Multi-objective Optimization Method and System for Process Production Technology Based on Digital Twin

Abstract

The present invention relates to the field of digital twin technology, and provides a multi-objective optimization method for process production technology based on digital twin. A multi-objective optimization problem with clear optimization objectives and constraint conditions is established according to actual requirements; relevant digital twin models are established to realize real-time interaction between the physical production system and the virtual production system to assist optimization; after obtaining the Pareto optimal solution, the dynamic parameter group correlation degree evaluation, entropy weight method and preference order method for similarity to the ideal solution are comprehensively utilized to form a fast and accurate process decision. The method and system disclosed by the present invention can solve technical problems such as multi-objective conflicts, untimely feedback, and inaccurate decision-making existing in related process plans, and can effectively improve the processing quality, increase the utilization rate of raw materials, and reduce the energy consumption of the whole process.

Description

Technical Field

[0001] The present invention relates to the technical field of digital twins, and particularly relates to a multi-objective optimization method and system for process production processes based on digital twins. Background Art

[0002] The process industry mainly includes basic raw material industries such as chemical engineering, iron and steel, non-ferrous metals, and building materials. It is the pillar and basic industry of the national economy and an important supporting force for the continuous growth of China's economy. Process industry enterprises usually consume a large amount of resources such as electricity, coal, and gas during the production process. In the process production process, under a certain working condition, different regulation schemes affect the types and total amounts of consumed energy. With the gradual development of the process industry towards large-scale, integrated, and continuous development, the increasingly fierce market competition at home and abroad has put higher requirements on comprehensive indicators such as product quality, overall cost, process safety, production efficiency, and energy consumption level of the process industry. Therefore, it is crucial to quickly obtain a more effective and accurate regulation scheme while considering multiple process production process objectives.

[0003] The optical fiber production process is a typical process production process. The optical fiber, fully known as the optical waveguide fiber, is a fiber made of glass or plastic and can be used as an optical conduction tool. Its transmission principle is total internal reflection of light. Optical fiber transmission has the advantages of wide frequency band, low loss, light weight, strong anti-interference ability, high fidelity, reliable performance, and continuously decreasing cost. The optical fiber preform is the core raw material for manufacturing quartz series optical fibers and is known in the industry as the "pearl on the crown" of the optical communication industry; the optical fiber preform has a specific refractive index profile and its diameter ranges from dozens of millimeters to hundreds of millimeters. Its processing is the most important part of the optical fiber process and has a decisive effect on the type and various properties of the optical fiber.

[0004] The primary drawing process of the optical fiber preform is a thermorheological process in which the quartz glass rod produces an axial elongation and radial contraction effect due to the differential drawing at the top and bottom in a high-temperature molten state. The specific process is as follows: The mother rod of the optical fiber preform is sent downward from the top of the drawing tower into the central area of the graphite high-temperature furnace. After being clamped at both ends by the tail stock, it is heated to the molten state at a temperature of about two thousand degrees Celsius, and then the mother rod is drawn to the target rod diameter under the combined action of low-speed feeding at the top and high-speed traction at the bottom; this process has the characteristics of large batch, continuity, and irreversibility and is a typical process type process.

[0005] In the prior art, relevant manufacturing enterprises will encounter the following problems when performing the primary drawing process of the optical fiber preform: The global improvement of the relevant process is a multi-objective optimization task, and the setting and adjustment of process parameters still largely rely on past experience, which leads to inaccurate working condition identification and process decision-making. Summary of the Invention

[0006] Through long-term practice, it has been found that in the prior art, the various objectives in the multi-objective optimization of the single-drawing process of optical fiber preforms have strong conflicts, interfere with each other, have high complexity and large scale differences; due to the continuity and irreversibility of the process, the on-site test cost of relevant process plans is high and the difficulty is great, and there is even a certain risk, which will delay the accurate feedback of the effectiveness of the process plan; the setting and adjustment of process parameters still largely rely on past experience, which leads to inaccurate working condition identification and process decision-making.

[0007] In view of this, the present invention aims to propose a multi-objective optimization method for process production based on digital twin. The multi-objective optimization method for process production based on digital twin includes:

[0008] Step S1, collecting the processing data in the optical fiber preform drawing process and performing data preprocessing; wherein, the data preprocessing includes outlier processing, interpolation processing, training set and test set division, and normalization processing;

[0009] Step S2, analyzing the correlation relationship between the optical fiber preform drawing process data and parameters, and determining that the optimization objective function at least includes the average rod diameter difference Δd after drawing, the total power consumption W during the whole process, and the raw material utilization rate δ; wherein, the optimization variables include the processing temperature T oven , the feeding speed V f , the drawing speed V d , and taking the value ranges of (T oven , V f , V d ), the magnitude of the equipment electric power P t , and the range of the mother rod diameter D as the constraint conditions;

[0010] Step S3, establishing a multi-physical field simulation model for the optical fiber preform drawing process; inputting solid heat transfer, laminar flow, fluid heat transfer, material properties and process parameters, wherein the material properties at least include rod density, surface tension coefficient, specific heat ratio, and Poisson's ratio; the process parameters include feeding speed, drawing speed, and hot furnace temperature; outputting the velocity field, pressure field, and temperature field distributions of the rod heat flow in a high-temperature environment;

[0011] Step S4, extracting the key setting steps and post-processing display interface of the multi-physical field simulation model of the optical fiber preform drawing process to form a control visualization interface for the virtual simulation system of the optical fiber preform drawing process and connecting it to the physical production system; wherein, the physical production system includes the monitoring response data of the on-site processing status of the workshop, and the difference between the monitoring response data of the on-site processing status of the workshop and the response data of the multi-physical field simulation model of the optical fiber preform drawing process is used to establish the objective function of the multi-physical field simulation model of the optical fiber preform drawing process;

[0012] Step S5: Calculate N Pareto optimal solutions and conduct a comprehensive evaluation. Based on the correlation degree P between the Pareto optimal solutions and the parameter group to be optimized currently cd and the changing trends of each index value, conduct a comprehensive evaluation;

[0013]

[0014] wherein, P cd represents the correlation degree, m is the index dimension, is the parameter group to be optimized, is the optimized parameter group, and N is a positive integer greater than 1;

[0015] Step S6: Based on N Pareto optimal solution candidate schemes and each index weight, construct a weighted decision matrix, calculate the positive and negative ideal solutions within the parameter interval in sequence, the distances between each candidate scheme and the positive and negative ideal solutions, and the relative closeness to the ideal solution, and generate the final order of the advantages and disadvantages of the schemes and the optimal operating parameters.

[0016] Preferably, in step S4, the visual control interface of the optical fiber preform drawing process virtual simulation system further includes a user interaction interface for constructing a multi-physical field simulation model of the optical fiber preform drawing process, setting feedback information, and designing a simulation effect evaluation window.

[0017] Preferably, the difference between the on-site processing state monitoring response data of the workshop and the response data of the multi-physical field simulation model of the optical fiber preform drawing process is

[0018]

[0019] The constraint condition is p lb ≤p≤p ub , wherein, p, p lb and p ub are respectively the parameter vector group and its upper and lower bounds, and R refers to the difference between the monitoring response data r m and the calculated simulation model response data r c .

[0020] Preferably, through the pseudo-inverse method Δp = R S -1 ΔR, iteratively correct the parameters of the multi-physical field simulation model of the optical fiber preform drawing process relative to the on-site processing state of the workshop, wherein, R S -1 is the pseudo-inverse matrix of R S ; wherein, the sensitivity matrix

[0021] Preferably, a digital twin model is constructed from a multi-physical field simulation model of the optical fiber preform drawing process. During the operation of the digital twin model, the optimization objectives, constraint conditions, and optimization variables are monitored in real time in the physical production system; according to the optimization result data, the historical working condition database is updated.

[0022] Preferably, based on N Pareto optimal solution candidate schemes and the weights of each index, a weighted decision matrix is constructed, including

[0023] Step S61, construct an initial decision matrix according to each optimization objective and perform normalization processing;

[0024] Step S62, calculate the entropy value e of each index in the initial decision matrix respectively j and the weight coefficient h j ,

[0025]

[0026]

[0027] where N is the number of samples, p ij is the normalized value of the j-th index of the i-th sample; when p ij =0, set p ij lnp ij =0, and k is 1 / lnN.

[0028] The present invention also discloses a system for executing the above-mentioned multi-objective optimization method for the process production process based on digital twin. The system includes

[0029] An acquisition unit for collecting processing data in the optical fiber preform drawing process and performing data preprocessing; wherein, the data preprocessing includes outlier processing, interpolation processing, training set and test set division, and normalization processing;

[0030] An initialization unit for analyzing the correlation between the optical fiber preform drawing process data and parameters, and determining that the optimization objective function at least includes the average rod diameter difference Δd after drawing, the total power consumption W during the whole process, and the raw material utilization rate δ; wherein, the optimization variables include the processing temperature T oven , the feeding speed V f , the drawing speed V d , and take the value ranges of (T oven , V f , V d ), the magnitude of the equipment electric power P t and the range of the mother rod diameter D as constraint conditions;

[0031] A model unit for establishing a multi-physical field simulation model of the optical fiber preform drawing process; inputting solid heat transfer, laminar flow, fluid heat transfer, material properties and process parameters, where the material properties include at least the rod density, surface tension coefficient, specific heat ratio, and Poisson's ratio; the process parameters include the feeding speed, drawing speed, and hot furnace temperature; outputting the velocity field, pressure field, and temperature field distributions of the rod heat flow in a high-temperature environment.

[0032] A twin unit for extracting the key setup steps and post-processing display interface of the multi-physical field simulation model of the optical fiber preform drawing process, forming a visual control interface for the virtual simulation system of the optical fiber preform drawing process, and connecting to the physical production system; where the physical production system includes the monitoring response data of the on-site processing status in the workshop, and the difference between the monitoring response data of the on-site processing status in the workshop and the response data of the multi-physical field simulation model of the optical fiber preform drawing process is used to establish the objective function of the multi-physical field simulation model of the optical fiber preform drawing process.

[0033] An evaluation unit for calculating N Pareto optimal solutions and conducting a comprehensive evaluation, and making a comprehensive evaluation based on the correlation degree P between the Pareto optimal solutions and the parameter group to be optimized currently cd and the changing trends of each index value.

[0034]

[0035] where P cd represents the correlation degree, m is the index dimension, is the parameter group to be optimized, is the optimized parameter group.

[0036] A screening unit for constructing a weighted decision matrix based on N Pareto optimal solution candidate schemes and each index weight, calculating the positive and negative ideal solutions within the parameter interval in sequence, the distances between each candidate scheme and the positive and negative ideal solutions, and the relative closeness to the ideal solution, and generating the order of the final scheme's quality and the optimal operating parameters.

[0037] Preferably, the twin unit includes a visualization module for constructing the user interaction interface of the multi-physical field simulation model of the optical fiber preform drawing process, setting feedback information, and designing a simulation effect evaluation window during the process of generating the visual control interface for the virtual simulation system of the optical fiber preform drawing process.

[0038] The present invention discloses an electronic device, including a memory and a processor: the memory is used for storing a computer program; the processor is used for implementing the above-mentioned multi-objective optimization method for the process production technology based on digital twin when executing the computer program.

[0039] The present invention also provides a computer-readable storage medium, on which a computer program is stored, and when the program is executed by a processor, it implements the method provided by the present invention.

[0040] Compared with the prior art, the multi-objective optimization method for the drawing process of optical fiber preforms provided by the present invention includes steps S1-S6, that is, collecting the processing data in the drawing process of optical fiber preforms and performing data preprocessing; analyzing the correlation relationship between the data and parameters of the drawing process of optical fiber preforms to determine the optimization objective function; establishing a multi-physical field simulation model for the drawing process of optical fiber preforms; inputting solid heat transfer, laminar flow, fluid heat transfer, material properties and process parameters; extracting the key setting steps and post-processing display interface of the multi-physical field simulation model for the drawing process of optical fiber preforms to form a visual control interface for the virtual simulation system of the drawing process of optical fiber preforms, and connecting it with the physical production system; calculating N Pareto optimal solutions and conducting a comprehensive evaluation, constructing a weighted decision matrix based on the N Pareto optimal solution candidate schemes and each index weight, calculating the positive and negative ideal solutions within the parameter interval in sequence, the distances between each candidate scheme and the positive and negative ideal solutions, and the relative closeness to the ideal solution, and generating the order of the pros and cons of the final scheme and the optimal operating parameters. This method establishes a multi-objective optimization problem with clear optimization objectives and constraint conditions according to actual needs; establishes relevant digital twin models to realize real-time interaction between the physical production system and the virtual production system to assist optimization; after obtaining the Pareto optimal solutions, comprehensively utilizes the dynamic parameter group correlation degree evaluation, entropy weight assignment method and the preference ranking method based on the similarity to the ideal solution to form a fast and accurate process decision. The method disclosed by the present invention can solve the technical problems such as multi-objective conflicts, untimely feedback and inaccurate decision-making existing in the relevant process schemes, and can effectively improve the processing quality, improve the utilization rate of raw materials and reduce the energy consumption of the whole process.

[0041] Other features and advantages of the present invention will be described in detail in the subsequent specific implementation part. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] The drawings constituting a part of the present invention are used to provide a further understanding of the present invention. The schematic embodiments and descriptions thereof of the present invention are used to explain the present invention and do not constitute an improper limitation to the present invention. In the drawings:

[0043] Figure 1 is a schematic flow chart of the multi-objective optimization method for the process production process based on digital twin of the present invention;

[0044] Figure 2 is a flow chart for constructing the multi-physical field simulation model of the drawing process of optical fiber preforms of the present invention;

[0045] Figure 3 is a flow chart for the comprehensive evaluation and decision-making of the Pareto optimal solutions of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0046] The specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are only for the purpose of illustrating and explaining the present invention, and are not intended to limit the present invention.

[0047] In order to enable those skilled in the art to better understand the solution of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without making creative efforts shall fall within the protection scope of the present invention.

[0048] It should be noted that the terms "first", "second", "third", "fourth", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that such data used in appropriate cases can be interchanged so as to implement the embodiments of the present invention described herein. In addition, the terms "comprising" and "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units not clearly listed or inherent to these processes, methods, products or devices.

[0049] In order to solve the problems in the prior art that in the multi-objective optimization of the single-stage drawing process of the optical fiber preform, each objective has strong conflicts, mutual interference, high complexity and large scale differences; due to the continuity and irreversibility of the process, the on-site test cost of the relevant process plan is high and the difficulty is great, and there is even a certain risk, which will delay the accurate feedback of the effectiveness of the process plan; the setting and adjustment of process parameters still largely rely on past experience, which leads to inaccurate working condition identification and process decision-making. The present invention provides a multi-objective optimization method for process production based on digital twin, as Figures 1 - 3 shown, the multi-objective optimization method for process production based on digital twin includes,

[0050] Step S1, collecting the processing data in the optical fiber preform drawing process and performing data preprocessing; wherein, the data preprocessing includes outlier processing, interpolation processing, training set and test set division, and normalization processing;

[0051] Step S2, analyzing the correlation relationship between the optical fiber preform drawing process data and parameters, and determining that the optimization objective function at least includes the average rod diameter difference Δd after drawing, the total power consumption W during the whole process, and the raw material utilization rate δ; wherein, the optimization variables include the processing temperature T oven , the feeding speed V f, drawing speed V d , taking the value ranges of (T oven , V f , V d ), the magnitude of the equipment electric power P t and the range of the mother rod diameter D as constraint conditions;

[0052] Step S3, establish a multi - physical - field simulation model for the fiber preform drawing process; input solid heat transfer, laminar flow, fluid heat transfer, material properties and process parameters. Among them, the material properties include at least the rod density, surface tension coefficient, specific heat ratio, Poisson's ratio; the process parameters include the feeding speed, drawing speed and hot furnace temperature; output the velocity field, pressure field and temperature field distributions of the rod heat flow under high - temperature environment;

[0053] Step S4, extract the key setting steps and post - processing display interface of the multi - physical - field simulation model for the fiber preform drawing process to form a visual control interface for the virtual simulation system of the fiber preform drawing process, and connect it to the physical production system; among them, the physical production system includes the monitoring response data of the on - site processing status of the workshop, and the difference between the monitoring response data of the on - site processing status of the workshop and the response data of the multi - physical - field simulation model for the fiber preform drawing process is used to establish the objective function of the multi - physical - field simulation model for the fiber preform drawing process;

[0054] Step S5, calculate N Pareto optimal solutions and conduct a comprehensive evaluation. Make a comprehensive evaluation according to the correlation degree P cd between the Pareto optimal solution and the parameter group to be optimized currently and the changing trends of each index value;

[0055]

[0056] Among them, P cd represents the correlation degree, m is the index dimension, is the parameter group to be optimized, is the optimized parameter group; N is a positive integer greater than 1;

[0057] Step S6, based on N Pareto - optimal - solution candidate schemes and the weights of each index, construct a weighted decision - making matrix, calculate the positive and negative ideal solutions within the parameter interval in turn, the distances between each candidate scheme and the positive and negative ideal solutions, and the relative closeness to the ideal solution, and generate the order of the final scheme quality and the optimal operating parameters.

[0058] The multi-objective optimization method for the drawing process of optical fiber preforms provided by the present invention includes steps S1 - S6, that is, collecting the processing data in the drawing process of optical fiber preforms and performing data preprocessing; analyzing the correlation relationship between the data and parameters of the drawing process of optical fiber preforms to determine the optimization objective function; establishing a multi-physical field simulation model for the drawing process of optical fiber preforms; inputting solid heat transfer, laminar flow, fluid heat transfer, material properties and process parameters; extracting the key setting steps and post-processing display interface of the multi-physical field simulation model for the drawing process of optical fiber preforms to form a visual control interface for the virtual simulation system of the drawing process of optical fiber preforms and connecting it to the physical production system; calculating N Pareto optimal solutions and conducting a comprehensive evaluation, constructing a weighted decision matrix based on the N Pareto optimal solution candidate schemes and the weights of each index, calculating the positive and negative ideal solutions within the parameter interval in turn, the distances between each candidate scheme and the positive and negative ideal solutions, and the relative closeness to the ideal solution, and generating the order of the pros and cons of the final scheme and the optimal operating parameters. This method establishes a multi-objective optimization problem with clear optimization objectives and constraint conditions according to actual requirements; establishes a relevant digital twin model to realize real-time interaction between the physical production system and the virtual production system to assist optimization; after obtaining the Pareto optimal solutions, comprehensively utilize the dynamic parameter group correlation degree evaluation, entropy weight method and ideal solution similarity order preference method to form a fast and accurate process decision. The method disclosed by the present invention can solve the technical problems such as multi-objective conflicts, untimely feedback and inaccurate decision-making existing in the relevant process schemes, and can effectively improve the processing quality, improve the utilization rate of raw materials and reduce the energy consumption of the whole process.

[0059] Combined with the analysis of the correlation relationship between the data and parameters of the drawing process of optical fiber preforms, the optimization of the primary drawing process scheme of optical fiber preforms takes the average rod diameter difference Δd after drawing, the total power consumption W of the whole process and the raw material utilization rate δ as the optimization objective functions, which respectively reflect the overall processing quality, energy consumption level and process advancement.

[0060] Taking the processing temperature T oven , the feeding speed V f , and the drawing speed V d as the optimization variables, and taking the value ranges of (T oven , V f , V d ), the magnitude of the equipment electric power P t and the range of the mother rod diameter D as the model constraint conditions. Among them, the optimization objective function is expressed as,

[0061]

[0062]

[0063]

[0064] Among them, when |d t-d target When |d t -d target | ≤ 3, s(t) = 1. For example, the constraint conditions are: 1700°C ≤ T oven ≤ 2100°C for the furnace temperature, 0 mm / min ≤ V f ≤ 20 mm / min for the feeding speed; 0 mm / min ≤ V d ≤ 100 mm / min for the stretching speed; and there is V f ≤ V d ; while the electric power P t of the equipment and the diameter D of the mother rod respectively satisfy 30 kW ≤ P t ≤ 50 kW and 80 mm ≤ D ≤ 140 mm. In f1 and f3, T is equal to the number of samples taken within a certain time period. For example, if samples are taken once per second, then T = 60 in 1 minute; while in f2, T represents the continuous time length of a certain time period.

[0065] Among them, d t and d target are respectively the value of the rod diameter after stretching obtained at a certain moment t and the target rod diameter value preset before processing. According to the tolerance requirements, d target = 55 ± 3 mm; s(t) is a characterization of the usability of the rod after stretching at a certain moment t. When s(t) = 1, it means that the rod diameter after stretching at this moment meets the tolerance requirements, that is, it contributes to improving the utilization rate of raw materials, while when s(t) = 0, it does not.

[0066] In order to establish a relevant digital twin model, realize real-time interaction between the physical production system and the virtual production system, and assist in optimization, it can effectively improve the processing quality, enhance the utilization rate of raw materials, and reduce the energy consumption of the whole process. Since the multi-physical field simulation model of the optical fiber preform stretching process is the basis for establishing the relevant process digital twin model. The optical fiber preform is an axisymmetric rod with a very small diameter fluctuation range. In order to reduce the simulation time, the original three-dimensional model is simplified to a two-dimensional axisymmetric model. As Figure 2 shown, use COMSOL Multiphysics 5.5 software to create a new model and select transient analysis.

[0067] Digital twin is based on high-fidelity simulation or mapping models, efficient data transmission and processing technologies, and fast feedback and control mechanisms, and can monitor the state, extract features, evaluate behaviors, guide regulation, and comprehensively optimize a component, product, or system throughout its life cycle. By constructing the digital twin system of the first stretching of the optical fiber preform, through the parallel operation, real-time interaction, and iterative optimization of the on-site production entity and the digital twin body, accurate working condition judgment, rapid abnormal feedback, and timely process decision-making can be realized, which can greatly improve the production quality and efficiency.

[0068] First, set the geometric parameters of the fiber preform master rod and the high-temperature furnace, construct geometric objects and form a geometric union, define the boundary coordinate system and the first tangential direction, and then create a global Cartesian space. The processes to be studied in the fiber preform extension process mainly include the thermal radiation heating of the high-temperature graphite furnace, the traction movement of the glass rod body, and the heat flow in the rod body. The parameters to be calculated include the velocity field, pressure field, and temperature field distributions in the fluid. Therefore, pre-define the solid heat transfer, laminar flow, and fluid heat transfer modules and enter the model developer. Set the material properties such as the rod density, surface tension coefficient, specific heat ratio, Poisson's ratio, and the process parameters such as the feeding speed, stretching speed, and heat furnace temperature, and then determine the relevant properties, boundary conditions, and numerical relationships of the entrances, exits, and interaction interfaces of each physical field.

[0069] Since the numerical simulation of the rheoforming process under consideration needs to deal with the strong deformation of the continuum, the traditional modes based on the Lagrangian description or Eulerian description are no longer applicable. The solution to this problem is a boundary tracking technique based on updating the underlying mesh, which is solved by the moving mesh module in the dynamic mesh. This module is based on the arbitrary Euler-Lagrange description and allows the boundary to move without the mesh following the material. After completing the model construction, select a personalized solver configuration, and finally generate post-processing results and perform visual analysis according to different requirements.

[0070] For the convenience of practical application, in a more preferred case of the present invention, in step S4, the control visualization interface of the fiber preform stretching process virtual simulation system further includes a user interaction interface for constructing a multi-physical field simulation model of the fiber preform stretching process, setting feedback information, and designing a simulation effect evaluation window. For example, extract the key setting steps and post-processing display interface in the model developer based on the COMSOL Multiphysics 5.5 software APP developer. In addition, add a user interaction interface, set feedback information, design a simulation effect evaluation window, and write methods, etc., and finally form the control visualization interface of the fiber preform stretching process virtual simulation system.

[0071] The control visualization interface of the virtual simulation system for the fiber preform drawing process consists of a main function area, a parameter input section, and a result viewing section; the main function area includes functions such as resetting all input box and table parameter values, drawing geometric figures, drawing simulation grids, calculating and drawing post-processing result diagrams, viewing simulation process documents, and viewing help documents; the parameter input section contains material parameters, such as temperature-independent parameters, temperature-dependent parameters, and processing parameters, such as input boxes and table parameter values for dimension parameters and process parameters, and data can be manually rewritten or imported from a data file; the result viewing section consists of geometry and grid, one-dimensional drawing, two-dimensional drawing, three-dimensional drawing group, real-time process display, and solution status, with functions such as customizing grid parameters, viewing the drawing situation of the selected item in real-time simulation, drawing by dragging the time progress bar, and point selection for value checking. The parameters observed and measured online include the necking amount of the rod, the tension of the extension part and the feeding part, the electrical power of the machine, the cross-sectional pressure, the flow rate of the rod, and the temperature distribution, etc. After connecting the on-site processing status monitoring to this control interface, the sensor measurement responses transmitted by the physical production system and the model calculation responses transmitted by the virtual simulation system will be used to establish relevant target optimization functions, with the aim of minimizing the error between the two and ensuring the fidelity of the digital twin model.

[0072] The difference between the on-site processing status monitoring response data in the workshop and the multi-physical field simulation model response data of the fiber preform drawing process is

[0073]

[0074] The constraint condition is p lb ≤p≤p ub where p, p lb and p ub are the parameter vector group and its upper and lower bounds respectively, and R refers to the difference between the monitored response data r m and the calculated simulation model response data r c Among them, is the sum of the squares of each R(p) vector.

[0075] Therefore, in a more preferred case of the present invention, the multi-physical field simulation model of the fiber preform drawing process is iteratively corrected with respect to the on-site processing status in the workshop by the pseudo-inverse method Δp = R S -1 ΔR, where R s -1 is the pseudo-inverse matrix of R S ; among them, the sensitivity matrix Once the sensitivity matrix of the original system is obtained, algorithms such as the nonlinear least squares method are used to obtain the correction amount of the original parameter group, thereby correcting p to ensure the consistency between the physical production system and the virtual simulation model. During the operation of the digital twin model, the digital twin model is constructed by the multi-physical field simulation model of the optical fiber preform drawing process. During the operation of the digital twin model, the optimization objectives, constraint conditions, and optimization variables are monitored in real time in the physical production system; according to the optimization result data, the historical working condition database is updated, so that the iterative construction of the multi-objective optimization problem becomes more accurate.

[0076] In order to balance convergence and diversity when selecting optimization observation points, the present invention adopts a dynamic acquisition function, which tends to explore in the early stage of the optimization process, that is, it is collected in areas far from the current optimal solution and with large uncertainties, so as to increase the possibility of obtaining other optimal solutions and avoid the optimization process from falling into local optima; in the later stage of the optimization process, it tends to exploit, focusing on collecting in areas adjacent to the current optimal solution and with high predicted expectations (means), so as to obtain possible global optimal solutions and make the optimization process tend to converge quickly.

[0077] When σ(x) = 0, the acquisition function takes the value of 0. When σ(x) ≠ 0, the acquisition function is expressed as

[0078]

[0079] where, f * and μ(x) are the function values at the current optimal solution and point x respectively; α is a coefficient that balances the proportion of exploitation and exploration behaviors, and its value depends on the current number of evaluations FE of the original objective function and the preset maximum number of evaluations FE max of the original objective function, and it shows an increasing trend during the optimization process. The relevant expression is

[0080]

[0081] where, δ is a coefficient used to avoid the optimization process from falling into local optima, and its value depends on the distance d(x, x * ) between the predicted sampling point x and the current sampling point x. If this distance is less than a certain distance threshold * , the coefficient change mechanism will be triggered to assist the optimization process to jump out of the local optimum, that is, when δ = 1, and when δ > 1; the distance penalty function is used to further avoid the sampling distances being too close before and after, and this function will help improve the diversity of the optimization solutions. When x * ∈{x1, x2, ···, x n}, P(x, x * ) = 0, and when :

[0082]

[0083] Among them, The introduction of the natural exponential function and the arctangent function can respectively rapidly reduce the influence of a point when the distance from x * is too far from x and avoid the excessive growth of the penalty function. Based on the traditional expected improvement acquisition function, the maximum mean μ(x) max and the maximum variance μ(x) max provided by the Gaussian process regression model are respectively normalized to eliminate the influence of the original objective function and constraint conditions of different scales on the optimization process.

[0084] To reduce the optimization process from falling into a local optimum, thus causing the optimization result to terminate. In a more preferred case of the present invention, in step S5, Bayesian optimization is used to calculate N Pareto optimal solutions and conduct a comprehensive evaluation. Bayesian optimization is a global optimization method based on a meta-model in sequential sampling with uncertainty. When using a meta-model to assist multi-objective optimization, usually a meta-model is constructed for each objective as a surrogate function, then a suitable acquisition function is established to guide the sampling of observation points, and finally a set of Pareto optimal solutions is obtained. The meta-model used to fit the unknown objective function is a Gaussian process regression model. A Gaussian process is a non-parametric stochastic process model in which observations occur in a continuous domain, which can fit a black-box function and give the confidence level of the fitting result.

[0085] After obtaining the set of Pareto optimal solutions, in order to comprehensively utilize the dynamic parameter group correlation degree evaluation, the entropy weight method, and the preference order method of the similarity degree to the ideal solution to form a fast and accurate process decision. In a more preferred case of the present invention, based on N Pareto optimal solution candidate schemes and the weights of each index, a weighted decision matrix is constructed, including

[0086] Step S61, construct an initial decision matrix according to each optimization objective and conduct a normalization process;

[0087] Step S62, calculate the entropy value e j of each index in the initial decision matrix and the weight coefficient h j ,

[0088]

[0089]

[0090] where N is the number of samples, p ij is the normalized value of the j-th index of the i-th sample; when p ij = 0, set p ij lnp ij= 0, where k is 1 / lnN. When the entropy weight coefficient h j is larger, the more information the index contains, that is, the greater its role in comprehensive evaluation.

[0091] Among them, the entropy weight method is a mathematical method used to judge the dispersion degree of a certain index. The greater the dispersion degree, the greater the influence of the index on comprehensive evaluation. Therefore, according to the variation degree of each index, the weights of each index can be calculated using information entropy, providing a basis for multi-index comprehensive evaluation.

[0092] Among them, the Technique for Order Preference by Similarity to an Ideal Solution (TOPSIS) can rank according to the degree of closeness of a finite number of evaluation objects to the idealized goal, that is, evaluate the relative advantages and disadvantages of existing objects. It ranks by detecting the distances between the evaluation objects and the optimal solution and the worst solution. If an evaluation object is closest to the optimal solution and farthest from the worst solution at the same time, it is the best; otherwise, it is not the optimal. Among them, the index values of the optimal solution reach the optimal values of each evaluation index. The index values of the worst solution reach the worst values of each evaluation index.

[0093] Based on the candidate solutions and the weights of each index, a normalized weighted decision matrix is constructed. The positive and negative ideal solutions within the parameter interval are calculated in turn, that is, the distances between the optimal solution and the worst solution, each solution and the positive and negative ideal solutions, and the relative closeness to the ideal solution. Finally, the order of the advantages and disadvantages of the solutions and the optimal operating parameters are generated.

[0094]

[0095]

[0096]

[0097] Among them, and C i are the distances between each solution and the optimal solution, the distances between each solution and the worst solution, and the relative closeness to the optimal solution respectively; and are the optimal solution and the worst solution respectively; ω j is the weight of the jth index.

[0098] The present invention also discloses a system for executing the above multi-objective optimization method for the process production process based on digital twins. The system includes,

[0099] An acquisition unit for collecting processing data in the fiber preform drawing process and performing data preprocessing; among them, the data preprocessing includes outlier processing, interpolation processing, training set and test set division, and normalization processing;

[0100] An initialization unit for analyzing the correlation relationship between the drawing process data and parameters of an optical fiber preform, and determining that the optimization objective function at least includes the average rod diameter difference Δd after drawing, the total power consumption W during the whole process, and the raw material utilization rate δ; among them, the optimization variables include the processing temperature T oven , the feeding speed V f , the drawing speed V d , taking the value ranges of (T oven , V f , V d ), the magnitude of the equipment electric power P t and the range of the mother rod diameter D as constraint conditions;

[0101] A model unit for establishing a multi-physical field simulation model of the optical fiber preform drawing process; inputting solid heat transfer, laminar flow, fluid heat transfer, material properties and process parameters, where the material properties at least include the rod density, surface tension coefficient, specific heat ratio, and Poisson ratio; the process parameters include the feeding speed, drawing speed, and hot furnace temperature; outputting the velocity field, pressure field, and temperature field distributions of the rod heat flow in a high-temperature environment;

[0102] A twin unit for extracting the key setting steps and post-processing display interface of the multi-physical field simulation model of the optical fiber preform drawing process, forming a visual control interface for the virtual simulation system of the optical fiber preform drawing process, and connecting it to the physical production system; among them, the physical production system includes the monitoring response data of the on-site processing status of the workshop, and the difference between the monitoring response data of the on-site processing status of the workshop and the response data of the multi-physical field simulation model of the optical fiber preform drawing process is used to establish the objective function of the multi-physical field simulation model of the optical fiber preform drawing process;

[0103] An evaluation unit for calculating N Pareto optimal solutions and conducting a comprehensive evaluation, and making a comprehensive evaluation based on the correlation degree P cd between the Pareto optimal solutions and the parameter group to be optimized currently and the changing trends of each index value;

[0104]

[0105] Among them, P cd represents the correlation degree, m is the index dimension, is the parameter group to be optimized, is the optimized parameter group;

[0106] A screening unit for constructing a weighted decision matrix based on N Pareto optimal solution candidate schemes and each index weight, calculating the positive and negative ideal solutions within the parameter interval in turn, the distances between each candidate scheme and the positive and negative ideal solutions, and the relative closeness to the ideal solution, and generating the order of the pros and cons of the final scheme and the optimal operating parameters.

[0107] The system for implementing the above multi-objective optimization method of the process production process based on digital twin provided by the present invention includes an acquisition unit, which is used to collect the processing data in the optical fiber preform drawing process and perform data preprocessing; an initialization unit is used for analyzing the correlation relationship between the optical fiber preform drawing process data and parameters to determine the optimization objective function; a model unit is used to establish a multi-physical field simulation model of the optical fiber preform drawing process; for inputting solid heat transfer, laminar flow, fluid heat transfer, material properties and process parameters; a twin unit is used to extract the key setting steps and post-processing display interface of the multi-physical field simulation model of the optical fiber preform drawing process, form a control visualization interface of the virtual simulation system for the optical fiber preform drawing process, and connect it with the physical production system; an evaluation unit is used to calculate and obtain N Pareto optimal solutions and conduct a comprehensive evaluation; a screening unit is used to construct a weighted decision matrix based on N Pareto optimal solution candidate schemes and each index weight, calculate the positive and negative ideal solutions within the parameter interval in turn, the distances between each candidate scheme and the positive and negative ideal solutions, and the relative closeness to the ideal solution, and generate the final order of the advantages and disadvantages of the scheme and the optimal operating parameters. This system establishes a multi-objective optimization problem with clear optimization objectives and constraint conditions according to actual needs; establishes a relevant digital twin model to realize real-time interaction between the physical production system and the virtual production system to assist optimization; after obtaining the Pareto optimal solution, comprehensively uses the dynamic parameter group correlation degree evaluation, entropy weight method and ideal solution similarity order preference method to form a fast and accurate process decision. The method disclosed by the present invention can solve technical problems such as multi-objective conflict, untimely feedback, and inaccurate decision-making existing in related process schemes, and can effectively improve the processing quality, improve the utilization rate of raw materials and reduce the energy consumption of the whole process.

[0108] In order to better present the control visualization interface of the virtual simulation system for the optical fiber preform drawing process, after being connected to the physical production system, the multi-physical field simulation model of the optical fiber preform drawing process and the monitoring response data of the on-site processing status in the workshop are better presented in the virtual environment. In a more preferred case of the present invention, the twin unit includes a visualization module, which is used to construct a user interaction interface of the multi-physical field simulation model of the optical fiber preform drawing process, set feedback information, and design a simulation effect evaluation window during the process of generating the control visualization interface of the virtual simulation system for the optical fiber preform drawing process.

[0109] The present invention also discloses an electronic device, including a memory and a processor: the memory is used to store a computer program; the processor is used to implement the above method when executing the computer program.

[0110] Furthermore, the present invention also provides a computer-readable storage medium, on which a computer program is stored, and when the program is executed by a processor, it implements the method provided by the present invention.

[0111] It should be noted that, for the foregoing method embodiments, for the sake of simple description, they are all expressed as a series of action combinations. However, those skilled in the art should understand that the present invention is not limited by the described action sequence, because according to the present invention, certain steps can be performed in other sequences or simultaneously. Secondly, those skilled in the art should also understand that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily essential to the present invention.

[0112] In the above embodiments, the descriptions of the respective embodiments have their own emphases. For the parts not detailed in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0113] In several embodiments provided by the present invention, it should be understood that the disclosed device can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of the units is only a logical function division. In actual implementation, there may be other division methods. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the displayed or discussed couplings or direct couplings or communication connections to each other can be through some interfaces. The indirect couplings or communication connections of the devices or units can be in electrical or other forms.

[0114] The units described as separate components may or may not be physically separated. The components displayed as units may or may not be physical units, that is, they can be located in one place, or can be distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment. [[ID=eleven]]

[0115] In addition, the functional units in each embodiment of the present invention can be integrated into one processing unit, or each unit can exist physically alone, or two or more units can be integrated into one unit. The above integrated units can be implemented in the form of hardware or in the form of software functional units.

[0116] When the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for causing a computer device (which can be a personal computer, a mobile terminal, a server, or a network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes: various media that can store program codes, such as USB flash drives, read-only memories (ROMs), random access memories (RAMs), mobile hard disks, magnetic disks, or optical discs.

[0117] The foregoing is only the preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention can have various modifications and changes. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A multi-objective optimization method for process production technology based on digital twin, characterized in that, The multi-objective optimization method for the process production process based on digital twin includes Step S1: Collect the processing data in the optical fiber preform drawing process and perform data preprocessing. Among them, the data preprocessing includes outlier processing, interpolation processing, training set and test set division, and normalization processing. Step S2: Analyze the correlation between the optical fiber preform drawing process data and parameters, and determine that the optimization objective function includes at least the average rod diameter difference after drawing , the total power consumption throughout the process , and the raw material utilization rate ; among them, the optimization variables include the processing temperature , the feeding speed , the drawing speed , and take 's value range, 's value range, 's value range, the size of the equipment electric power , and the range of the mother rod diameter as the constraint conditions; Step S3: Establish a multi-physical field simulation model for the optical fiber preform drawing process. Input solid heat transfer, laminar flow, fluid heat transfer, material properties, and process parameters. Among them, the material properties at least include rod density, surface tension coefficient, specific heat ratio, and Poisson's ratio. The process parameters include feeding speed, drawing speed, and hot furnace temperature. Output the velocity field, pressure field, and temperature field distributions of the rod heat flow in a high-temperature environment. Step S4: Extract the key setting steps and post-processing display interface of the multi-physical field simulation model for the optical fiber preform drawing process to form a control visualization interface for the virtual simulation system of the optical fiber preform drawing process and connect it to the physical production system. Among them, the physical production system includes the monitoring response data of the on-site processing status in the workshop, and the difference between the monitoring response data of the on-site processing status in the workshop and the response data of the multi-physical field simulation model for the optical fiber preform drawing process is used to establish the objective function of the multi-physical field simulation model for the optical fiber preform drawing process. Step S5, calculate N Pareto optimal solutions and conduct a comprehensive evaluation, and conduct a comprehensive evaluation based on the correlation between the Pareto optimal solutions and the parameter group to be optimized currently and the changing trends of each index value; Among them, represents the degree of association, is the index dimension, is the parameter group to be optimized, is the optimized parameter group; Step S6: Based on N Pareto optimal solution candidate schemes and the weights of each index, construct a weighted decision matrix, calculate the positive and negative ideal solutions within the parameter interval in turn, the distances between each candidate scheme and the positive and negative ideal solutions, and the relative closeness to the ideal solution, and generate the order of the final scheme's advantages and disadvantages and the optimal operating parameters.

2. The multi-objective optimization method for process production technology based on digital twin according to claim 1, wherein, In step S4, the control visualization interface of the virtual simulation system for the optical fiber preform drawing process also includes a user interaction interface for constructing the multi-physical field simulation model of the optical fiber preform drawing process, setting feedback information, and designing a simulation effect evaluation window.

3. The multi-objective optimization method for process production technology based on digital twin according to claim 1, wherein, The difference between the monitoring response data of the on-site processing status in the workshop and the response data of the multi-physical field simulation model for the optical fiber preform drawing process is The constraint conditions are , where , and are the parameter vector group and its upper and lower bounds respectively, refers to the difference between the monitored response data and the response data of the computational simulation model .

4. The multi-objective optimization method for process production technology based on digital twin according to claim 3, wherein, By the pseudo-inverse method Iteratively correct the parameters of the multi-physical field simulation model of the optical fiber preform drawing process relative to the processing state on the workshop floor, where is the pseudo-inverse matrix of; where the sensitivity matrix .

5. The multi-objective optimization method for process production technology based on digital twin according to any one of claims 1-4, characterized in that, Construct a digital twin model from the multi-physical field simulation model of the optical fiber preform drawing process. During the operation of the digital twin model, monitor the optimization objectives, constraint conditions, and optimization variables in the physical production system in real time. Update the historical working condition database according to the optimization result data.

6. The multi-objective optimization method for process production technology based on digital twin according to any one of claims 1-4, characterized in that Based on N Pareto optimal solution candidate schemes and the weights of each index, constructing a weighted decision matrix includes Step S61: Construct an initial decision matrix according to each optimization objective and perform normalization processing. Step S62, calculate the entropy value of each index in the initial decision matrix respectively and the weight coefficient , Among them, is the number of samples, is the -th sample's -th normalized value of the index; when , set , is 1 / lnN.

7. A system for implementing the multi-objective optimization method of the digital-twin-based process production process according to any one of claims 1-6, characterized in that, The system includes An acquisition unit for collecting the processing data in the optical fiber preform drawing process and performing data preprocessing. Among them, the data preprocessing includes outlier processing, interpolation processing, training set and test set division, and normalization processing. An initialization unit for analyzing the correlation between the drawing process data and parameters of an optical fiber preform, and determining that the optimization objective function includes at least the average rod diameter difference after drawing , the total power consumption throughout the process and the raw material utilization rate ; among them, the optimization variables include the processing temperature , the feeding speed , the drawing speed , and taking 's value range, 's value range, 's value range, the size of the equipment electric power and the range of the mother rod diameter as constraint conditions; A model unit for establishing a multi-physical field simulation model for the optical fiber preform drawing process. Input solid heat transfer, laminar flow, fluid heat transfer, material properties, and process parameters. Among them, the material properties at least include rod density, surface tension coefficient, specific heat ratio, and Poisson's ratio. The process parameters include feeding speed, drawing speed, and hot furnace temperature. Output the velocity field, pressure field, and temperature field distributions of the rod heat flow in a high-temperature environment. A twin unit, which is used to extract the key setup steps and post - processing display interface of the multi - physical - field simulation model for the fiber preform drawing process, form the control visualization interface of the virtual simulation system for the fiber preform drawing process, and connect it to the physical production system; wherein, the physical production system includes the monitoring response data of the on - site processing status in the workshop, and the difference between the monitoring response data of the on - site processing status in the workshop and the response data of the multi - physical - field simulation model for the fiber preform drawing process is used to establish the objective function of the multi - physical - field simulation model for the fiber preform drawing process. An evaluation unit is used to calculate N Pareto optimal solutions and conduct a comprehensive evaluation, and conduct a comprehensive evaluation based on the correlation between the Pareto optimal solutions and the parameter group to be optimized currently and the changing trends of each index value; Among them, represents the degree of association, is the index dimension, is the parameter group to be optimized, is the optimized parameter group; A screening unit, which is used to construct a weighted decision matrix based on N Pareto - optimal solution candidate schemes and each index weight, calculate the positive and negative ideal solutions within the parameter interval in sequence, the distances between each candidate scheme and the positive and negative ideal solutions, and the relative closeness to the ideal solution, and generate the final order of the scheme's advantages and disadvantages and the optimal operating parameters.

8. The system according to claim 7, wherein The twin unit includes a visualization module, which is used to construct the user interaction interface of the multi - physical - field simulation model for the fiber preform drawing process, set feedback information, and design a simulation effect evaluation window during the process of generating the control visualization interface of the virtual simulation system for the fiber preform drawing process.

9. An electronic device, characterized in that, It includes a memory and a processor: the memory is used to store computer programs; the processor is used to implement the multi - objective optimization method for the process production technology based on digital twin as described in any one of claims 1 - 6 when executing the computer programs.

10. A machine - readable storage medium, on which instructions are stored, and these instructions are used to make a machine execute the multi - objective optimization method for the process production technology based on digital twin as described in any one of claims 1 - 6 of the present application.

Images