A method and system for intelligent self-optimization and prediction of pre-tightening force of high-strength bolt connection with corrosion

Abstract

This invention discloses an intelligent self-optimizing prediction method and system for the preload of rusted high-strength bolt connections, belonging to the field of bolt preload detection technology. Specifically, it includes: establishing a finite element database of bolt connection preload loss and training a machine learning prediction model based on a neural network; when the recognition accuracy of the machine learning prediction model reaches a preset value, real-time monitoring of the preload loss of the target bolt is performed to obtain real-time geometric feature information of the target bolt connection; when the similarity between the real-time geometric feature information of the target bolt connection and the finite element model in the finite element database is greater than or equal to a preset value, the preload loss from the existing finite element database is retrieved; when the similarity is lower than a preset value, the actual bolt connection preload loss is output in real-time using the trained machine learning prediction model; this achieves automated, marker-free detection of bolt looseness, with high tolerance for bolt image acquisition angles and distances, and can meet the requirements for bolt looseness detection at different locations.

Description

Technical Field

[0001] This invention belongs to the field of bolt preload detection technology, specifically relating to an intelligent self-optimization prediction method and system for preload of rusted high-strength bolt connections. Background Technology

[0002] As a key connecting element in steel structure systems, bolts are widely used in the construction and installation of various steel structure projects, such as buildings, bridges, and towers, due to their excellent load-bearing performance, fatigue resistance, seismic resistance, high-strength connections, ease of construction, and maintainability. However, with the increase in service time, especially in the harsh marine and industrial atmospheric corrosion environments, corrosion of bolted connections has become increasingly common and severe, posing a significant hidden danger to the structural performance and safety of steel structures.

[0003] High-strength bolts exposed to corrosive environments for extended periods, especially those in structures with a service life of 10 years or more, often exhibit extensive and deep corrosion, manifesting as noticeable blistering, cracking, and spalling. Taking the Nanpanjiang Bridge as an example, after 10 years of construction, approximately 80% of the bolted connections in its main truss showed moderate to severe corrosion, significantly more severe than that of the structural members. This corrosion not only damages the bolt surface material and reduces the effective cross-sectional area of ​​the structural materials but also directly leads to a significant loss of bolt preload. Preload is a core element in maintaining the stability and load-bearing capacity of high-strength bolt connections, and its reduction is closely related to the degree of corrosion of the bolt head and nut. A substantial decrease in preload directly affects the stiffness, load-bearing capacity, and fatigue life of the connection, thereby jeopardizing the stability and safety of the entire steel structure system.

[0004] Faced with the severe challenges posed by corroded bolt connections, existing inspection methods often fall short of meeting the demands for efficiency, accuracy, and economy. Traditional inspection methods, such as manual visual inspection, hammer testing, and torque wrench measurement, while simple to operate, are inefficient, reliant on human experience, susceptible to subjective influences, and pose safety hazards when inspecting bolts at heights or inaccessible areas. Emerging acoustic, electrical, and magnetic sensor detection technologies, while theoretically providing more refined data, are expensive, sensitive to environmental conditions, and their effectiveness in practical applications is less than ideal. Image recognition-based detection technologies also suffer from issues such as easily damaged markers and limited image acquisition angles, making them unsuitable for complex field conditions. Summary of the Invention

[0005] The technical problem to be solved by the present invention is to provide an intelligent self-optimization prediction method and system for preload of high-strength rusted bolts, which addresses the shortcomings of the prior art. This method solves the technical problem of the limitations of existing detection of preload loss of various rusted bolts, realizes automated and marker-free detection of bolt looseness, has high tolerance for bolt image acquisition angle and distance, and can meet the requirements of bolt looseness detection in different positions.

[0006] The present invention adopts the following technical solution:

[0007] A method for intelligent self-optimization prediction of preload in high-strength rusted bolt connections includes the following steps:

[0008] Establish a finite element database of preload loss in bolted connections and train a machine learning prediction model based on neural networks;

[0009] Determine whether the recognition accuracy of the obtained machine learning prediction model reaches the preset value;

[0010] When the recognition accuracy of the machine learning prediction model reaches the preset value, the preload loss of the target bolt is monitored in real time to obtain the real-time geometric feature information of the target bolt connection.

[0011] The similarity between the real-time geometric feature information of the target bolt connection and the finite element model in the finite element database is compared. When the similarity is greater than or equal to the preset value, the preload loss of the existing finite element database is called.

[0012] When the similarity is lower than the preset value, the trained machine learning prediction model is used to output the actual bolt connection preload loss in real time.

[0013] Preferably, establishing a database of bolt connection preload loss specifically involves:

[0014] Establish a three-dimensional solid geometric model of the uncorroded bolt connection;

[0015] The first analysis model is obtained by dividing the three-dimensional solid geometric model of the uncorroded bolt connection into elements and applying boundary conditions.

[0016] Apply the first load to the first analysis model and calculate the preload of the first analysis model;

[0017] Based on the dimensions of the uncorroded bolt connections and the actual principles of bolt corrosion, components are randomly defined as corrosion units.

[0018] The relevant physical properties of bolt connections are randomly set based on the actual bolt corrosion principle;

[0019] The modification units are killed using the birth and death unit method to obtain the second analysis model;

[0020] Apply the second step load to the second analysis, calculate the preload of the second analysis model, and calculate the preload loss value of the corroded bolt;

[0021] Record the relevant dimensions, physical information, and preload loss values ​​of the corroded bolts;

[0022] Repeat the above steps, organize the data obtained each time, establish a database of bolt connection preload loss, and train a machine learning prediction model based on neural networks.

[0023] Preferably, the prediction model based on the neural network uses the particle swarm optimization algorithm to optimize the BP neural network.

[0024] Preferably, when the recognition accuracy of the prediction model does not reach the preset value, the finite element database is added, and the training of the neural network-based machine learning prediction model is returned.

[0025] Preferably, obtaining the real-time geometric feature information of the bolt connection specifically involves:

[0026] The obtained bolt connection images are converted to grayscale to obtain grayscale images;

[0027] The grayscale image is denoised by Gaussian filtering to obtain the denoised image.

[0028] The binarization threshold is determined using the minimum error method, and the denoised image is binarized to obtain a binary image.

[0029] Harris corner detection is used to identify and mark the center of the bolt head;

[0030] Edge detection technology is used to accurately define the overall outline of the bolt, ensuring that the bolt boundary can still be accurately defined even under the influence of corrosion;

[0031] Compile and output the geometric dimensions of each bolt connection, including diameter, length, and degree of corrosion.

[0032] Preferably, the similarity between the real-time geometric feature information of the target bolt connection and the finite element model in the finite element database is compared. When the similarity is greater than or equal to a preset value, the preload loss from the existing finite element database is called. Specifically:

[0033] The similarity is obtained by comparing the geometric features of the target bolt connection with the dimensional information recorded in the database.

[0034] Set a threshold for the target comparison similarity, and judge the degree of similarity between the model and the real-time bolt connection based on the comparison similarity;

[0035] If the value exceeds the threshold, the preload loss from the existing finite element database is used directly as the prediction result.

[0036] Preferably, the real-time output of the actual bolt connection preload loss using a trained machine learning prediction model is as follows:

[0037] The geometric features and related physical information of the target bolt are normalized and regularized and input into the machine learning model to predict the preload. If the preload loss exceeds the set value, a signal is sent to notify maintenance personnel to repair the target bolt; otherwise, normal data is detected in real time and archived.

[0038] Based on the obtained preload results, a finite element model is established by combining the corresponding geometric features and physical information, and a finite element database is added.

[0039] Secondly, embodiments of the present invention provide an intelligent self-optimizing prediction system for the preload of rusted high-strength bolt connections, comprising:

[0040] The data module establishes a finite element database of bolt connection preload loss and trains a neural network-based machine learning prediction model.

[0041] The judgment module determines whether the recognition accuracy of the machine learning prediction model has reached the preset value.

[0042] The monitoring module monitors the preload loss of the target bolt in real time and obtains real-time geometric feature information of the target bolt connection when the recognition accuracy of the machine learning prediction model reaches a preset value.

[0043] The comparison module compares the real-time geometric feature information of the target bolt connection with the finite element model in the finite element database. When the similarity is greater than or equal to the preset value, it calls the preload loss from the existing finite element database.

[0044] The prediction module uses a trained machine learning prediction model to output the actual bolt connection preload loss in real time when the similarity is lower than the preset value.

[0045] Thirdly, a computer device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the computer program, implements the steps of the above-described intelligent self-optimization prediction method for preload of rust-resistant high-strength bolt connections.

[0046] Fourthly, embodiments of the present invention provide a computer-readable storage medium including a computer program, which, when executed by a processor, implements the steps of the above-described intelligent self-optimization prediction method for preload of rusted high-strength bolt connections.

[0047] Compared with the prior art, the present invention has at least the following beneficial effects:

[0048] A method for intelligent self-optimizing prediction of preload in high-strength rusted bolt connections is proposed. First, a preload loss database is established using neural network technology, and a prediction model is trained to capture complex correlations between features and improve prediction accuracy. Next, accuracy verification is used to ensure model reliability. Bolt geometric features are monitored and extracted in real time, and combined with database comparison, rapid prediction or analysis using machine learning models can be performed. This process not only optimizes resource allocation and reduces safety risks through a rapid response mechanism, but also continuously enhances the model's generalization ability by integrating new data into the database, forming a closed-loop system of continuous learning and optimization. This fundamentally improves the scientific rigor and foresight of rusted bolt monitoring. The aim is to construct an efficient and adaptive intelligent self-optimizing prediction system for preload in high-strength rusted bolt connections, ensuring structural safety and maintenance efficiency.

[0049] Furthermore, by establishing a database of bolt preload loss, a 3D model of the uncorroded bolt is first created and finite element analysis is performed. Then, the impact of the corrosion process on the mechanical properties of the model is simulated, and the changes in preload before and after corrosion are recorded, ultimately accumulating to form a database. This step aims to capture the specific impact of corrosion on bolt preload through accurate simulation and data analysis. Its core advantage lies in providing a quantitative assessment method that can accurately predict and prevent preload loss due to bolt corrosion, ensuring structural safety. In principle, it combines computer-aided design, finite element analysis, and materials corrosion science. By gradually applying loads and adjusting corrosion parameters, it reflects the degradation mechanism of bolt performance in a real environment. The establishment of the database provides a solid data foundation for further statistical analysis, predictive model optimization, and practical engineering applications.

[0050] Furthermore, a particle swarm optimization (PSO-BP) algorithm is employed to optimize the backpropagation (BP) neural network used in the prediction model. This aims to leverage the global search advantage of PSO and the powerful fitting ability of BP neural networks to significantly improve the accuracy and efficiency of intelligent prediction of preload in corroded high-strength bolt connections. The PSO algorithm, by simulating swarm intelligence, guides the efficient optimization of network weights, overcoming the limitation of traditional BP neural networks easily getting trapped in local optima and achieving a closer approximation of the global optimum. Optimized by the PSO algorithm, the BP neural network, through multi-level feedforward and backpropagation, can better learn the complex patterns implicit in the input data, especially the nonlinear mapping relationship between corrosion features and preload loss. Unlike CNNs, which rely on convolutional and pooling layers to capture local features and reduce dimensionality, the PSO-BP neural network focuses on in-depth learning of the overall data, dynamically adjusting the weight matrix through particle swarm optimization to achieve effective feature integration and optimization. This process not only enhances the model's sensitivity to corrosion features but also, through continuous weight updates, enables the model to learn and adapt to various corrosion conditions in a large number of training samples, ensuring the broad applicability and high reliability of the prediction results and providing a scientific basis for structural safety assessment and maintenance decisions.

[0051] Furthermore, when the prediction model's accuracy falls below a preset standard, the training process is repeated by adding a finite element database to enrich data diversity and improve the model's generalization ability. This strategy leverages real physical data from finite element analysis to enhance the model's understanding of complex corrosion conditions, optimizes model parameters to mitigate overfitting, and further improves prediction accuracy to the desired level. This process integrates the advantages of physical simulation and machine learning, ensuring that the intelligent self-optimizing prediction of preload in high-strength rusted bolt connections is more accurate and effective.

[0052] Furthermore, the techniques used in acquiring real-time geometric feature information of bolted connections, including image grayscale conversion, noise reduction, binarization, corner detection, and edge recognition, aim to automatically extract the geometric dimensions and corrosion status of bolts through precise image processing, ensuring the real-time nature and accuracy of the data. This series of processes not only accelerates the detection process and improves monitoring efficiency but also provides solid data support for the preload loss prediction model by accurately defining the bolt outline and assessing the degree of corrosion. By reducing environmental interference through step-by-step image processing, key features are accurately identified, enabling precise measurement of bolt diameter, length, and corrosion areas, thereby providing a scientific basis for maintenance decisions and ensuring structural safety and reliability.

[0053] Furthermore, by comparing the real-time detected bolt geometric features with database records, the system can quickly and accurately predict preload loss, improving prediction efficiency and practicality. The principle behind this is to use feature matching technology to quantify the similarity between the real-time acquired bolt connection geometric data and the model data in the database, and to set threshold distinction rules and processing methods: when the similarity exceeds the threshold, the preload loss value of the matching item in the database is directly used as the prediction result. This utilizes past data to accelerate the decision-making process, ensuring the reliability and timeliness of the prediction. This not only optimizes resource allocation and concentrates efforts on solving complex problems, but also significantly improves the overall efficiency and adaptability of the prediction system through an intelligent screening mechanism.

[0054] Furthermore, by inputting the normalized characteristics and physical information of specific bolts into a machine learning model to predict preload loss, real-time early warnings for abnormal situations and archiving of routine data are achieved, ensuring structural safety while accumulating valuable maintenance data. Furthermore, by using the prediction results and actual characteristics to construct a finite element model and expand the database, not only is the accuracy of real-time prediction optimized, but model optimization is also promoted. By utilizing intelligent technology to improve maintenance efficiency and through continuous data feedback loops, the understanding of the preload loss mechanism of corroded bolts and corresponding countermeasures are deepened, fundamentally strengthening the practical value and long-term benefits of the prediction system.

[0055] It is understood that the beneficial effects of the second to fourth aspects mentioned above can be found in the relevant descriptions in the first aspect mentioned above, and will not be repeated here.

[0056] In summary, this invention achieves efficient and accurate prediction of bolt corrosion, greatly enhancing the timeliness and accuracy of structural safety monitoring, enabling rapid response and providing a scientific basis for maintenance decisions, while ensuring the long-term reliability and practicality of the prediction.

[0057] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0058] Figure 1 This is a schematic diagram of the process of the present invention;

[0059] Figure 2 A schematic diagram of a computer device provided in an embodiment of the present invention;

[0060] Figure 3 This is a block diagram of a chip according to an embodiment of the present invention;

[0061] Figure 4 The BP neural network structure diagram is optimized for the particle swarm optimization algorithm used in this invention;

[0062] Figure 5This is a schematic diagram illustrating the process of acquiring the target bolt connection dimensions using an industrial camera and outputting preload force, as provided in an embodiment of the present invention. Detailed Implementation

[0063] 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, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0064] In the description of this invention, it should be understood that the terms "comprising" and "including" indicate the presence of the described features, integrals, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or collections thereof.

[0065] It should also be understood that the terminology used in this specification is for the purpose of describing particular embodiments only and is not intended to limit the invention. As used in this specification and the appended claims, the singular forms “a,” “an,” and “the” are intended to include the plural forms unless the context clearly indicates otherwise.

[0066] It should also be further understood that the term "and / or" as used in this specification and the appended claims refers to any combination and all possible combinations of one or more of the associated listed items, and includes such combinations. For example, A and / or B can represent three cases: A alone, A and B simultaneously, and B alone. Additionally, the character " / " in this invention generally indicates that the preceding and following objects have an "or" relationship.

[0067] It should be understood that although terms such as first, second, third, etc., may be used in the embodiments of the present invention to describe the preset range, these preset ranges should not be limited to these terms. These terms are only used to distinguish the preset ranges from one another. For example, without departing from the scope of the embodiments of the present invention, the first preset range may also be referred to as the second preset range, and similarly, the second preset range may also be referred to as the first preset range.

[0068] Depending on the context, the word "if" as used here can be interpreted as "when," "when," "in response to determination," or "in response to detection." Similarly, depending on the context, the phrase "if determination" or "if detection (of the stated condition or event)" can be interpreted as "when determination," "in response to determination," "when detection (of the stated condition or event)," or "in response to detection (of the stated condition or event)."

[0069] The accompanying drawings illustrate various structural schematic diagrams according to embodiments disclosed in this invention. These drawings are not to scale, and some details have been enlarged for clarity, and some details may have been omitted. The shapes of the various regions and layers shown in the drawings, as well as their relative sizes and positional relationships, are merely exemplary and may deviate from reality due to manufacturing tolerances or technical limitations. Furthermore, those skilled in the art can design regions / layers with different shapes, sizes, and relative positions as needed.

[0070] This invention provides an intelligent self-optimizing prediction method for the preload of high-strength bolt connections subject to corrosion. By constructing an intelligent self-optimizing prediction system for the preload of high-strength bolt connections subject to corrosion based on neural networks and finite element analysis, it achieves efficient and accurate prediction of bolt corrosion, greatly enhancing the timeliness and accuracy of structural safety monitoring. Utilizing deep learning technology to automatically extract features, combined with real-time monitoring and image processing technology, the system can quickly respond and provide a scientific basis for maintenance decisions. Simultaneously, through continuous learning mechanisms and database optimization, the model's generalization ability under complex corrosion conditions is continuously improved, ensuring the long-term reliability and practicality of the predictions.

[0071] Please see Figure 1 This invention discloses an intelligent self-optimization prediction method for the preload of high-strength rusted bolt connections, comprising the following steps:

[0072] S1. Establish a database of bolt connection preload loss and train a neural network model using the existing data in the database.

[0073] The specific steps for establishing a database of bolt connection preload loss are as follows:

[0074] S101. Use finite element software to establish a three-dimensional solid geometric model of the uncorroded bolt connection;

[0075] S102. Perform element division and apply boundary conditions on the three-dimensional solid geometric model of the uncorroded bolt connection to obtain the first analysis model;

[0076] S103. Apply the first step load to the first analysis model and calculate the preload of the first analysis model;

[0077] S104. Based on the dimensions of the uncorroded bolt connections and the actual bolt corrosion principle, randomly define components as corrosion units;

[0078] S105. Randomly set the relevant physical properties of bolt connections based on the actual bolt corrosion principle;

[0079] S106. Using the birth and death unit method, the modifier units are killed to obtain the second analysis model;

[0080] S107. Apply the second step load to the second analysis, calculate the preload of the second analysis model, and calculate the preload loss value of the corroded bolt.

[0081] S108. Record the relevant dimensions, physical information, and preload loss value of the corroded bolts;

[0082] S109. Repeat the above steps, organize the data obtained each time, and establish a database of bolt connection preload loss. In this case, ANSYS finite element software is used to initially establish a database with 1000 finite element models.

[0083] The neural network model uses the particle swarm optimization algorithm to optimize the BP neural network. The neural network structure is a three-layer network with the number of input neurons equal to the number of target bolted features, 5 hidden layer neurons, and 1 output neuron.

[0084] When simulating actual corrosion using finite element method (FEM) software, it is necessary to consider as much dimensional information as possible to ensure the accuracy of the FEM simulation results.

[0085] The preload loss obtained from the finite element method simulation is used as the output, and multiple physical information is used as the input. The model is trained by dividing the model into training, validation and test sets to obtain a machine learning model.

[0086] S2. Determine whether the recognition accuracy of the prediction model has reached the preset value. If yes, proceed to step S3; otherwise, add the finite element database and return to step S1.

[0087] S3. Real-time monitoring of the preload loss of the target bolt is performed using an image acquisition device and a non-destructive testing device to obtain real-time geometric feature information of the bolt connection;

[0088] In this embodiment of the invention, the image acquisition device uses a CCD industrial camera, which is relatively inexpensive. The image acquisition device is installed near the target bolt connection, and the image is clear and has a slight angle, which can completely capture the relevant geometric dimensions of the bolt connection.

[0089] In this embodiment of the invention, the image acquisition device can acquire bolt connection images when the lighting conditions in the set area are good, such as acquiring one image at each of the following time periods: 8:00, 10:00, 12:00, 14:00, 16:00, and 18:00, and transmitting it to a remote server.

[0090] The specific steps for obtaining real-time geometric feature information of bolted connections are as follows:

[0091] S301. Perform grayscale processing on the obtained bolt connection image to obtain a grayscale image;

[0092] S302. Denoise the grayscale image by using Gaussian filtering to obtain the denoised image;

[0093] S303. Determine the binarization threshold using the minimum error method, and perform binarization operation on the denoised image to obtain a binary image.

[0094] S304. Use Harris corner detection to identify and mark the center of the bolt head;

[0095] S305. Use edge detection technology (such as the improved Canney operator or active profile model) to accurately define the overall profile of the bolt, ensuring that the bolt boundary can still be accurately defined under the influence of corrosion.

[0096] S306. Compile and output the geometric dimensions of each bolt connection, including diameter, length, and degree of corrosion.

[0097] S4. Compare the real-time bolt connection with the finite element model. If the similarity is lower than the preset value, proceed to step S5. Otherwise, directly call the preload loss from the existing finite element database.

[0098] S401. Based on the geometric features of the target bolt connection obtained in step S3, compare them with the dimensional information recorded in the database and obtain the comparison similarity.

[0099] S402. Set the threshold for target comparison similarity, and judge the degree of similarity between the model and the real-time bolt connection based on the comparison similarity.

[0100] S403. If the value is below the threshold, proceed to step S5; otherwise, directly use the preload loss from the existing finite element database as the prediction result.

[0101] S5. Utilize the trained machine learning model to output the actual bolt connection preload loss in real time.

[0102] S501. Normalize and regularize the geometric features and related physical information of the target bolt and input them into the machine learning model to predict the preload. If the preload loss exceeds the set value, a signal is sent to notify the maintenance personnel to inspect the target bolt. Otherwise, the normal data is detected in real time and archived.

[0103] S502. For the obtained preload result, establish a finite element model based on the corresponding geometric features and physical information, and add a finite element database.

[0104] In another embodiment of the present invention, an intelligent self-optimizing prediction system for the preload of rusted high-strength bolt connections is provided. This system can be used to implement the above-mentioned intelligent self-optimizing prediction method for the preload of rusted high-strength bolt connections. Specifically, the intelligent self-optimizing prediction system for the preload of rusted high-strength bolt connections includes a data module, a judgment module, a monitoring module, a comparison module, and a prediction module.

[0105] Among them, the data module establishes a finite element database of bolt connection preload loss and trains a machine learning prediction model based on neural networks.

[0106] The judgment module determines whether the recognition accuracy of the machine learning prediction model has reached the preset value.

[0107] The monitoring module monitors the preload loss of the target bolt in real time and obtains real-time geometric feature information of the target bolt connection when the recognition accuracy of the machine learning prediction model reaches a preset value.

[0108] The comparison module compares the real-time geometric feature information of the target bolt connection with the similarity of the finite element model in the finite element database. When the similarity is greater than or equal to the preset value, it calls the preload loss from the existing finite element database.

[0109] The prediction module uses a trained machine learning prediction model to output the actual bolt connection preload loss in real time when the similarity is lower than the preset value.

[0110] In another embodiment of the present invention, a terminal device is provided, comprising a processor and a memory. The memory stores a computer program, which includes program instructions. The processor executes the program instructions stored in the computer storage medium. The processor may be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. It is the computing and control core of the terminal, suitable for implementing one or more instructions, specifically suitable for loading and executing one or more instructions to achieve a corresponding method flow or corresponding function. The processor described in this embodiment can be used in the operation of an intelligent self-optimization prediction method for the preload of rust-corroded high-strength bolt connections, including:

[0111] A finite element database of bolt preload loss is established, and a machine learning prediction model based on a neural network is trained. The accuracy of the machine learning prediction model's recognition reaches a preset value. When the machine learning prediction model's recognition accuracy reaches the preset value, the preload loss of the target bolt is monitored in real time to obtain the real-time geometric feature information of the target bolt connection. The similarity between the real-time geometric feature information of the target bolt connection and the finite element model in the finite element database is compared. When the similarity is greater than or equal to a preset value, the preload loss from the existing finite element database is retrieved. When the similarity is lower than a preset value, the actual bolt connection preload loss is output in real time using the trained machine learning prediction model.

[0112] In another embodiment of the present invention, a storage medium is also provided, specifically a computer-readable storage medium (memory). This computer-readable storage medium is a memory device in a terminal device used to store programs and data. It is understood that the computer-readable storage medium here can include both the built-in storage medium in the terminal device and extended storage media supported by the terminal device. The computer-readable storage medium provides storage space that stores the terminal's operating system. Furthermore, this storage space also stores one or more instructions suitable for loading and execution by a processor. These instructions can be one or more computer programs (including program code). It should be noted that the computer-readable storage medium here can be high-speed RAM or non-volatile memory, such as at least one disk storage device.

[0113] One or more instructions stored in a computer-readable storage medium can be loaded and executed by a processor to implement the corresponding steps of the intelligent self-optimization prediction method for preload of rusted high-strength bolt connections in the above embodiments; one or more instructions in the computer-readable storage medium are loaded and executed by the processor in the following steps:

[0114] A finite element database of bolt preload loss is established, and a machine learning prediction model based on a neural network is trained. The accuracy of the machine learning prediction model's recognition reaches a preset value. When the machine learning prediction model's recognition accuracy reaches the preset value, the preload loss of the target bolt is monitored in real time to obtain the real-time geometric feature information of the target bolt connection. The similarity between the real-time geometric feature information of the target bolt connection and the finite element model in the finite element database is compared. When the similarity is greater than or equal to a preset value, the preload loss from the existing finite element database is retrieved. When the similarity is lower than a preset value, the actual bolt connection preload loss is output in real time using the trained machine learning prediction model.

[0115] Please see Figure 2The terminal device is a computer device. In this embodiment, the computer device 60 includes a processor 61, a memory 62, and a computer program 63 stored in the memory 62 and executable on the processor 61. When executed by the processor 61, the computer program 63 implements the fluid composition calculation method in the reservoir stimulation wellbore of this embodiment. To avoid repetition, these details are not elaborated here. Alternatively, when executed by the processor 61, the computer program 63 implements the functions of each model / unit in the fluid composition calculation system in the reservoir stimulation wellbore of this embodiment. To avoid repetition, these details are not elaborated here.

[0116] Computer device 60 can be a desktop computer, laptop, handheld computer, cloud server, or other computing device. Computer device 60 may include, but is not limited to, a processor 61 and a memory 62. Those skilled in the art will understand that... Figure 2 This is merely an example of computer device 60 and does not constitute a limitation on computer device 60. It may include more or fewer components than shown, or combine certain components, or different components. For example, computer device may also include input / output devices, network access devices, buses, etc.

[0117] The processor 61 may be a central processing unit (CPU), or other general-purpose processors, CPUs, graphics processing units (GPUs), digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, quantum computing-based data processing logic units, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor.

[0118] The memory 62 can be an internal storage unit of the computer device 60, such as a hard disk or RAM of the computer device 60. The memory 62 can also be an external storage device of the computer device 60, such as a plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, etc. equipped on the computer device 60.

[0119] Furthermore, the memory 62 may include both internal storage units of the computer device 60 and external storage devices. The memory 62 is used to store computer programs and other programs and data required by the computer device. The memory 62 can also be used to temporarily store data that has been output or will be output.

[0120] Any references to memory, databases, or other media used in the embodiments provided in this application may include at least one of non-volatile and volatile memory. Non-volatile memory may include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory may include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM may be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc.

[0121] The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to these.

[0122] Please see Figure 3 The terminal device is a chip. In this embodiment, the chip 600 includes a processor 622, which may be one or more, and a memory 632 for storing computer programs executable by the processor 622. The computer program stored in the memory 632 may include one or more modules, each corresponding to a set of instructions. Furthermore, the processor 622 may be configured to execute the computer program to perform the generalizable monocular absolute depth map estimation method described above.

[0123] Additionally, chip 600 may also include a power supply component 626 and a communication component 650. The power supply component 626 can be configured to perform power management of chip 600, and the communication component 650 can be configured to enable communication of chip 600, such as wired or wireless communication. Furthermore, chip 600 may also include an input / output interface 658. Chip 600 can operate on an operating system stored in memory 632.

[0124] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, 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, not all, of the embodiments of the present invention. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0125] Example

[0126] The following is an example of the prediction process for the preload loss of a certain type of corroded bolt;

[0127] This embodiment provides an intelligent self-optimization prediction method for the preload force of high-strength bolt connections with corrosion, specifically including the following steps:

[0128] Based on the bolt connection types used in the current engineering structure, an ANSYS finite element model is established, and a preload loss simulation is performed to establish a bolt connection preload loss database. The existing database is then used to optimize the BP neural network for training using the particle swarm optimization algorithm.

[0129] After the database contains 1000 finite element models, it is determined whether the prediction accuracy of the trained neural network has reached the preset value. If the preset value is reached, the target bolt connection is monitored in real time. If it is not sufficient, the database is further optimized.

[0130] Using a CCD industrial camera to capture images of bolted connections during the day when the light is good at the target site, the images are transmitted to a remote server. Finally, computer vision techniques, including but not limited to Gaussian filtering for noise reduction, binarization, and Harris corner detection, are used to obtain the geometric feature information of the bolted connections in the target component.

[0131] By comparing the obtained geometric feature information with the dimensions of finite element models in the existing database, a target similarity threshold of 95% is set. The similarity between the model and the real-time bolted connection is determined based on the comparison similarity. If the similarity is greater than or equal to 95%, the preload loss in the database is directly used as the prediction result for the target bolted connection.

[0132] Meanwhile, for bolt connections below the threshold, the preload loss is predicted using a trained neural network model based on the real-time geometric features of the bolt connection; a finite element model is established based on the geometric features to simulate the preload loss, and finally the result is added to the existing database.

[0133] Please see Figure 4 The figure illustrates the process of optimizing a backpropagation (BP) neural network using the particle swarm optimization (PSO) algorithm. This process includes initializing parameters such as particle velocity, position, inertia factor, and number of iterations. Then, iteratively calculating the fitness value, individual extreme values, and swarm extreme values ​​of the particles ultimately finds the global optimum. During this process, the PSO algorithm helps the BP neural network perform a global search in the weight space, thus avoiding getting trapped in local optima. This means that PSO-BP neural networks are more likely to find a global optimum or a solution close to the global optimum during training, improving the model's generalization ability and prediction accuracy.

[0134] Adjusting the weights and thresholds of the BP neural network during the process can effectively improve the convergence speed and accuracy of the BP neural network, thereby improving the accuracy of preload prediction for rusted high-strength bolt connections.

[0135] For the overall data structure, the PSO optimization algorithm has a stronger learning ability. By simulating swarm intelligence, the PSO algorithm can guide the efficient optimization of network weights. This enables the BP neural network to better learn complex patterns in the input data during training, especially nonlinear mapping relationships, such as the relationship between corrosion features and preload loss.

[0136] pass Figure 4 As can be seen, the advantage of this invention lies in optimizing the training process of the BP neural network using an improved particle swarm optimization algorithm, significantly improving the model's training efficiency and prediction accuracy, especially when dealing with nonlinear problems. This optimization method is particularly suitable for applications requiring high-precision prediction, such as the intelligent prediction of preload in corroded high-strength bolt connections.

[0137] Please see Figure 5 The figure shows a flowchart of obtaining the dimensional features of the target bolt connection and predicting the preload in an embodiment of the present invention. The figure uses an industrial camera to take pictures and computer vision technology to obtain the dimensions of the target bolt connection in real time.

[0138] Then, the obtained dimensions are compared with those in the existing finite element database. If there is a model with a similarity greater than the preset value, the preload loss of that model is directly used as the prediction result.

[0139] Otherwise, the preload value is predicted using a trained PSO-BP neural network model, and the obtained feature information is used for finite element modeling to simulate preload loss. The result is then added to the database to achieve self-optimization.

[0140] In summary, this invention provides an intelligent self-optimizing prediction method and system for the preload of corroded high-strength bolt connections. By constructing an intelligent self-optimizing prediction system for the preload of corroded high-strength bolt connections based on finite element analysis and particle swarm optimization (PSO) algorithm-optimized BP neural network, it achieves fully automated management of the entire process from model simulation and real-time monitoring to intelligent prediction. Furthermore, the detection process does not cause any damage or destruction to existing steel structure nodes and bolt connections, resulting in low cost and high practicality. It also provides a theoretical basis for monitoring the performance of existing bolt connections. The embodiments demonstrate that by establishing a database with a precise model, optimizing the neural network to a predetermined accuracy, combining image processing technology with on-site data collection, and establishing a high-threshold comparison strategy for rapid classification, it ensures both immediate and accurate prediction of preload loss in similar bolt connections, effectively handles complex cases using the neural network model, and continuously iterates and upgrades the database by adding new simulation data, continuously improving the accuracy and generalization ability of the prediction system.

[0141] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is merely an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiments can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit. Furthermore, the specific names of the functional units and modules are only for easy differentiation and are not intended to limit the scope of protection of this application. The specific working process of the units and modules in the above system can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0142] In the above embodiments, the descriptions of each embodiment have different focuses. For parts that are not described in detail or recorded in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0143] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed in this invention can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0144] In the embodiments provided by this invention, it should be understood that the disclosed devices / terminals and methods can be implemented in other ways. For example, the device / terminal embodiments described above are merely illustrative. For instance, the division of modules or units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical, mechanical, or other forms.

[0145] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0146] Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0147] If the integrated module / unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the methods of the above embodiments can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include: any entity or device capable of carrying the computer program code, recording media, USB flash drives, portable hard drives, magnetic disks, optical disks, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signals, telecommunication signals, and software distribution media, etc. It should be noted that the content included in the computer-readable medium can be appropriately added or removed according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, computer-readable media do not include electrical carrier signals and telecommunication signals.

[0148] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0149] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0150] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0151] The above content is only for illustrating the technical concept of the present invention and should not be construed as limiting the scope of protection of the present invention. Any modifications made to the technical solution based on the technical concept proposed in this invention shall fall within the scope of protection of the claims of this invention.

Claims

1. A method for intelligent self-optimization and prediction of pre-tightening force of high-strength bolted connections against corrosion, characterized by, The method comprises the following steps: A finite element database of bolt connection preload loss is established, and a machine learning prediction model based on a neural network is trained, specifically as follows: A three-dimensional entity geometric model of a non-rust bolt connection is established; The three-dimensional entity geometric model of the non-rust bolt connection is divided into units and boundary conditions are applied to obtain a first analysis model; A first load is applied to the first analysis model, and the preload of the first analysis model is calculated; According to the size of the non-rust bolt connection and the actual bolt rusting principle, a component is randomly defined as a rusting unit; According to the actual bolt rusting principle, the relevant physical properties of the bolt connection are randomly set; The modified unit is killed by using the birth and death unit method to obtain a second analysis model; A second load is applied to the second analysis model, and the preload of the obtained second analysis model is calculated to obtain the preload loss value of the rusted bolt; The relevant size information, physical information and preload loss value of the rusted bolt are recorded; The above steps are repeated, and the data obtained each time is sorted to establish a finite element database of bolt connection preload loss and train a machine learning prediction model based on a neural network; It is judged whether the recognition accuracy of the obtained machine learning prediction model reaches a preset recognition accuracy value; When the recognition accuracy of the machine learning prediction model reaches the preset value, the preload loss of the target bolt is monitored in real time, and real-time geometric feature information of the target bolt connection is obtained; The similarity of the obtained real-time geometric feature information of the target bolt connection and the finite element model in the finite element database is compared, and when the similarity is greater than or equal to a similarity threshold value, the preload loss of the existing finite element database is called, specifically as follows: According to the real-time geometric feature information of the target bolt connection, the size information recorded in the database is compared, and the comparison similarity is obtained; the target comparison similarity threshold value is set, and the similarity of the model and the real-time bolt connection is judged according to the comparison similarity; if it is higher than the target comparison similarity threshold value, the preload loss of the existing finite element database is directly called as the prediction result; When the similarity is lower than the similarity threshold value, the trained machine learning prediction model is used to output the actual bolt connection preload loss in real time.

2. The method of claim 1, wherein the method is characterized by: The prediction model based on the neural network uses a particle swarm algorithm to optimize the BP neural network.

3. The method of claim 1, wherein the method is characterized by: When the recognition accuracy of the prediction model does not reach the preset value, the finite element database is increased, and the machine learning prediction model based on the neural network is returned to train.

4. The method of claim 1, wherein the method is characterized by: The real-time geometric feature information of the bolt connection is obtained as follows: The obtained picture of the bolt connection is subjected to picture gray processing to obtain a gray picture; The gray picture is subjected to noise reduction by a Gaussian filter to obtain a picture after noise reduction; A binary threshold value is determined by using a minimum error method, and a binary operation is performed on the denoised image to obtain a binary image; Harris corner detection is used to identify and mark the center of the bolt head; Edge detection technology is used to accurately determine the overall profile of the bolt to ensure that the bolt boundary can be accurately defined under the influence of rust; The diameter, length, rust degree and other geometric sizes of each bolt connection are sorted and output.

5. The method of claim 1, wherein the method is characterized by: The trained machine learning prediction model is used to output the actual bolt connection preload loss in real time as follows: The geometric feature information and related physical information of the target bolt are normalized and regularly input into a machine learning model for pre-tightening force prediction. If the pre-tightening force loss exceeds the set value, a signal is sent to the maintenance personnel to maintain the target bolt, otherwise the real-time detection normal data is backed up; For the obtained pre-tightening force result, a finite element model is established combining the corresponding geometric features and physical information to increase the finite element database.

6. A rust high-strength bolt connection pre-tightening force intelligent self-optimization prediction system, characterized in that, It comprises: A data module for establishing a finite element database of bolt connection pre-tightening force loss and training a machine learning prediction model based on neural network, specifically: Establish a three-dimensional solid geometric model of uncorroded bolt connection; Divide the uncorroded bolt connection three-dimensional solid geometric model into units and apply boundary conditions to obtain a first analysis model; Apply the first step load to the first analysis model to calculate the pre-tightening force of the first analysis model; Randomly define components as corrosion units according to the size of uncorroded bolt connection and the actual bolt corrosion principle; Randomly set the physical properties of bolt connection according to the actual bolt corrosion principle; Kill the modified units using the birth-death unit method to obtain a second analysis model; Apply the second step load to the second analysis model to calculate the pre-tightening force of the obtained second analysis model to calculate the pre-tightening force loss value of the corroded bolt; Record the size information, physical information and pre-tightening force loss value of the relevant corroded bolt; Repeat the above steps to organize the data obtained each time to establish a finite element database of bolt connection pre-tightening force loss and train a machine learning prediction model based on neural network; A judgment module for judging whether the recognition accuracy of the machine learning prediction model reaches the preset value; A monitoring module for monitoring the pre-tightening force loss of the target bolt in real time when the recognition accuracy of the machine learning prediction model reaches the recognition accuracy preset value to obtain the real-time geometric feature information of the target bolt connection; A comparison module for comparing the similarity of the real-time geometric feature information of the target bolt connection with the finite element model in the finite element database, and calling the pre-tightening force loss of the existing finite element database when the similarity is greater than or equal to the similarity threshold, specifically: Compare the size information recorded in the database according to the real-time geometric feature information of the target bolt connection and obtain the comparison similarity; set a target comparison similarity threshold, judge the similarity degree of the model and the real-time bolt connection according to the comparison similarity; if it is higher than the target comparison similarity threshold, directly call the pre-tightening force loss of the existing finite element database as the prediction result; A prediction module for outputting the actual bolt connection pre-tightening force loss in real time using the trained machine learning prediction model when the similarity is lower than the similarity threshold.

7. A computer-readable storage medium storing one or more programs, the one or more programs comprising instructions that when executed by a computer cause the computer to perform a method of any of claims 1-6. The one or more programs include instructions that, when executed by a computing device, cause the computing device to perform the method of any one of claims 1 to 5.

8. A computing device, comprising: It comprises: One or more processors, memories and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, the one or more programs comprising steps for performing the method of any one of claims 1 to 5.

Images