next unit time.

[0058] For example, data on the tensile force variation over time of composite material wall panel components under different working conditions were acquired, with data collected every 0.1 seconds. In the data preprocessing stage, Gaussian smoothing was used to reduce noise in the original tensile force-time curves. The Gaussian smoothing algorithm, based on the Gaussian kernel function, effectively suppresses noise interference and preserves the true trend of signal variation by weighted averaging of data points and their neighborhood values. The window size for Gaussian smoothing was set to 61.

[0059] The time unit was set to 0.5 seconds. This setting takes into account various factors, including the characteristics of stress changes in composite panel components and the frequency of data acquisition. Because in actual tensile processes, the mechanical properties of composite panel components do not change instantaneously, but evolve gradually over a certain time scale. Setting the time unit to 0.5 seconds captures the key characteristics of material stress changes without making the data too fragmented, thus ensuring data continuity and integrity.

[0060] Based on this unit of time setting, the originally continuous set of data was reasonably divided into 5 groups. This grouping method is to prepare for subsequent model training and validation, following the common principles of training and validation set partitioning in machine learning, to ensure that the model has good generalization ability.

[0061] Next, one set of data will be selected from these five sets as the validation set, and the other four sets will be used as the training set. The purpose of the validation set is to evaluate the trained model, check its performance on unseen data, and thus determine the model's accuracy and stability. The training set, on the other hand, is used for the model's learning process, allowing the model to learn from these data the pattern of tensile force changing over time when the composite panel component is under tension.

[0062] To enable the model to better learn the time-series features in the data, a sliding window operation was applied to both the training and validation sets, with a window size of 10. The sliding window operation is a common technique in time-series data processing; it acts like a moving window that slides across the data sequence, capturing a portion of the data within the window as a sample each time. Specifically, for both the training and validation sets, the window starts from the beginning of the data and moves forward one unit of time each time, capturing a sample of data for 10 units of time (i.e., 5 seconds). This sliding window operation generates multiple samples from the original data sequence, each containing tensile force data for 10 consecutive time units. These samples serve as input to the model, allowing it to learn the patterns and trends of tensile force changes over different time periods. The tensile force at the eleventh time unit is the output.

[0063] For the already trained LSTM prediction model, the ultimate bearing capacity prediction was carried out. The specific procedure was as follows: From the tensile test data of composite panel components under nine different working conditions, validation set data was selected. The tensile force data of the first ten time units of each validation set were used as initial inputs and fed into the trained LSTM prediction model. Based on the input data, the model outputs the predicted value of the tensile force for the eleventh time unit.

[0064] Subsequently, the input data is updated. The predicted tensile force data for the 11th time unit is appended to the end of the original input data sequence, while the first value of the original input data sequence is removed, thus forming a new input sequence containing tensile force data for ten time units. The new input sequence is then fed back into the LSTM prediction model for the next round of tensile force prediction.

[0065] Stopping condition determination: After each iteration of prediction, calculate the increase in the current predicted value compared to the previous predicted value. The formula for calculating the increase is:

[0066] The above data update and prediction process is repeated continuously, monitoring the changes in the tensile force increase in real time. When the increase is less than 10%, the external force growth is considered to have stabilized and reached its limit, at which point the iterative prediction stops. The external force value obtained from the last prediction at this point is the predicted ultimate bearing capacity of the composite material wall panel assembly.

[0067] The relative errors of the ultimate bearing capacity on the verification set under nine different working conditions are shown in Table 1.

[0068] Table 1

[0069] As shown in Table 1, the LSTM prediction model performs well on the validation set, with the largest error in the fifth working condition being 0.0232. The average relative error on the validation set under the nine different working conditions is 0.0123.

[0070] In this embodiment, by inputting the updated input sequence into the ultimate bearing capacity prediction model and calculating the increase of the predicted value of the external force in the next unit time and the increase of the external force value at the current time, the input of the ultimate bearing capacity prediction model can be dynamically adjusted based on the increase, thereby effectively improving the adaptability and accuracy of the ultimate bearing capacity prediction model in the ultimate bearing capacity prediction process.

[0071] In this embodiment, by acquiring external force data over a certain time period and inputting it into the ultimate bearing capacity prediction model, the change in external force of the composite material wall panel assembly in the next time unit can be predicted. By continuously updating the input sequence and performing iterative prediction, the predicted value of the bearing capacity can be updated in real time, resulting in a more accurate predicted value of the ultimate bearing capacity. This model has better applicability, can more comprehensively mine data features, and improves the accuracy of the ultimate bearing capacity prediction for the composite material wall panel assembly.

[0072] In some embodiments, the training process of the external force prediction model includes: Construct a pre-defined external force prediction model; Multiple stress images of the composite material wall panel assembly were acquired and preprocessed to obtain the first dataset; Perform a sliding window operation on the first dataset to obtain a sliding window dataset; Based on the sliding window dataset, the preset external force prediction model is trained using the mean squared error loss function to obtain the trained external force prediction model.

[0073] The process is easy to understand: build a CNN-Transformer prediction model, acquire multiple stress images of the composite panel component and preprocess them to obtain the first dataset, where the label is external force. The first dataset includes photos of the panel component and load data. The load data includes the external force corresponding to the time each photo was taken, as well as the specific time each photo was taken.

[0074] The images in the first dataset are subjected to a sliding window operation to obtain a sliding window dataset. The window size is 4, meaning that the images at 4 adjacent time steps are used as input and the external force at the last time step is used as output.

[0075] For example, a CNN-Transformer model is trained using a dataset containing 2868 groups of sliding window data under different working conditions to obtain a trained CNN-Transformer prediction model, which can predict the external force corresponding to each image. The optimizer uses the Adam optimizer, with a learning rate of 0.001, a decay rate of 0.9 for first-order moment estimation, a decay rate of 0.99 for second-order moment estimation, a weight decay rate of 0.0005, and a loss function of mean squared error. The maximum number of early stopping events is 30.

[0076] In some embodiments, a CNN network was trained using the same training settings and training set. The average relative errors of the CNN model and the CNN-Transformer model were validated using data from nine different testing scenarios, as shown in Table 2. The CNN model had the largest average relative error in the first testing scenario (0.4283), and the smallest average relative error in the ninth testing scenario (0.0379). The average relative error of the CNN model on the validation set across the nine different testing scenarios was 0.1242. The CNN-Transformer model had the largest error in the third testing scenario (0.1512), and the smallest average relative error in the eighth testing scenario (0.0306). The average relative error of the CNN-Transformer model on the validation set across the nine different testing scenarios was 0.0779.

[0077] Table 2

[0078] As shown in Table 2, the CNN-Transformer model outperforms the CNN model in 6 out of 9 different working conditions on the validation set. The average relative error on the validation set of the nine different working conditions is 37.3% higher than that of the CNN model.

[0079] In this embodiment, by acquiring and preprocessing multiple stress images of the composite material wall panel assembly, a first dataset is obtained. A sliding window dataset is then generated through a sliding window operation, which effectively enhances the quality of the training data. The preset external force prediction model is trained using the mean square error loss function, which enables the prediction of the stress condition of the composite material wall panel assembly and improves the accuracy and reliability of external force prediction.

[0080] In some embodiments, training the ultimate bearing capacity prediction model based on the external force-time variation curve includes: A preset ultimate bearing capacity prediction model is constructed, wherein the preset ultimate bearing capacity prediction model is an LSTM prediction model; Multiple data points are obtained based on the external force-time variation curve. The multiple data points are then cleaned to obtain cleaned data points. The 3σ principle is used to identify and remove the cleaned data points, resulting in the removed data points. Linear interpolation is used to fill in the removed data points to obtain a second dataset; Based on the second dataset, the preset ultimate bearing capacity prediction model is trained using the mean squared error loss function to obtain the trained ultimate bearing capacity prediction model.

[0081] The training process of the ultimate bearing capacity prediction model includes the following steps: (1) Construct a preset ultimate bearing capacity prediction model, which is an LSTM prediction model; (2) Based on the external force-time change curve, multiple data points are obtained. The data points are cleaned to remove outliers and noise. Statistical analysis methods, such as the 3σ principle, are used to identify and remove data points that deviate significantly from the normal range. For missing data, linear interpolation can be used to fill in the missing data based on the trend of the preceding and following data to obtain the second dataset. (3) Determine key parameters such as the number of network layers and the number of hidden layer neurons. For example, selecting 2-3 LSTM layers can better capture the long-term dependencies of time series data; the number of hidden layer neurons can be adjusted according to the complexity of the data and computing resources, generally between 64-256. The mean squared error is selected as the loss function to measure the difference between the model's predicted value and the true value. The Adam optimizer is selected, whose adaptive learning rate adjustment mechanism can automatically adjust the learning rate during training to accelerate the convergence speed of the model. The initial learning rate is set to 0.001 and adjusted in a timely manner according to the training situation. The exhaustion rate of the first moment estimation is set to 0.9, the exhaustion rate of the second moment estimation is 0.99, the weight decay is set to 0.0005, and the maximum number of early stopping mechanisms is 10. Based on the second dataset, the preset ultimate bearing capacity prediction model is trained using the mean squared error loss function to obtain the trained ultimate bearing capacity prediction model.

[0082] Figure 4 This is a second schematic flowchart of the method for predicting the ultimate bearing capacity of composite material wall panel components provided in this application embodiment, as shown below. Figure 4 As shown, the method includes the following steps: Multiple stress images of composite material wall panel components are collected by acquisition equipment to obtain an image dataset, and the image dataset is preprocessed to obtain the target image set; The target image set is input into the trained CNN-Transformer model to obtain the predicted external force value for each image in the target image set. Based on the predicted external force value for each image, the external force-time change curve is obtained. (3) Organize the external force-time variation curves into a training set, ensuring that the data are arranged in chronological order and that the data format meets the model input requirements, so as to provide basic data for subsequent model training; (4) Construct an LSTM prediction model, and set the hyperparameters such as the number of hidden layers and the number of neurons in the model in a reasonable manner according to the task requirements, so as to build a prediction framework that can process time series data. (5) Based on the external force data of the first ten units of time in the training set, input it into the constructed LSTM prediction model, and output the predicted value of the external force for the eleventh unit of time through the calculation and learning of the model. (6) Use the external force values ​​predicted by the model in (5) as new input data, and combine them with the original time series data to form a new input sequence. Input the data of the first ten time units in the new sequence into the model again to obtain the predicted external force value for the next time unit (i.e., the twelfth time unit). Repeat this process of updating data and inputting it into the model for prediction; (7) After each iteration, calculate the increase of the predicted external force value at the next moment compared to the current external force value. When the increase is less than 10%, stop the iteration process. The external force value obtained in the last prediction is the predicted value of the ultimate bearing capacity of the composite material wall panel assembly.

[0083] In this embodiment, a pre-defined ultimate bearing capacity prediction model is constructed, and multiple data points are obtained based on the external force-time variation curve. The data is then cleaned, filtered, and interpolated to obtain a second dataset, effectively improving the quality of the model training data. By training the pre-defined ultimate bearing capacity prediction model using the mean square error loss function, the prediction of the ultimate bearing capacity of composite material wall panel components can be achieved, improving the accuracy and reliability of the ultimate bearing capacity prediction model.

[0084] The method for predicting the ultimate bearing capacity of composite material wall panel components provided in this application can be executed by a device for predicting the ultimate bearing capacity of composite material wall panel components. This application uses the example of a device for predicting the ultimate bearing capacity of composite material wall panel components executing the method for predicting the ultimate bearing capacity of composite material wall panel components to illustrate the device for predicting the ultimate bearing capacity of composite material wall panel components provided in this application.

[0085] This application also provides a device for predicting the ultimate bearing capacity of composite material wall panel components, such as... Figure 5 As shown, the device for predicting the ultimate bearing capacity of the composite material wall panel assembly includes: an acquisition module 510, a first processing module 520, a second processing module 530, and a prediction module 540.

[0086] The acquisition module 510 is used to acquire multiple stress images of the composite material wall panel assembly through the acquisition device to obtain an image dataset. Each stress image corresponds to a time step. Each stress image in the image dataset is preprocessed to obtain a target image set. The first processing module 520 is used to input the target image set into the trained external force prediction model to obtain the external force prediction value corresponding to each image in the target image set. The second processing module 530 is used to obtain the external force-time change curve based on the predicted external force value corresponding to each image; The prediction module 540 is used to train the ultimate bearing capacity prediction model based on the external force-time variation curve, and to perform iterative prediction through the ultimate bearing capacity prediction model to obtain the predicted value of the ultimate bearing capacity of the composite material wall panel assembly.

[0087] According to the method for predicting the ultimate bearing capacity of composite material wall panel components provided in this application, multiple stress images of the composite material wall panel components are acquired by an acquisition device, which can quickly obtain the overall deformation information of the wall panel, reduce interference with the structure, improve data acquisition efficiency and convenience, and realize non-contact acquisition. By using an external force prediction model to perform local and global feature depth extraction and analysis on the tensile images of the composite material wall panel components, the predicted external force value corresponding to each image in the target image set is obtained. This value is then used as input to the ultimate bearing capacity prediction model, and the ultimate bearing capacity is obtained through iterative prediction. This achieves efficient conversion from image data to mechanical performance indicators, and improves the accuracy and reliability of the prediction of the ultimate bearing capacity of composite material wall panel components.

[0088] The device for predicting the ultimate bearing capacity of composite material wall panel components provided in this application embodiment can achieve... Figures 1 to 5 The various processes implemented in the method for predicting the ultimate bearing capacity of composite material wall panel components will not be described again here to avoid repetition.

[0089] In some embodiments, such as Figure 6 As shown, this application embodiment also provides an electronic device 600, including a processor 601, a memory 602, and a computer program stored in the memory 602 and executable on the processor 601. When the program is executed by the processor 601, it implements the various processes of the above-described method embodiment for predicting the ultimate bearing capacity of composite material wall panel components and can achieve the same technical effect. To avoid repetition, it will not be described again here.

[0090] It should be noted that the electronic devices in the embodiments of this application include the mobile electronic devices and non-mobile electronic devices described above.

[0091] This application also provides a non-transitory computer-readable storage medium storing a computer program. When the computer program is executed by a processor, it implements the various processes of the above-described method embodiment for predicting the ultimate bearing capacity of composite material wall panel components and achieves the same technical effect. To avoid repetition, it will not be described again here.

[0092] The processor is the processor in the electronic device described in the above embodiments. The readable storage medium includes computer-readable storage media, such as computer read-only memory (ROM), random access memory (RAM), magnetic disk, or optical disk.

[0093] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the above-described method for predicting the ultimate bearing capacity of composite material wall panel components.

[0094] The processor is the processor in the electronic device described in the above embodiments. The readable storage medium includes computer-readable storage media, such as computer read-only memory (ROM), random access memory (RAM), magnetic disk, or optical disk.

[0095] This application embodiment also provides a chip, which includes a processor and a communication interface. The communication interface and the processor are coupled. The processor is used to run programs or instructions to implement the various processes of the above-described method embodiment for predicting the ultimate bearing capacity of composite material wall panel components, and can achieve the same technical effect. To avoid repetition, it will not be described again here.

[0096] It should be understood that the chip mentioned in the embodiments of this application may also be referred to as a device-level chip, device chip, chip device, or on-chip device chip, etc.

[0097] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element. Furthermore, it should be noted that the scope of the methods and apparatuses in the embodiments of this application is not limited to performing functions in the order shown or discussed, but may also include performing functions substantially simultaneously or in the reverse order, depending on the functions involved. For example, the described methods may be performed in a different order than described, and various steps may be added, omitted, or combined. Additionally, features described with reference to certain examples may be combined in other examples.

[0098] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a computer software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) and includes several instructions to cause a terminal (which may be a mobile phone, computer, server, or network device, etc.) to execute the method for predicting the ultimate bearing capacity of composite material wall panel components of various embodiments of this application.

[0099] In the description of this application, "first feature" and "second feature" may include one or more of the features.

[0100] In the description of this application, "multiple" means two or more.

[0101] The embodiments of this application have been described above with reference to the accompanying drawings. However, this application is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of this application without departing from the spirit and scope of the claims, and all of these forms are within the protection scope of this application.

[0102] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "illustrative embodiment," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of this application. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0103] Although embodiments of this application have been shown and described, those skilled in the art will understand that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of this application, the scope of which is defined by the claims and their equivalents.

Claims

1. A method for predicting the ultimate bearing capacity of a composite material wall panel assembly, characterized in that, The method includes: Multiple stress images of the composite material wall panel assembly are acquired by the acquisition device to obtain an image dataset. Each stress image corresponds to a time step. Each stress image in the image dataset is preprocessed to obtain a target image set. The target image set is input into the trained external force prediction model to obtain the external force prediction value corresponding to each image in the target image set; Based on the predicted external force value corresponding to each image, the external force-time variation curve is obtained; The ultimate bearing capacity prediction model is trained based on the external force-time variation curve, and the predicted value of the ultimate bearing capacity of the composite material wall panel assembly is obtained by iterative prediction through the ultimate bearing capacity prediction model.

2. The method for predicting the ultimate bearing capacity of composite material wall panel components according to claim 1, characterized in that, The step of preprocessing each stress image in the image dataset to obtain the target image set includes: A mask is created based on the part of the material panel that deforms most significantly under external force in each stress image, and the image parts outside the mask are all set to black. The black portion of each stress image is cropped to obtain the cropped stress image; By smoothing the boundaries of each stress image after cropping using the dilation operation, the edges of each stress image are made smoother and the images are more continuous, resulting in the target image set.

3. The method for predicting the ultimate bearing capacity of composite material wall panel components according to claim 1, characterized in that, The step of inputting the target image set into the trained external force prediction model to obtain the predicted external force value for each image in the target image set includes: The target image set is input into the CNN module to obtain the local features of each image in the target image set; The target image set is input into the Transformer module to obtain the global features of each image in the target image set; Based on the gating mechanism, the local and global features of each image are fused to obtain the fused features corresponding to each image; The fusion features corresponding to each image are input into a fully connected network to obtain the predicted external force value for each image in the target image set. The external force prediction model is a CNN-Transformer model, which includes a CNN module and a Transformer module.

4. The method for predicting the ultimate bearing capacity of composite material wall panel components according to claim 1, characterized in that, The step of iteratively predicting the ultimate bearing capacity of the composite material wall panel assembly using the ultimate bearing capacity prediction model includes: The external force data for 10 unit time intervals is obtained as an input sequence. The input sequence is then input into the ultimate bearing capacity prediction model to obtain the predicted value of the external force for the next unit time interval. Add the predicted value of the external force for the next unit time to the end of the original input sequence, and delete the external force data for the first unit time at the beginning of the original input sequence to obtain the updated input sequence. The updated input sequence is fed into the ultimate bearing capacity prediction model for iterative prediction to obtain the predicted value of the ultimate bearing capacity of the composite panel assembly.

5. The method for predicting the ultimate bearing capacity of composite material wall panel components according to claim 4, characterized in that, The step of inputting the updated input sequence into the ultimate bearing capacity prediction model for iterative prediction to obtain the predicted value of the ultimate bearing capacity of the composite material wall panel assembly includes: The updated input sequence is fed into the ultimate bearing capacity prediction model to obtain the predicted value of the external force per unit time. Calculate the predicted value of the external force in the next unit of time and the increase in the current value of the external force; When the increase is less than the preset threshold, the predicted value of the external force in the next unit time will be used as the predicted value of the ultimate bearing capacity of the composite material wall panel assembly. When the increase is greater than or equal to the preset threshold, the predicted value of the external force in the next unit time is added to the end of the original input sequence, the external force data in the first unit time at the beginning of the original input sequence is deleted, and the updated input sequence is obtained. Then, the process is repeated to input the updated input sequence into the ultimate bearing capacity prediction model to obtain the predicted value of the external force in the next unit time.

6. The method for predicting the ultimate bearing capacity of composite material wall panel components according to claim 1, characterized in that, The training process of the external force prediction model includes: Construct a pre-defined external force prediction model; Multiple stress images of the composite material wall panel assembly were acquired and preprocessed to obtain the first dataset; Perform a sliding window operation on the first dataset to obtain a sliding window dataset; Based on the sliding window dataset, the preset external force prediction model is trained using the mean squared error loss function to obtain the trained external force prediction model.

7. The method for predicting the ultimate bearing capacity of composite material wall panel components according to claim 1, characterized in that, The training of the ultimate bearing capacity prediction model based on the external force-time variation curve includes: A preset ultimate bearing capacity prediction model is constructed, wherein the preset ultimate bearing capacity prediction model is an LSTM prediction model; Multiple data points are obtained based on the external force-time variation curve. The multiple data points are then cleaned to obtain cleaned data points. The 3σ principle is used to identify and remove the cleaned data points, resulting in the removed data points. Linear interpolation is used to fill in the removed data points to obtain a second dataset; Based on the second dataset, the preset ultimate bearing capacity prediction model is trained using the mean squared error loss function to obtain the trained ultimate bearing capacity prediction model.

8. A device for predicting the ultimate bearing capacity of a composite material wall panel assembly, implemented using the method for predicting the ultimate bearing capacity of a composite material wall panel assembly as described in any one of claims 1 to 7, characterized in that, The device includes: The acquisition module is used to acquire multiple stress images of composite material wall panel components through an acquisition device to obtain an image dataset. Each stress image corresponds to a time step. Each stress image in the image dataset is preprocessed to obtain a target image set. The first processing module is used to input the target image set into the trained external force prediction model to obtain the external force prediction value corresponding to each image in the target image set; The second processing module is used to obtain the external force-time variation curve based on the predicted external force value corresponding to each image; The prediction module is used to train the ultimate bearing capacity prediction model based on the external force-time variation curve, and to perform iterative prediction through the ultimate bearing capacity prediction model to obtain the predicted value of the ultimate bearing capacity of the composite material wall panel assembly.

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 method for predicting the ultimate bearing capacity of the composite material wall panel assembly as described in any one of claims 1 to 7.

10. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the method for predicting the ultimate bearing capacity of the composite material wall panel assembly as described in any one of claims 1 to 7.

Images