Optical fiber cable weaving process parameter optimization method, equipment and medium

Abstract

The invention discloses an optical fiber cable weaving process parameter optimization method and device and a medium, and relates to the technical field of optical fiber cable manufacturing, and the method comprises the steps: building an analytical method model of optical fiber cable mechanical properties; sampling in the process parameter space to generate training sample points, inputting the training sample points into the analytical method model to obtain a mechanical property response data set, and fitting a polynomial proxy model based on a least square method; constructing a multi-objective optimization function based on the polynomial agent model, and performing global optimization in the process parameter space based on the multi-objective optimization function to obtain an optimal process parameter combination; the optimal process parameter combination is put into actual production, actual mechanical properties are tested, a process-actual mechanical property data pair is formed, a database is updated, a least square method is adopted to introduce a proportion correction coefficient and an offset correction coefficient to dynamically calibrate an analytical method model, the steps are repeated, a self-learning closed loop is formed, and a self-learning result is obtained. And rapid performance prediction and accurate parameter optimization are realized.

Description

Technical Field

[0001] This application relates to the field of optical fiber and cable manufacturing technology, and in particular to a method, equipment and medium for optimizing optical fiber and cable braiding process parameters. Background Technology

[0002] Currently, the braiding process of optical fiber cables typically relies on experience for parameter setting, with common parameters including braiding angle and yarn thickness. In traditional optical fiber cable production, the design and adjustment of the braiding process depend on repeated experimental verification or simulation calculations. Experimental verification methods rely on experience and numerous experiments, resulting in long design and manufacturing cycles, high costs, and significant computational demands. Experience-based methods lack systematicity, while experimental verification requires repeated trial and error, making it difficult to quickly adapt to the braiding requirements of different optical fiber cables. Furthermore, simulation methods such as the finite element method (FEM) can accurately simulate the mechanical properties of optical cables, but their simulations require detailed modeling and substantial computational resources, making them complex and time-consuming, and difficult to meet the needs of rapid response and efficient optimization. Therefore, there is an urgent need for a method, equipment, and medium for optimizing optical fiber cable braiding process parameters to achieve rapid performance prediction and accurate parameter optimization, thereby improving efficiency, reducing costs, and enhancing product performance, and meeting the mechanical performance design requirements of different types of optical fiber cables. Summary of the Invention

[0003] The purpose of this application is to provide a method, equipment, and medium for optimizing the process parameters of optical fiber and cable braiding, which can achieve rapid performance prediction and accurate parameter optimization, thereby improving efficiency, reducing costs, and enhancing product performance.

[0004] To achieve the above objectives, this application provides the following solution: Firstly, this application provides a method for optimizing the braiding process parameters of optical fibers and cables, including: S1: Establish an analytical model of the mechanical properties of optical fiber cables; the analytical model includes a two-stage tensile deformation model, an equivalent thin-walled circular tube bending stiffness model, and an equivalent thin-walled circular tube torsional stiffness model; S2: Sample points are generated in the process parameter space. The training sample points are input into the analytical model to obtain the mechanical performance response dataset. A polynomial surrogate model is fitted based on the least squares method. Each training sample point is a combination of process parameters. The polynomial surrogate model includes a tensile strength surrogate model, a bending stiffness surrogate model, and a torsional stiffness surrogate model. S3: Construct a multi-objective optimization function based on the polynomial surrogate model, and perform global optimization in the process parameter space based on the multi-objective optimization function to obtain the optimal combination of process parameters; S4: Put the optimal combination of process parameters into actual production, test the actual mechanical properties of the finished optical cable, form a process-actual mechanical properties data pair and update the database, use the least squares method to introduce proportional correction coefficients and offset correction coefficients to dynamically calibrate the analytical model, replace the analytical model with the corrected analytical model and repeat steps S2-S4 to form a self-learning closed loop.

[0005] Secondly, this application provides a computer device, including: a memory, a processor, and a computer program stored in the memory and capable of running on the processor, wherein the processor executes the computer program to implement the above-described method for optimizing optical fiber and cable braiding process parameters.

[0006] Thirdly, this application provides a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the above-mentioned method for optimizing optical fiber and cable braiding process parameters.

[0007] According to the specific embodiments provided in this application, the following technical effects are disclosed: This application provides a method, equipment, and medium for optimizing optical fiber cable braiding process parameters. By establishing an analytical model to predict the mechanical properties of optical fiber cables, and combining it with a polynomial surrogate model to quickly fit the mechanical properties and optimize process parameters, this achieves rapid performance prediction and accurate process parameter optimization in the optical fiber cable braiding process, significantly improving braiding efficiency, reducing costs, and enhancing product consistency and reliability. By feeding back actual production data and dynamically updating the analytical model and optimization algorithm, a complete closed loop is formed from virtual prediction to actual production, and then self-improvement through feedback, improving the accuracy of optical fiber cable mechanical property prediction and process adaptability. Attached Figure Description

[0008] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0009] Figure 1 This is an application environment diagram of a method for optimizing optical fiber and cable braiding process parameters according to an embodiment of this application.

[0010] Figure 2 This is a flowchart illustrating a method for optimizing optical fiber and cable braiding process parameters according to an embodiment of this application.

[0011] Figure 3 This is a schematic diagram illustrating the specific process of a method for optimizing optical fiber and cable braiding process parameters according to an embodiment of this application.

[0012] Figure 4 This is a schematic diagram of the structure of a computer device provided in an embodiment of this application. Detailed Implementation

[0013] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0014] To make the above-mentioned objectives, features and advantages of this application more apparent and understandable, the application will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0015] The fiber optic cable braiding process parameter optimization method provided in this application embodiment can be applied to, for example... Figure 1 In the application environment shown, terminal 102 communicates with server 104 via a network. A data storage system can store the data that server 104 needs to process. The data storage system can be set up independently, integrated into server 104, or placed in the cloud or on another server. Terminal 102 can send requests to be processed to server 104. After receiving the requests, server 104 establishes an analytical model of the mechanical properties of optical fiber cables, samples and generates training sample points in the process parameter space, inputs the training sample points into the analytical model to obtain a mechanical performance response dataset, fits a polynomial surrogate model based on the least squares method, constructs a multi-objective optimization function based on the polynomial surrogate model, performs global optimization in the process parameter space based on the multi-objective optimization function to obtain the optimal process parameter combination, puts the optimal process parameter combination into actual production, tests the actual mechanical properties, forms a process-actual mechanical performance data pair and updates the database, uses the least squares method to introduce proportional correction coefficients and offset correction coefficients to dynamically calibrate the analytical model, and repeats the above steps to form a self-learning closed loop. Server 104 can feed back the obtained optimal process parameter combination to terminal 102. In addition, in some embodiments, the method for optimizing the process parameters of optical fiber cable braiding can also be implemented by the server 104 or the terminal 102 separately. For example, the terminal 102 can directly optimize the process parameters for the request to be processed, or the server 104 can obtain the request to be processed from the data storage system and optimize the process parameters for the request to be processed.

[0016] The terminal 102 can be, but is not limited to, various desktop computers and laptops. The server 104 can be implemented using a standalone server or a server cluster consisting of multiple servers, or it can be a cloud server.

[0017] In one exemplary embodiment, such as Figure 2 As shown, a method for optimizing optical fiber cable braiding process parameters is provided. This method is executed by computer equipment, specifically by a terminal or server alone, or by both a terminal and a server. In this embodiment, the method is applied to... Figure 1 Taking server 104 as an example, the explanation includes the following steps S1 to S4.

[0018] S1: Establish an analytical model of the mechanical properties of optical fiber cables; the analytical model includes a two-stage tensile deformation model, an equivalent thin-walled circular tube bending stiffness model, and an equivalent thin-walled circular tube torsional stiffness model.

[0019] S2: Sample points are generated in the process parameter space. The training sample points are input into the analytical model to obtain the mechanical performance response dataset. A polynomial surrogate model is fitted based on the least squares method. Each training sample point is a combination of process parameters. The polynomial surrogate model includes a tensile strength surrogate model, a bending stiffness surrogate model, and a torsional stiffness surrogate model.

[0020] S3: Construct a multi-objective optimization function based on the polynomial surrogate model, and perform global optimization in the process parameter space based on the multi-objective optimization function to obtain the optimal combination of process parameters.

[0021] S4: Put the optimal combination of process parameters into actual production, test the actual mechanical properties of the finished optical cable, form a process-actual mechanical properties data pair and update the database, use the least squares method to introduce proportional correction coefficients and offset correction coefficients to dynamically calibrate the analytical model, replace the analytical model with the corrected analytical model and repeat steps S2-S4 to form a self-learning closed loop.

[0022] By implementing steps S1 to S4 above, a fiber optic cable braiding process parameter optimization system integrating analytical modeling, data-driven approaches, and dynamic optimization was established. The entire scheme first achieves rapid and reliable prediction of the mechanical properties of fiber optic cables by establishing an analytical model. Then, the data generated by this analytical model is used to train a machine learning multinomial surrogate model, achieving rapid performance prediction and accurate parameter optimization. This improves efficiency, reduces costs, and enhances product performance, meeting the mechanical performance design requirements of different types of fiber optic cables. Finally, by combining optimization algorithms and a self-learning mechanism, accurate, automatic, and continuous optimization of process parameters is achieved, forming a complete closed loop from virtual prediction to actual production, and then self-improvement through feedback.

[0023] 1. Analytical Model: This refers to a model that directly calculates the mechanical properties of optical fibers and cables using mathematical analytical formulas. This avoids complex finite element simulations and experimental testing to obtain performance data, saving costs and improving computational efficiency. Performance prediction models can be established based on other mechanical theories (such as the energy method), but these may have higher computational complexity.

[0024] 2. Surrogate Models: A "simplified and efficient computational model" that replaces complex, time-consuming, or difficult-to-solve original models for rapid performance prediction, replacing costly experiments or simulations. Common types include response surfaces, Kriging, and radial basis functions (RBR).

[0025] 3. Braiding process parameters: including braiding angle, yarn width, etc. These parameters directly affect the mechanical properties of optical fiber cables after braiding and curing.

[0026] 4. Mechanical properties: including tensile strength, bending stiffness and torsional stiffness, are key indicators for evaluating the quality of optical fiber cables.

[0027] 5. Two-stage tensile deformation model: used to describe the initial structural deformation and subsequent co-load-bearing behavior of high-coverage braided structures during the tensile process.

[0028] 6. Equivalent thin-walled circular tube model: The braided layer is equivalent to a thin-walled circular tube, which simplifies the calculation of bending and torsional stiffness.

[0029] 7. Self-learning mechanism: By feeding back actual production data, the model and optimization algorithm are dynamically updated to improve prediction accuracy and process adaptability.

[0030] Step S1 is used for establishing and experimentally calibrating the rapid analytical model. This application first establishes an analytical model of the mechanical properties of optical cables to quickly calculate the tensile strength, bending stiffness, and torsional stiffness of the optical fiber cable. For example... Figure 3 As shown. Taking the process parameters including weaving angle and yarn width as an example, the method provided in this application will be explained. However, the process parameters may also include the total number of weaving spindles, which is not limited here.

[0031] In step S1 above, the two-stage tensile deformation model includes the following steps S101~S103.

[0032] S101: Determine the coverage of fiber optic cable braided structures based on the line coverage formula. The line coverage formula is used to determine the current cable coverage. The line coverage formula is as follows: (1); in, For the coverage of the woven structure; The total number of spinning spindles; This refers to the width of a single strand of yarn; Pi; The median diameter of the braided layer; It is a cosine function; This is the initial weaving angle.

[0033] S102: When the coverage is greater than 100%, the tensile strength is calculated using a two-stage tensile deformation model, including calculating the new weaving angle using the weaving angle calculation formula, and calculating the first-stage tensile load and the second-stage tensile load using the first-stage load calculation formula and the second-stage load calculation formula respectively; and calculating the tensile strength based on the first-stage tensile load and the second-stage tensile load.

[0034] S103: When the coverage is less than or equal to 100%, the tensile strength is calculated using the second stage load calculation formula.

[0035] If the coverage exceeds 100%, the woven structure will deform under tensile load. The new weave angle when the woven layer structure deforms to a blocked state is calculated using the weave angle calculation formula. The weave angle calculation formula is as follows: (2); in For the new weaving angle.

[0036] The load borne by the optical cable when the braided layer structure deforms to a blocked state is calculated using the first-stage load calculation formula. The first-stage load calculation formula is as follows: (3); in, This is the first stage of tensile load; The tensile modulus of the optical fiber; The outer diameter of the optical fiber; the strain of the optical fiber during structural deformation. for: (4).

[0037] The parameters such as the tensile load in the first stage, the new braiding angle, the initial braiding angle, the tensile modulus of the optical fiber, the strain of the optical fiber during the structural deformation stage, and the outer diameter of the optical fiber were obtained through measurement.

[0038] If the coverage is less than or equal to 100%, the woven structure will not deform under tensile load. (5); Once the braided structure of the optical cable reaches a blocked state, structural deformation will cease. The optical fiber and the braided structure will share the tensile load until deformation occurs. The tensile force on the optical cable under the combined load is calculated using the second-stage load calculation formula, as follows: (6); (7); in, This is the second stage of tensile load; This refers to the tensile modulus of the fiber. This refers to the total cross-sectional area of ​​the woven structure in the axial direction. This represents the axial strain in the second stage. The cross-sectional area of ​​the optical fiber; Elongation at break of the fiber; This represents the breaking elongation of the optical fiber. Parameters such as the fiber's tensile modulus, the total cross-sectional area of ​​the braided structure in the axial direction, the axial strain in the second stage, and the optical fiber's cross-sectional area are obtained through measurement.

[0039] The tensile strength of the optical cable is calculated using the total load calculation formula, which is as follows: (8); in, Tensile strength.

[0040] When an optical cable is subjected to bending loads, for ease of analysis, the complex braided structure is equivalent to a homogeneous thin-walled circular tube model to simplify the geometry and stress relationship. The bending stiffness of the braided layer is calculated using the formula for calculating the bending stiffness of a circular tube. The optical fiber in the middle is considered a uniform circular rod. The formula for estimating the bending stiffness of the optical cable, i.e., the equivalent thin-walled circular tube bending stiffness model, is as follows: (9); in, The bending stiffness of the optical cable; Pi; This refers to the tensile modulus of the fiber. The thickness of the fiber; The median diameter of the braided layer; It is the cotangent function; For a new weaving angle; The total number of spinning spindles; This refers to the width of a single yarn strand. It is a cosine function; This is the initial weaving angle; The tensile modulus of the optical fiber; Where is the fiber radius.

[0041] When an optical cable is subjected to torsional loads, for ease of analysis, the complex braided structure is equivalent to a homogeneous thin-walled circular tube model to simplify the geometry and stress relationships. The braided structure is calculated using an equivalent shear model, with the central optical fiber considered as a uniform circular rod. The torsional stiffness calculation model of a circular tube is used to calculate the equivalent shear modulus of the braided structure. The formula for calculating the torsional stiffness of the optical cable, i.e., the equivalent thin-walled circular tube torsional stiffness model, is as follows: (10); in, This refers to the torsional stiffness of the optical cable. Pi; This is the shear modulus of the fiber; The thickness of the fiber; The median diameter of the braided layer; It is the cotangent function; For a new weaving angle; The total number of spinning spindles; This refers to the width of a single yarn strand. It is a cosine function; This is the initial weaving angle; It is the fiber shear modulus; Where is the fiber radius.

[0042] To overcome the computational bottleneck of analytical models in massive iterative optimization, this application constructs a surrogate model based on a multinomial response surface to achieve millisecond-level prediction performance. The key to its construction lies in the generation of high-quality training samples and efficient model fitting.

[0043] In step S2 above, sampling is performed in the process parameter space to generate training sample points. Specifically, this includes: for each process parameter, the range of each process parameter is divided into several equal probability intervals using the hypercubic Latin sampling method; a value of a process parameter is randomly selected in each equal probability interval, and the values ​​of each selected process parameter are randomly combined to form several process parameter combinations, and each process parameter combination is a training sample point.

[0044] First, the hypercubic Latin sampling experimental design method is adopted to perform intelligent sampling within the process parameter space (such as weaving angle, yarn width, etc.).

[0045] This sampling method assumes there are m process parameters (weaving angle and yarn width) and plans to sample n points. LHS divides the range of each parameter into n equally probable intervals. Then, a value is randomly selected from each interval of each parameter, and these values ​​are randomly combined across all dimensions to ensure that the training sample points are uniformly distributed across each parameter dimension; and that there is no obvious clustering of training sample points in the m-dimensional space.

[0046] For an m-dimensional parameter space, an LHS design containing n samples can be represented as an n The matrix X of m is as follows: (11); in, x is the element in the nth row and mth column of matrix X; i This represents the value of the j-th process parameter at the i-th training sample point. Using the LHS strategy, each column of matrix X is a uniform sample over its parameter domain, forming a combination of process parameters.

[0047] After obtaining high-quality training sample points, the response is calculated using a high-fidelity analytical model, and a multinomial surrogate model is fitted. First, for each m-dimensional training sample point in the LHS sample matrix X... Run the fast analytical model to obtain the corresponding mechanical performance response values. To form a complete dataset .

[0048] Establish an independent polynomial surrogate model for each mechanical performance index. For indices containing multiple process parameters (such as braiding angle)... Yarn width The general formula for the higher-order model (etc.), the polynomial surrogate model, is as follows: (12); in, These are the predicted mechanical properties calculated using a polynomial surrogate model. It is the coefficient of the constant term; These are the coefficients of the linear terms of each parameter; For the first One process parameter; These are the coefficients of the second-order interaction terms; For the first One process parameter; These are the coefficients of the third-order interaction term; For the first One process parameter; It is the model error term.

[0049] To determine the polynomial coefficients, the least squares method is used, and the optimal parameters are solved by minimizing the sum of squared residuals. After the model form is determined, it is necessary to learn from the data to determine all the coefficients in the polynomial. (in p (This refers to the total number of coefficients), and this process is called model fitting.

[0050] The least squares method is used for fitting. Its core objective is to find a set of coefficients that make the multinomial surrogate model's predictions accurate for all training sample points. Compared with the true value The sum of the differences is minimized. This difference is measured by the sum of squared residuals: (13); in, The sum of squared residuals between the predicted and actual mechanical properties; the true value. These are the actual mechanical property values ​​from the experiment; Represents training sample points The corresponding polynomial basis function terms; For the first The value of the first process parameter in the training sample points.

[0051] By constructing a design matrix from the equations of the residual sum of squares, the problem is transformed into a matrix solution, allowing for the rapid and direct calculation of the optimal coefficient vector. : (14); Where X is the extended design matrix (containing constant, linear, squared, and cross terms), and Y is a vector consisting of the true response values ​​of all samples; superscript... Indicates transpose. Once calculated... The polynomial proxy model is then fully determined.

[0052] The following section presents a multi-objective optimization based on weighted combination and equipotential surface analysis: This involves obtaining multiple performance indicators (tensile strength, etc.). Bending stiffness Torsional stiffness After applying the polynomial proxy model, multi-objective collaborative optimization is performed. A comprehensive multi-objective optimization function is constructed: the multi-objective problem is transformed into a single-objective problem using a linear weighted sum method. Each objective is assigned a weight coefficient to reflect its relative importance. The multi-objective optimization function is expressed as follows: (15); in, The overall performance index aims to maximize... ; Let be the weighting coefficient, satisfying and For example, if tensile strength is the most important factor, then it can be set as follows: ; For the pre-built polynomial proxy model corresponding to the performance, For tensile strength surrogate model; A proxy model for bending stiffness; For torsional stiffness proxy model; These are the training sample points.

[0053] It should be noted that alternatives to the multinomial surrogate model, besides MLP, GPR, and SVR, can also be achieved using machine learning methods such as random forests and neural networks, but a balance between accuracy and computational resources is required.

[0054] In step S3 above, global optimization is performed in the process parameter space based on the multi-objective optimization function to obtain the optimal combination of process parameters. Specifically, this includes: using the multi-objective optimization function as the fitness function, using a genetic algorithm to perform global optimization in the process parameter space, obtaining the Pareto front non-dominated solution set, and screening the optimal combination of process parameters.

[0055] By performing multi-objective comprehensive optimization on the constructed multi-objective optimization function, a genetic algorithm is employed to perform global optimization within the process parameter space. The genetic algorithm uses the multi-objective optimization function as its fitness function and iteratively updates the population through an evolutionary process of "selection-crossover-mutation," gradually bringing individuals closer to the optimal solution. After multiple generations of evolution, the algorithm ultimately obtains the optimal combination of process parameters. Under these parameter conditions, key performance indicators such as tensile strength, bending stiffness, and torsional stiffness all reach their optimal comprehensive performance.

[0056] Through iterative evolution involving selection, crossover, and mutation operations, and by running the genetic algorithm multiple times under different weighting coefficient combinations, a series of non-dominated solutions (i.e., optimal compromise solutions where none of the performance indicators are superior or inferior to the others) were obtained. These solutions formed a Pareto front in the performance space. The Pareto front intuitively reflects the trade-offs between key performance parameters such as tensile strength, bending stiffness, and torsional stiffness, and can provide a decision-making basis for selecting process parameters under different design preferences. Finally, through comprehensive performance weight analysis and Pareto front screening, the optimal combination of process parameters was obtained. Under these conditions, the overall performance indicators reach an optimal balance.

[0057] It should be noted that particle swarm optimization, simulated annealing, and other algorithms can also be used to replace genetic algorithms for adjusting model parameters.

[0058] (1) High-quality sample generation: First, within the design space of key process parameters such as weaving angle and yarn width, intelligent sampling is performed using the hypercubic Latin sampling (LHS) method. This method divides the range of each parameter into equally probable intervals and randomly selects combinations to ensure that the samples are evenly distributed in the multidimensional space and avoid clustering, thereby obtaining a training sample set with better coverage. These training sample points together constitute a parameter-performance correspondence matrix, which is used for subsequent model training.

[0059] (2) Construction and fitting of polynomial surrogate model: After obtaining the samples, they are input into the high-fidelity analytical model to calculate the corresponding performance response data. For each mechanical performance index (such as tensile strength, bending stiffness, torsional stiffness, etc.), a polynomial response surface model is established to describe the nonlinear relationship between process parameters and performance.

[0060] This model includes linear terms, squared terms, and interaction terms among the parameters. By minimizing the sum of squared errors between the predicted and actual values, it automatically determines the optimal values ​​for each coefficient in the model. In other words, the model learns from sample data to extract the patterns of how parameter changes affect performance, achieving rapid performance prediction. Once the model is trained, inputting any set of process parameters will yield the corresponding performance prediction results within milliseconds.

[0061] (3) Multi-objective collaborative optimization and weight analysis: After establishing the polynomial surrogate model for each performance index, this application further conducts multi-objective collaborative optimization. By introducing a comprehensive performance evaluation function, multiple performance objectives (such as tensile strength, bending stiffness, and torsional stiffness) are combined according to different weights to reflect the comprehensive performance under different design focuses. For example, when tensile performance is the most critical, it can be given a higher weight.

[0062] Based on this comprehensive performance index, a genetic algorithm is employed to conduct a global search and iterative evolution across the entire process parameter space. Through operations such as selection, crossover, and mutation, the algorithm gradually brings the parameter combinations closer to the optimal state. After multiple rounds of evolution, a series of schemes achieving an optimal balance among different performance characteristics are obtained. These schemes form a Pareto front in the performance space, intuitively reflecting the trade-offs between different performance indicators.

[0063] Ultimately, through comprehensive performance weight analysis and Pareto front screening, the optimal combination of process parameters was obtained, enabling the optical cable to achieve overall optimal performance in key aspects such as tension, bending, and torsion.

[0064] The specific implementation process of the self-learning mechanism is as follows: The optimized combination of process parameters is then applied to actual production. Samples of the produced optical cables are taken and subjected to standard mechanical testing to precisely measure their actual structural parameters (such as final braiding angle and yarn density) and corresponding mechanical properties (tensile strength, bending stiffness, and torsional stiffness), forming new "process parameter-structural parameter-mechanical property" data pairs. These data pairs are then added to the system's core database. This database thus becomes a continuously growing, dynamic knowledge base containing both historical and the latest experimental data.

[0065] To improve the prediction accuracy of analytical models for the mechanical properties of optical cables under actual working conditions, a method for correcting and optimizing analytical models based on experimental data feedback was established. This method combines theoretical calculations with measured data to achieve joint calibration of analytical models for tension, bending, and torsion. The specific operation steps are as follows: 1. Data Preprocessing and Model Comparison: The collected data is processed, outliers are removed, and the average response curves under each working condition are calculated. The experimentally measured mechanical response is compared with the results calculated by the analytical model to obtain the deviation relationship between the theoretical prediction value (the mechanical performance prediction value calculated by the analytical model) and the measured value (i.e., the actual mechanical performance value in the experiment).

[0066] Experimental data acquisition method: Training data is generated through virtual simulation to replace some physical experiments, but the accuracy of the simulation must be ensured.

[0067] A model correction equation was established based on the least squares method, and regression fitting was performed on the main equivalent parameters (such as braided layer modulus, equivalent thickness, shear modulus, etc.) in the analytical model. A proportional correction coefficient was introduced during the model correction process. With offset correction This is to simultaneously correct for the overall stiffness difference of the model and the initial load offset.

[0068] Assume there is The test data for optical cables (tensile strength, bending stiffness, torsional stiffness) are provided for each set of data, along with the actual experimental mechanical property values. and predicted mechanical properties The objective of the least squares method is to minimize the sum of squared errors: (16); In the formula, Let n be the sum of squared errors of the n sets of data. This is a proportional correction parameter. For offset correction parameters, by adjusting... and By taking the partial derivative and setting it to zero, we can obtain an analytical solution.

[0069] In step S4 above, the analytical model is dynamically calibrated using the least squares method by introducing proportional correction coefficients and offset correction coefficients. Specifically, the standard formula for the least squares solution is as follows: (17); (18); in, This is the proportional correction factor; For the first Predicted mechanical properties of each training sample point; For the first The actual mechanical performance values ​​of each training sample point; , These represent the average values ​​of the predicted mechanical properties and the actual experimental mechanical properties, respectively. For offset correction parameters; A general correction equation is established based on the least squares method to calibrate the analytical model. This general correction equation is then used for dynamic calibration of the analytical model. .

[0070] Using the same principle, the analytical models of the tensile strength, bending stiffness, and torsional stiffness of optical cables were corrected three times using the same method.

[0071] The revised model is as follows: The revised two-stage tensile deformation model is as follows: (19); In the formula, This is a correction factor for the tensile strength formula. This is a parameter for offset correction in the tensile strength formula.

[0072] The corrected equivalent thin-walled circular tube bending stiffness model is as follows: (20); In the formula, This is a correction factor for the bending stiffness formula. This is the offset correction parameter for the bending stiffness formula.

[0073] The modified equivalent thin-walled circular tube torsional stiffness model is as follows: (twenty one); In the formula, This is a correction factor for the torsional stiffness formula. The offset correction parameter is used for the torsional stiffness formula.

[0074] The analytical model is used for parameter optimization and mechanical performance prediction in subsequent optical cable design stages. The specific process is as follows: The tensile, bending, and torsional properties of the optical cable are recalculated using the modified analytical model within the system, generating new performance prediction results, which are then used as input for the next round of process optimization. This process forms a closed-loop feedback loop with actual production experiments, enabling the optimization algorithm to continuously iterate and update with the assistance of a polynomial surrogate model. This gradually improves the prediction accuracy and the matching degree with process parameters, achieving the rapid, high-precision, and dynamically adaptive computational capabilities of the analytical model, thereby constructing a sustainable, self-learning optical cable design and optimization system.

[0075] This application includes the following technical points: 1. Establishment of analytical performance prediction model: including two-stage tensile deformation model, equivalent thin-walled circular tube model (for bending and torsional stiffness calculation), and fast performance prediction method based on analytical formula.

[0076] 2. Experimental optimization of analytical model: Adjusting model parameters through experimental data to improve prediction accuracy, involving the application of optimization algorithms (such as least squares method and genetic algorithm).

[0077] 3. Construction of polynomial surrogate model and performance prediction: Using a polynomial surrogate model and a fast prediction framework of hypercubic Latin sampling, a fast mapping from process parameters to mechanical properties is achieved.

[0078] 4. Dynamic optimization and self-learning mechanism: By feeding back actual production data, the model and optimization algorithm are dynamically updated to achieve continuous optimization of process parameters.

[0079] 5. Overall method integration: Combining analytical models, polynomial surrogate models, and self-learning mechanisms to form a complete process parameter optimization and control system.

[0080] This application has comprehensive advantages in terms of efficiency, accuracy, and practicality, and has the following benefits: 1) To address the issues of existing technologies relying on experience and incurring high experimental costs: This application replaces a large number of experiments with analytical models and polynomial surrogate models, achieving rapid prediction and significantly reducing design cycle and cost.

[0081] 2) To address the issues of complex and time-consuming calculations in simulation methods: the analytical model in this application has high computational efficiency, and the polynomial surrogate model responds quickly, overcoming the bottlenecks of simulation methods such as the finite element method.

[0082] 3) Addressing the lack of rapid optimization methods: This application combines dynamic optimization and self-learning mechanisms to achieve precise adjustment of process parameters, thereby improving product consistency and reliability.

[0083] 4) Therefore, this application achieves efficient and precise optimization of the optical fiber and cable braiding process, improving production efficiency and product quality.

[0084] This application also provides an application scenario in which the above-mentioned method for optimizing optical fiber and cable braiding process parameters is applied. Specifically, the method for optimizing optical fiber and cable braiding process parameters provided in this embodiment can be applied in a braiding process parameter optimization scenario. The braiding process parameter optimization scenario includes a request issuance stage and a braiding process parameter optimization stage; the request to be processed enters the braiding process parameter optimization stage from the request issuance stage, and obtains the corresponding content features (including the optimal combination of process parameters) through human-machine collaboration. The method for optimizing optical fiber and cable braiding process parameters provided in this embodiment belongs to the braiding process parameter optimization stage. Specifically, in the optimization of braiding process parameters for the requested process, an analytical model of the mechanical properties of optical fiber cables can be established. Training sample points are generated by sampling within the process parameter space. These training sample points are then input into the analytical model to obtain a mechanical performance response dataset. A polynomial surrogate model is fitted using the least squares method. A multi-objective optimization function is constructed based on the polynomial surrogate model. Global optimization is then performed within the process parameter space based on the multi-objective optimization function to obtain the optimal combination of process parameters. This optimal combination of process parameters is then put into actual production to test the actual mechanical properties, forming a process-actual mechanical performance data pair and updating the database. The analytical model is then dynamically calibrated using the least squares method by introducing proportional correction coefficients and offset correction coefficients. The above steps are repeated to form a self-learning closed loop.

[0085] In one exemplary embodiment, a computer device is provided, which may be a server or a terminal, and its internal structure diagram may be as follows. Figure 4 As shown, this computer device includes a processor, memory, input / output (I / O) interfaces, and a communication interface. The processor, memory, and I / O interfaces are connected via a system bus, and the communication interface is also connected to the system bus via the I / O interfaces. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a database. The internal memory provides the environment for the operating system and computer programs stored in the non-volatile storage media. The database stores optimization data for fiber optic cable braiding process parameters. The I / O interfaces are used for information exchange between the processor and external devices. The communication interface is used for communication with external terminals via a network connection. When the computer program is executed by the processor, it implements a method for optimizing fiber optic cable braiding process parameters.

[0086] Those skilled in the art will understand that Figure 4The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0087] In one exemplary embodiment, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above-described method embodiments.

[0088] In one exemplary embodiment, a computer-readable storage medium is provided storing a computer program that, when executed by a processor, implements the steps in the above-described method embodiments.

[0089] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data must comply with relevant regulations.

[0090] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can 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 can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM).

[0091] 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.

[0092] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0093] This document uses specific examples to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the methods and core ideas of this application. Furthermore, those skilled in the art will recognize that, based on the ideas of this application, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. An optical fiber cable weaving process parameter optimization method, characterized by, The optical fiber cable weaving process parameter optimization method comprises: S1: an analytical model of mechanical properties of the optical fiber cable is established; the analytical model comprises a two-stage tensile deformation model, an equivalent thin-walled circular tube bending stiffness model and an equivalent thin-walled circular tube torsional stiffness model; S2: training sample points are generated by sampling in a process parameter space, the training sample points are input into the analytical model to obtain a mechanical property response data set, and a polynomial surrogate model is fitted based on a least square method; each training sample point is a process parameter combination; the polynomial surrogate model comprises a tensile strength surrogate model, a bending stiffness surrogate model and a torsional stiffness surrogate model; S3: a multi-objective optimization function is constructed based on the polynomial surrogate model, and global optimization is performed in the process parameter space based on the multi-objective optimization function to obtain an optimal process parameter combination; S4: the optimal process parameter combination is put into actual production, actual mechanical properties of the cable product are tested, a process-actual mechanical property data pair is formed, and a database is updated, a proportional correction coefficient and an offset correction coefficient are introduced into the analytical model by using a least square method to dynamically calibrate the analytical model, the analytical model is replaced by the analytical model after correction, and steps S2-S4 are repeated to form a self-learning closed loop.

2. The optical fiber cable weaving process parameter optimization method of claim 1, wherein, The two-stage tensile deformation model comprises: The coverage of the optical fiber cable weaving structure is determined based on a line coverage formula; When the coverage is greater than 100%, the tensile strength is calculated using the two-stage tensile deformation model, including calculating a new weaving angle by a weaving angle calculation formula, and calculating first-stage tensile load and second-stage tensile load by first-stage load calculation formula and second-stage load calculation formula respectively; the tensile strength is calculated according to the first-stage tensile load and the second-stage tensile load; When the coverage is less than or equal to 100%, the tensile strength is calculated using the second-stage load calculation formula.

3. The optical fiber cable weaving process parameter optimization method of claim 1, wherein, The equivalent thin-walled circular tube bending stiffness model is as follows: ； wherein, is the bending stiffness of the optical cable; is the ratio of the circumference of a circle to its diameter; is the tensile modulus of the fiber; is the thickness of the fiber; is the diameter in the braid layer; is the cotangent function; is the new braid angle; is the total number of braid spindles; is the single yarn width, is the cosine function; is the initial braid angle; is the tensile modulus of the optical fiber; is the radius of the optical fiber.

4. The optical fiber cable weaving process parameter optimization method of claim 1, wherein, The equivalent thin-walled circular tube torsional stiffness model is as follows: ； wherein, is the torsional stiffness of the optical cable; is pi; is the shear modulus of the fiber; is the thickness of the fiber; is the diameter in the weave layer; is the cotangent function; is the new weave angle; is the total number of weave spindles; is the single yarn width, is the cosine function; is the initial weave angle; is the fiber shear modulus; is the fiber radius.

5. The optical fiber cable weaving process parameter optimization method of claim 1, wherein, The training sample points are generated by sampling in the process parameter space, specifically comprising: For each process parameter, the range of each process parameter is divided into several equal probability intervals by using a hypercube Latin sampling method; A value of each process parameter is randomly extracted in each equal probability interval, and each extracted value of the process parameter is randomly combined to form several process parameter combinations, each process parameter combination being a training sample point.

6. The optical fiber cable weaving process parameter optimization method of claim 1, wherein, The multi-objective optimization function is expressed as follows: ； wherein, is a comprehensive performance index, the goal is to maximize ; is a weight coefficient; is a tensile strength proxy model; is a bending stiffness proxy model; is a torsional stiffness proxy model; is a training sample point.

7. The optical fiber cable weaving process parameter optimization method of claim 1, wherein, Global optimization is performed in the process parameter space based on the multi-objective optimization function to obtain an optimal process parameter combination, specifically comprising: A genetic algorithm is used to perform global optimization in the process parameter space by taking the multi-objective optimization function as a fitness function, a Pareto front non-dominated solution set is obtained, and an optimal process parameter combination is screened.

8. The optical fiber cable weaving process parameter optimization method of claim 1, wherein, The analytical model is dynamically calibrated by introducing a proportional correction coefficient and an offset correction coefficient by using a least square method, specifically comprising: ； ； wherein, is a proportional correction factor; is a predicted mechanical property value of the i-th training sample point; is a predicted mechanical property value of the i-th training sample point; is an experimental actual mechanical property value of the i-th training sample point; is an experimental actual mechanical property value of the i-th training sample point; , respectively represent the average values of the predicted mechanical property value and the experimental actual mechanical property value; is a bias correction parameter; The general correction equation of the analytical model is established based on the least square method, and the general correction equation is used for dynamic calibration of the analytical model, and the general correction equation is .

9. A computer device comprising: A memory, a processor, and a computer program stored on the memory and capable of running on the processor, characterized in that the processor executes the computer program to implement the optical fiber cable braiding process parameter optimization method of any one of claims 1-8.

10. A computer-readable storage medium having stored thereon a computer program, characterized in that, The computer program is executed by the processor to implement the optical fiber cable braiding process parameter optimization method of any one of claims 1-8.

Images