Lining cloth composite structure optimization method and system using twinborn simulation

Abstract

The invention provides a lining cloth composite structure optimization method and system using twinborn simulation, and relates to the technical field of digital twinborn, and the method comprises the steps: carrying out the twinborn simulation of a lining cloth structure initial scheme, and constructing an initial lining cloth model; performing application scene feature mining on the target garment, and constructing a plurality of simulation application fields; introducing a lining cloth risk detection channel, performing risk detection on the initial lining cloth model in combination with a plurality of simulation application fields, and constructing an initial lining cloth risk detection matrix; and if the lining cloth risk constraint is not met, constructing a lining cloth structure optimization guide space, generating a lining cloth structure adjustment scheme group, and performing multi-field joint optimization to obtain a lining cloth structure optimization strategy. The technical problems that in the prior art, due to the fact that diversified actual application scene factors are not fully combined and a risk detection and evaluation mechanism is lacked, the optimization result adaptability is insufficient, potential risks exist, and the optimization efficiency is low are solved, and adaptability optimization of the lining cloth structure in various application environments is achieved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of digital twinning, in particular to a lining composite structure optimization method and system using twinning simulation. BACKGROUND

[0002] In the field of garment manufacturing, lining plays a key role in enhancing the stiffness of garments, maintaining the stability of the shape, and improving the comfort of wearing. With the growing demand for functional and personalized garments, the lining structure has gradually developed from traditional single material and single layer arrangement to multi-layer composite and multi-functional superimposed lining composite structure. In order to improve the performance of garments in different use scenarios, the optimization of lining composite structure has become an important research direction in garment engineering. At present, the optimization methods of lining composite structure mainly rely on experience-based parameter adjustment, finite element simulation analysis or material selection test. These methods have improved the design rationality to some extent, but generally lack deep combination with actual garment application scenarios, and the optimization process is difficult to fully reflect the complex environmental factors such as stress and deformation of garments in the dynamic wearing process. In addition, the risk assessment means for the optimization results is limited, and it is difficult to find potential failure problems of lining composite structure in multiple scenarios, such as wrinkles, deformation, delamination and other problems. These deficiencies lead to poor adaptability of the optimized lining structure in actual application, and quality problems are prone to occur in the finished garments, which also makes the optimization process inefficient, seriously restricting the development of the garment manufacturing industry. SUMMARY

[0003] The present application provides a lining composite structure optimization method and system using twinning simulation, which solves the technical problems in the prior art that the optimization process of lining composite structure does not fully combine the diversified factors of actual application scenarios and lacks comprehensive detection and evaluation of risks, resulting in insufficient adaptability of the optimization results, structural instability risk and low optimization efficiency. The technical effects of realizing the adaptability optimization of lining structure in multiple application environments and improving the optimization efficiency of lining structure and the stability of lining structure are achieved.

[0004] In view of the above problems, in one aspect, the present application provides a lining composite structure optimization method using twin simulation, which comprises: performing twin simulation according to a lining structure initial scheme of a target garment to construct an initial lining model; performing lining application scene feature mining on the target garment according to an application scene factor set to construct multiple simulation application scenes; introducing a lining risk detection channel, combining the multiple simulation application scenes to perform risk detection on the initial lining model to construct an initial lining risk detection matrix; if the initial lining risk detection matrix does not satisfy a lining risk constraint, constructing a lining structure optimization guide space and adjusting the lining structure initial scheme according to the lining structure optimization guide space to generate a lining structure adjustment scheme group; based on the lining risk detection channel and the lining risk constraint, performing multi-scene joint optimization on the lining structure adjustment scheme group according to the multiple simulation application scenes to obtain a lining structure optimization strategy.

[0005] Preferably, the lining application scene feature mining on the target garment according to the application scene factor set to construct multiple simulation application scenes comprises: performing lining application scene retrieval on the target garment to obtain an application scene retrieval set; performing trend analysis on the application scene retrieval set according to the application scene factor set to construct multiple scene factor parameter domains; performing random value combination according to the multiple scene factor parameter domains to obtain an initial application scene set; performing pairwise difference evaluation on the initial application scene set to obtain a scene difference evaluation set, and performing clustering fusion on the initial application scene set according to the scene difference evaluation set to obtain multiple characteristic application scenes; and building the multiple simulation application scenes according to the multiple characteristic application scenes.

[0006] Preferably, the introduction of the lining risk detection channel, the combination of the multiple simulation application scenes, and the risk detection on the initial lining model to construct an initial lining risk detection matrix comprises: performing simulation application on the initial lining model according to the multiple simulation application scenes to obtain multiple groups of lining application simulation data; inputting the multiple groups of lining application simulation data into the lining risk detection channel to obtain multiple lining risk detection results; constructing multiple application risk detection matrices according to the multiple lining risk detection results, and integrating the multiple application risk detection matrices to generate the initial lining risk detection matrix.

[0007] Preferably, the multiple sets of lining application simulation data are input into the lining risk detection channel to obtain multiple lining risk detection results, including: the lining risk detection channel includes a lining deformation risk detection model, a lining fatigue risk detection model, and a lining recovery risk detection model; a first set of lining application simulation data is extracted based on the multiple sets of lining application simulation data, and features are extracted from the first set of lining application simulation data to obtain first lining deformation simulation data, first lining fatigue simulation data, and first lining recovery simulation data; the first lining deformation simulation data is input into the lining deformation risk detection model to obtain a first lining deformation risk coefficient; the first lining fatigue simulation data is input into the lining fatigue risk detection model to obtain a first lining fatigue risk coefficient; the first lining recovery simulation data is input into the lining recovery risk detection model to obtain a first lining recovery risk coefficient; the first lining deformation risk coefficient and the first lining fatigue risk coefficient are combined to generate a first lining risk detection result, and the first lining risk detection result is added to the multiple lining risk detection results.

[0008] Preferably, if the initial lining risk detection matrix does not meet the lining risk constraints, constructing a lining structure optimization guidance space includes: generating an lining risk anomaly region if the initial lining risk detection matrix does not meet the lining risk constraints; identifying attributes of the lining risk anomaly region to obtain lining risk anomaly factors; performing correlation evaluation on the lining structure variable set of the initial lining structure scheme based on the lining risk anomaly factors to obtain a risk-related structure variable set that meets a predetermined correlation degree; retrieving value records of the risk-related structure variable set based on the target garment to establish multiple correlation variable value constraints; and combining parameter values ​​of the risk-related structure variable set based on the multiple correlation variable value constraints to generate the lining structure optimization guidance space.

[0009] Preferably, based on the lining risk detection channel and the lining risk constraints, a multi-field joint optimization is performed on the lining structure adjustment scheme group according to the multiple simulation application fields to obtain the lining structure optimization strategy, including: performing optimization analysis on the lining structure adjustment scheme group according to the multiple simulation application fields based on the lining risk detection channel and the lining risk constraints, generating multiple single-field lining structure optimization groups; performing intersection analysis on the multiple single-field lining structure optimization groups to obtain a multi-field joint lining structure optimization group; constructing a lining comprehensive risk analysis function based on the lining multidimensional risk indicators, where the lining multidimensional risk indicators include lining deformation risk, lining fatigue risk, and lining recovery risk; and performing lining comprehensive risk minimization optimization on the multi-field joint lining structure optimization group according to the lining comprehensive risk analysis function to generate the lining structure optimization strategy.

[0010] Preferably, based on the lining risk detection channel and the lining risk constraints, the lining structure adjustment scheme group is optimized according to the multiple simulation application fields to generate multiple single-field lining structure optimization groups, including: extracting a first simulation application field according to the multiple simulation application fields; extracting the Pth lining structure adjustment scheme according to the lining structure adjustment scheme group, where P is a positive integer; performing simulation application on the Pth lining structure adjustment scheme according to the first simulation application field to obtain the Pth lining application simulation data; inputting the Pth lining application simulation data into the lining risk detection channel to obtain the Pth lining risk detection result; if the Pth lining risk detection result satisfies the lining risk constraints, adding the Pth lining structure adjustment scheme to the first single-field lining structure optimization group; based on the lining risk detection channel and the lining risk constraints, iteratively optimizing the lining structure adjustment scheme group according to the first simulation application field to construct the first single-field lining structure optimization group, and adding the first single-field lining structure optimization group to the multiple single-field lining structure optimization groups.

[0011] Preferably, the application scenario factor set includes multiple application scenario factors, including environmental factors, time factors, mechanical factors, and behavioral factors.

[0012] Preferably, the lining risk constraints include lining deformation risk constraints, lining fatigue risk constraints, and lining recovery risk constraints.

[0013] On the other hand, the present invention also provides a lining composite structure optimization system utilizing twin simulation. The system includes: a lining structure simulation module for performing twin simulation based on an initial lining structure scheme of a target garment to construct an initial lining model; a scene feature mining module for mining lining application scene features of the target garment based on an application scene factor set to construct multiple simulation application fields; a risk detection module for introducing a lining risk detection channel and performing risk detection on the initial lining model in conjunction with the multiple simulation application fields to construct an initial lining risk detection matrix; an adjustment scheme generation module for constructing a lining structure optimization guidance space if the initial lining risk detection matrix does not meet the lining risk constraints, and adjusting the initial lining structure scheme according to the lining structure optimization guidance space to generate a group of lining structure adjustment schemes; and a lining structure optimization module for performing multi-field joint optimization on the group of lining structure adjustment schemes based on the lining risk detection channel and the lining risk constraints, according to the multiple simulation application fields, to obtain an lining structure optimization strategy.

[0014] One or more technical solutions provided in this invention have at least the following beneficial effects: Based on the initial design of the lining structure of the target garment, a twin simulation is performed to construct an initial lining model, providing reliable structural data and dynamic behavior simulation capabilities for subsequent simulation, evaluation, and optimization. Application scenario features of the lining are mined based on a set of application scenario factors, constructing multiple simulation application scenarios. By mining physical features in specific wearing or usage scenarios, a basis is provided for subsequent performance evaluation and adaptation optimization under multiple scenarios. A lining risk detection channel is introduced, and risk detection is performed on the initial lining model in conjunction with the multiple simulation application scenarios, constructing an initial lining risk detection matrix to provide quantitative feedback and control standards for the optimization process. If the initial lining risk detection matrix does not meet the lining risk constraints, an lining structure optimization guidance space is constructed, and the initial lining structure design is adjusted according to the lining structure optimization guidance space, generating a group of lining structure adjustment schemes to improve optimization efficiency and avoid blind searching. Based on the lining risk detection channel and the lining risk constraints, the group of lining structure adjustment schemes is jointly optimized in multiple simulation application scenarios. Global optimization is performed by integrating the performance feedback of multiple simulation scenarios to achieve a trade-off between schemes. Finally, a lining structure optimization strategy that meets the needs of multiple scenarios and risk constraints is output.

[0015] In summary, this invention constructs a realistic and reliable initial lining model, combines feature mining and simulation of multi-dimensional application scenarios to establish a performance evaluation environment under multiple scenarios, and introduces a lining risk detection channel to achieve comprehensive identification of potential structural failures. Based on this, by constructing a structural optimization guidance space to generate a multi-scheme adjustment group and conducting joint optimization in multiple simulation fields, the adaptability, safety, and stability of the lining structure design under different usage scenarios are effectively improved. This solves the problems of insufficient response to actual scenarios, poor risk identification ability, and low optimization efficiency of existing optimization methods, significantly enhancing the optimization efficiency of the lining composite structure and the stability of the lining structure.

[0016] The above description is merely an overview of the technical solution of the present invention. In order to better understand the technical means of the present invention and to implement it in accordance with the contents of the specification, and in order to make the above and other objects, features and advantages of the present invention more apparent and understandable, specific embodiments of the present invention are described below. Attached Figure Description

[0017] Figure 1 This is a flowchart illustrating the method for optimizing a lining composite structure using twin simulation, as provided in an embodiment of the present invention.

[0018] Figure 2 This is a schematic diagram of the process for constructing the initial lining risk detection matrix in the lining composite structure optimization method using twin simulation provided in an embodiment of the present invention.

[0019] Figure 3This is a schematic diagram of the structure of the lining composite structure optimization system using twin simulation provided in an embodiment of the present invention.

[0020] Figure labeling: Lining structure simulation module 10, scene feature mining module 20, risk detection module 30, adjustment scheme generation module 40, lining structure optimization module 50. Detailed Implementation

[0021] This invention provides a method and system for optimizing lining composite structures using twin simulation. By introducing a multi-scenario modeling and risk detection mechanism based on twin simulation, it solves the technical problems of insufficient adaptability of optimization results, structural instability risk, and low optimization efficiency in the optimization process of lining composite structures due to the lack of comprehensive risk detection and assessment and the failure to fully incorporate the diverse factors of actual application scenarios in the optimization process of lining composite structures. This achieves the technical effect of realizing the adaptability optimization of lining structures in various application environments, improving the optimization efficiency of lining structures, and enhancing the stability of lining structures.

[0022] Example 1, as Figure 1 As shown, this embodiment of the invention provides a method for optimizing a lining composite structure using twin simulation, the method comprising: Step S100: Perform twin simulation based on the initial design of the lining structure of the target garment to construct an initial lining model.

[0023] Specifically, the target garment refers to the specific garment object that requires lining structure optimization, such as suits, trench coats, and sportswear. The initial lining structure scheme refers to the set of design parameters initially configured in the target garment design, including lining material, thickness, texture direction, and layering structure. The initial lining model refers to the digital model constructed through twin simulation based on the initial lining structure scheme, used for subsequent analysis and optimization.

[0024] The design drawings and physical parameters of the target garment are obtained, and the corresponding initial lining structure is extracted, including material parameters (such as fusible interlining and non-woven interlining), process parameters (such as bonding methods and seam boundaries), and geometric shapes. Then, based on a finite element analysis platform or multiphysics simulation software, a corresponding 3D simulation model is constructed. Through material modeling, mesh generation, and boundary condition settings, dynamic mechanical simulation and stress behavior simulation are achieved to obtain the initial lining model. This initial lining model can respond to external forces, deformations, or stress distributions in digital space, providing a simulation basis for subsequent scene adaptation and risk assessment.

[0025] Step S200: Based on the application scenario factor set, perform lining application scenario feature mining on the target garment to construct multiple simulated application scenarios.

[0026] Furthermore, the application scenario factor set includes multiple application scenario factors, including environmental factors, time factors, mechanical factors, and behavioral factors.

[0027] Specifically, the application scenario factor set refers to the multi-dimensional influencing factors that constitute the actual wearing scenario, covering environmental factors (temperature, humidity), time factors (season, usage duration), mechanical factors (stress type, load intensity), and behavioral factors (human movement, posture changes), etc. The simulation application field refers to the virtual test environment constructed in the simulation system that reflects specific scenario conditions, used to simulate the performance response of the lining composite structure under specific conditions.

[0028] By conducting surveys or analyzing historical databases related to clothing usage scenarios, an application scenario retrieval set is extracted. This set is then combined with information on environment, time, mechanics, and behavior to construct a preliminary factor set, i.e., the application scenario factor set. Subsequently, trend analysis and parameter domain division are performed on multiple application scenario factors to construct multiple simulation dimensions (such as combinations of high temperature-humidity, low temperature-dryness, and high-frequency bending and stretching), generating an initial application scenario set. The initial application scenario set is evaluated for difference using methods such as Euclidean distance and KL divergence. Clustering algorithms are then used to fuse the initial application scenarios, selecting representative characteristic scenarios for simulation modeling. Finally, multiple simulation application fields are built as optimization constraint inputs. For example, if a sportswear lining needs to adapt to spring and summer training scenarios, representative scenarios such as "high temperature and humidity + large bending" and "low temperature + long-term static holding" are extracted by analyzing typical environmental temperatures, humidity, and human movement range, serving as input conditions for constructing the simulation environment.

[0029] Step S300: Introduce a lining risk detection channel, combine the multiple simulation application fields to perform risk detection on the initial lining model, and construct an initial lining risk detection matrix.

[0030] Specifically, the lining risk detection channel refers to an analytical mechanism that integrates multiple risk models (such as deformation, fatigue, and recovery performance) to identify potential problems in the lining structure. The initial lining risk detection matrix is ​​a two-dimensional matrix data structure that records the risk indicators or risk coefficients corresponding to the initial lining model in various simulation application scenarios, which are used to subsequently determine whether the constraints are met.

[0031] The aforementioned simulation scenarios are sequentially applied to the initial lining model, and a complete simulation process is run in each scenario, outputting structural response data such as stress, deformation, and energy dissipation. This simulation data is then input into a pre-constructed lining risk detection channel, which includes a lining deformation risk detection model (determining whether the elastic response capability is exceeded), a lining fatigue risk detection model (predicting performance degradation after multiple loading cycles), and a lining recovery risk detection model (measuring short-term recovery efficiency). The risk detection results for each scenario are output separately and integrated into a unified initial lining risk detection matrix for subsequent decision-making.

[0032] Step S400: If the initial lining risk detection matrix does not meet the lining risk constraints, construct a lining structure optimization guidance space, and adjust the initial lining structure scheme according to the lining structure optimization guidance space to generate a group of lining structure adjustment schemes.

[0033] Furthermore, the lining risk constraints include lining deformation risk constraints, lining fatigue risk constraints, and lining recovery risk constraints.

[0034] Specifically, the lining risk constraint is a preset risk tolerance threshold used to determine whether the structure is safe and acceptable in the simulation scenario. It includes lining deformation risk constraints, fatigue risk constraints, and recovery risk constraints. Lining deformation risk constraints limit the degree to which the lining changes shape under stress or environmental changes, measured by indicators such as deformation amount and strain. Lining fatigue risk constraints limit the degree to which the lining, under repeated stress or long-term use, experiences fatigue failure phenomena such as breakage and loosening due to gradual deterioration of material properties, measured by indicators such as fatigue life and number of cyclic loads. Lining recovery risk constraints limit the ability of the lining to return to its original state after the removal of external forces or environmental influences, characterized by indicators such as recovery rate and residual deformation. The lining structure optimization guidance space refers to the adjustment range and direction of the lining structure parameters determined based on the risk detection results and optimization objectives. The lining structure adjustment scheme group refers to a series of lining structure adjustment schemes generated based on the optimization guidance space.

[0035] If any indicator or risk coefficient in the initial lining risk detection matrix exceeds the corresponding risk constraint, the current structure is considered to have unacceptable risk, meaning the initial lining risk detection matrix does not meet the lining risk constraints. At this point, high-risk areas are identified, and key influencing factors within these areas (such as insufficient interlayer thickness, poor stress transmission paths, etc.) are extracted to establish a set of risk anomaly factors. Then, through data backtracking and historical sample analysis, related structural variables (such as material type, overlapping methods, etc.) are mined, and a variable combination constraint interval is established based on the design boundaries of the target garment, forming a guiding space for lining structure optimization. Within this space, parameter combination searches are performed to generate multiple lining structure adjustment schemes that meet the initial constraints for subsequent optimization.

[0036] Step S500: Based on the lining risk detection channel and the lining risk constraint, perform multi-field joint optimization on the lining structure adjustment scheme group according to the multiple simulation application fields to obtain the lining structure optimization strategy.

[0037] Specifically, the constructed group of lining structure adjustment schemes is run in multiple simulation scenarios, repeating the risk assessment process and generating multiple single-scenario optimization results. These results are then integrated using an intersection method to obtain a set of candidate schemes that simultaneously satisfy constraints from multiple scenarios. To improve the optimization accuracy, a weighted function of multi-dimensional risk indicators for the lining is introduced, assigning different weights to factors such as deformation, fatigue, and recovery to construct a comprehensive risk function. This function is then used to minimize the candidate schemes, ultimately determining the lining structure optimization strategy. This achieves a balance between global consistency and risk minimization in the lining structure schemes, improving the practicality, stability, and deployability of the optimized schemes in diverse scenarios.

[0038] Furthermore, step S200 includes: Step S210: Perform a lining application scenario search based on the target garment to obtain an application scenario search set.

[0039] Step S220: Perform trend analysis on the application scenario retrieval set based on the application scenario factor set, and construct multiple scenario factor parameter domains.

[0040] Step S230: Randomly combine the multiple scenario factor parameter domains to obtain an initial application scenario set.

[0041] Step S240: Perform pairwise difference evaluation on the initial application scenario set to obtain a scenario difference evaluation set, and perform clustering and fusion on the initial application scenario set based on the scenario difference evaluation set to obtain multiple feature application scenarios.

[0042] Step S250: Based on the multiple characteristic application scenarios, build the multiple simulation application fields.

[0043] Specifically, based on the target garment's design attributes, target wearer, intended use, and seasonal suitability, corresponding candidate scenarios are matched from a pre-defined application scenario knowledge base. Design attributes include garment type (e.g., suits, workwear, sportswear), functional type (e.g., warmth, windproof, crispness), applicable gender and age, and daily or special usage needs. Application scenarios highly correlated with the target garment are extracted using a feature tag matching mechanism, forming a preliminary candidate set, denoted as the application scenario retrieval set.

[0044] For all application scenarios in the application scenario retrieval set, data on associated environmental factors (such as temperature and humidity), time factors (such as season and usage duration), mechanical factors (such as activity range and load direction), and behavioral factors (such as frequency of sitting and bending) are extracted. Data analysis tools are used to perform statistical and frequency distribution analysis on these factors, observing their distribution patterns and trends across different application scenarios. Based on the analysis results, a reasonable value range for each factor is determined, i.e., multiple scenario factor parameter domains are constructed. For example, for temperature among environmental factors, its highest, lowest, and common ranges are analyzed across multiple retrieved application scenarios to determine the parameter domain for the temperature factor as a specific interval.

[0045] Based on multiple scenario factor parameter domains, sampling methods such as random sampling, Latin hypercube sampling, or Sobol sequence are used to generate multiple combinations of each scenario factor parameter domain to obtain an initial application scenario set with broad coverage. Each generated scenario combination is a parameter vector representing a potential use case. A scenario difference evaluation matrix is ​​constructed by calculating the Euclidean distance, Manhattan distance, or cosine similarity between each pair of initial application scenario combinations. Taking Euclidean distance as an example, for every two scenarios in the initial application scenario set, the square root of the sum of the squares of their numerical differences across various scenario factors is calculated to obtain their difference values, which are recorded in the scenario difference evaluation matrix. Then, clustering algorithms such as K-means, DBSCAN, or hierarchical clustering are used to fuse and group the scenario data according to the difference values ​​in the scenario difference evaluation set, dividing the initial application scenario set into several clusters. Application scenarios within each cluster have high similarity, while application scenarios between different clusters have significant differences. Representative samples are extracted from each cluster as feature application scenarios. These representative samples are the scenario data closest to the cluster center, representing the average or typical characteristics of that cluster.

[0046] The environmental parameters, time behavior characteristics, and mechanical conditions of each characteristic application scenario are input into the simulation modeling software to establish corresponding boundary conditions and excitation configurations, such as setting temperature field (constant or fluctuating), humidity field (static or alternating), periodic force path (simulating wearing or taking off or movement), friction coefficient, impact load, etc., to form multiple simulation application fields covering typical usage situations, which serve as the input environment for subsequent risk detection of the lining model.

[0047] The following is a complete example process of steps S210 to S250, using "business men's suits" as the target garment: Target garment attributes: Type = Men's suit; Function = Crisp, durable; Target audience = Business men; Scenarios = Daily commuting and meetings. The initial matching application scenario retrieval set is as follows: Scenario 1: Seated meetings, Scenario 2: Subway commuting, Scenario 3: Daily office activities, Scenario 4: Business meals, Scenario 5: Business flights. Factors extracted from the retrieval set and parameter domains established are shown in Table 1: Table 1 - Examples of Scene Factor Parameter Domains Factor Type Parameter Name Parameter Domain Range Environment Temperature (°C) [18, 26] Environment Humidity (%) [35, 75] Time Wearing Duration (h) [2, 10] Mechanics Upper Limb Activity Frequency (times / h) [10, 80] Behavior Standing Frequency (times / h) [1, 12] Behavior Bending Frequency (times / h) [0, 5] Latin hypercube sampling was used to generate an initial set of application scenarios. Random values ​​were selected from the parameter domains of the multiple scenario factors mentioned above, resulting in the initial application scenario set shown in Table 2. Table 2 - Examples of Initial Application Scenarios Scene Number Temperature (°C) Humidity (%) Wearing Duration (h) Upper Limb Activity Frequency Standing Frequency Bending Frequency A 19.5 40 6 20 3 1 B 22.0 60 8 60 8 2 C 24.5 50 4 40 2 1 D 21.0 35 10 75 10 4 E 26.0 70 5 50 5 0 A difference matrix was constructed based on Euclidean distance and clustered into two classes. Representative scenarios were extracted for each class: Feature Scenario F1 (representing low activity, stable environment) = Scenario A; Feature Scenario F2 (representing high-frequency movement, humid environment) = Scenario B. Simulation application fields were constructed based on these two feature application scenarios: Simulation Scenario F1 configuration: Temperature = 19.5℃, Humidity = 40%, Action period = 6h, Low-frequency upper limb movement, slight standing, Simulated loading of low-amplitude sinusoidal reciprocating tension, 3mm displacement per cycle, 10 cycles / h, Constant temperature. Simulation Scenario F2 configuration: Temperature = 22℃, Humidity = 60%, Action period = 8h, High-frequency upper limb movement, frequent standing, Loading of random disturbance stress, Periodic buckling, Humid-thermal coupled environmental field application.

[0048] Furthermore, such as Figure 2 As shown, step S300 includes: Step S310: Based on the multiple simulation application scenarios, simulate the initial lining model to obtain multiple sets of lining application simulation data.

[0049] Step S320: Input the multiple sets of lining application simulation data into the lining risk detection channel to obtain multiple lining risk detection results.

[0050] Step S330: Based on the multiple lining risk detection results, construct multiple application risk detection matrices, and integrate the multiple application risk detection matrices to generate the initial lining risk detection matrix.

[0051] Specifically, the aforementioned multiple simulation application scenarios are sequentially loaded into the initial lining model, and simulation software is run to solve the simulations. For each simulation application scenario, key lining performance indicators such as stress distribution, strain concentration, fatigue accumulation, deformation response, and structural recovery capacity are recorded during the simulation period, forming a set of lining application simulation data. This lining application simulation data is stored in time series or statistical form for easy subsequent unified risk detection.

[0052] Multiple sets of lining application simulation data are input into a constructed lining risk detection channel. This channel, composed of a trained multi-task deep neural network or an expert evaluation system incorporating logical rules, is capable of identifying the risk category and level of the lining structure in specific application scenarios. The detection content includes, but is not limited to: lining deformation risk (e.g., permanent deformation due to excessive shear or bending stress); lining fatigue risk (e.g., strength reduction under frequent load cycles); and lining recovery risk (e.g., shape memory failure or recovery delay under high temperature or high humidity conditions). Finally, multiple lining risk detection results are obtained, with one result corresponding to each simulation field, including quantitative values ​​for various risk indicators.

[0053] Based on multiple lining risk detection results, an application risk detection matrix is ​​generated for each simulation scenario. This matrix represents the numerical results of various risk indicators in a row-column format, for example, recording each lining risk as a two-dimensional matrix of [scenario number] × [risk category]. Based on these multiple application risk detection matrices, a matrix integration method using weighted averaging, maximum risk focusing, or scenario probability distribution is employed to obtain a unified initial lining risk detection matrix. This initial lining risk detection matrix serves as the overall performance evaluation of the current lining structure's risk response capability under comprehensive application scenarios, and is used in subsequent structural adjustment and optimization steps.

[0054] Furthermore, step S320 includes: Step S321: The lining risk detection channel includes a lining deformation risk detection model, a lining fatigue risk detection model, and a lining recovery risk detection model.

[0055] Step S322: Extract the first set of lining application simulation data based on the multiple sets of lining application simulation data, and perform feature extraction on the first set of lining application simulation data to obtain the first lining deformation simulation data, the first lining fatigue simulation data, and the first lining recovery simulation data.

[0056] Step S323: Input the first lining deformation simulation data into the lining deformation risk detection model to obtain the first lining deformation risk coefficient.

[0057] Step S324: Input the first lining fatigue simulation data into the lining fatigue risk detection model to obtain the first lining fatigue risk coefficient.

[0058] Step S325: Input the first lining fabric recovery simulation data into the lining fabric recovery risk detection model to obtain the first lining fabric recovery risk coefficient. Combine the first lining fabric deformation risk coefficient and the first lining fabric fatigue risk coefficient to generate the first lining fabric risk detection result, and add the first lining fabric risk detection result to the plurality of lining fabric risk detection results.

[0059] Specifically, the lining risk detection channel includes three models: lining deformation risk detection model, lining fatigue risk detection model, and lining recovery risk detection model. The lining deformation risk detection model assesses the tendency of the lining to undergo permanent deformation under complex loads and bending conditions; the lining fatigue risk detection model identifies the trend of structural performance degradation after multiple cyclic loadings; and the lining recovery risk detection model evaluates whether the recovery capability of the lining material meets the standards after typical deformation or environmental disturbances. All three models are built based on machine learning algorithms (such as support vector machines, convolutional neural networks, and XGBoost), trained using supervised learning with historical simulation data and experimental datasets, and output risk coefficient values ​​corresponding to various risk indicators. A model training example is shown below: The dataset includes, but is not limited to, finite element simulation and measured data of various lining materials (such as nonwoven fabrics, composite linings, and thermally bonded linings) under different load conditions (such as tension, compression, bending, and cyclic loading). Each dataset contains the following fields: stress-strain curve, residual deformation, springback rate, strain energy density, number of load cycles, fatigue crack length variation, recovery time, etc., with corresponding risk levels manually labeled as training tags. The total dataset contains no fewer than 3000 datasets, with independent training and validation sets constructed for each risk model, in a ratio of 8:2. Dimensionality reduction of continuous variable features in the simulation data is performed using Z-score normalization and principal component analysis (PCA). To capture the time-series response characteristics of the materials, a sliding window extraction (window size of 10 frames, step size of 1 frame) is further applied to some features (such as load stress fluctuation sequences) to obtain nested sequence features for use in subsequent deep learning models.

[0060] The lining deformation risk detection model adopts a hybrid deep learning structure based on CNN-BiLSTM: the front end uses 3 layers of convolution to extract local stress field distribution image features, and the back end connects to a bidirectional long short-term memory network to capture the deformation evolution trend, outputting the deformation risk level (a continuous value of 0-1). The loss function is mean squared error (MSE), the optimizer is Adam, the initial learning rate is 0.001, the number of training epochs is 100, and the early stopping tolerance is 5 epochs.

[0061] The fatigue risk detection model for the lining fabric employs a traditional machine learning method based on Support Vector Regression (SVR). The input consists of five dimensions: cycle number, maximum stress, stress amplitude, residual energy, and crack growth rate. The output is a fatigue risk score. The kernel function used is the Radial Basis Function (RBF), with a penalty coefficient C of 10 and γ of 0.1. Five-fold cross-validation is used to adjust the parameters.

[0062] The lining recovery risk detection model employs an XGBoost-based gradient boosting regression tree model. The input consists of 8 features: maximum strain, recovery rate, deformation recovery time, and the difference before and after environmental disturbance. The output is the recovery risk level. The model tree depth is set to 6, the learning rate to 0.05, the training epochs to 500, and the early stopping mechanism to be triggered in 10 epochs. Minimizing the mean squared error of the validation set is used as the parameter tuning metric.

[0063] The first set of lining application simulation data is extracted from multiple sets of lining application simulation data. Here, "first set" refers to any one of the multiple sets of lining application simulation data; it can be extracted in the storage order or randomly. Feature extraction is performed on the first set of lining application simulation data to obtain the first lining deformation simulation data, the first lining fatigue simulation data, and the first lining recovery simulation data. Specifically, this includes deriving force-displacement curves, stress-strain fields, energy dissipation spectra, springback rate, and residual deformation from the first simulation application field, and then selecting physical feature quantities related to deformation trend, fatigue accumulation, and recovery performance. The feature vectors are compressed and structured using methods such as normalization, principal component analysis, or convolutional feature encoding to meet the input requirements of subsequent models.

[0064] The deformation simulation data of the first lining fabric is input into the lining fabric deformation risk detection model to obtain the first lining fabric deformation risk coefficient. This first lining fabric deformation risk coefficient reflects the probability and degree of inelastic deformation or material yielding of the lining fabric under the simulation scenario, and its value ranges from 0 to 1. The closer the value is to 1, the higher the deformation risk. The fatigue simulation data of the first lining fabric is input into the lining fabric fatigue risk detection model to obtain the first lining fabric fatigue risk coefficient. This first lining fabric fatigue risk coefficient quantifies the possibility of material performance deterioration and damage of the lining fabric under repeated loading, and its value usually ranges from 0 to 1. The closer the value is to 1, the higher the fatigue risk. The recovery simulation data of the first lining fabric is input into the lining fabric recovery risk detection model to obtain the first lining fabric recovery risk coefficient. This first lining fabric recovery risk coefficient comprehensively considers the degree of recovery of the lining fabric after temperature and humidity changes or instantaneous deformation, and its value ranges from 0 to 1. The closer the value is to 1, the higher the recovery risk. Combining the aforementioned first lining fabric deformation risk coefficient, first lining fabric fatigue risk coefficient, and first lining fabric recovery risk coefficient, the first lining fabric risk detection result is generated. The risk detection result of the first lining fabric is represented in the form of a three-dimensional vector or a risk assessment table and added to the current multiple lining fabric risk detection results for subsequent integrated analysis.

[0065] Furthermore, step S400 includes: Step S410: If the initial lining risk detection matrix does not meet the lining risk constraint, generate an abnormal lining risk area.

[0066] Step S420: Identify the attributes of the lining risk anomaly area to obtain the lining risk anomaly factor.

[0067] Step S430: Based on the lining risk anomaly factor, perform correlation evaluation on the lining structure variable set of the initial lining structure scheme to obtain a risk correlation structure variable set that meets the predetermined correlation degree.

[0068] Step S440: Based on the target clothing, retrieve the value records of the risk-related structural variable set and establish multiple value constraints for related variables.

[0069] Step S450: Combine the parameter values ​​of the risk-related structural variable set according to the multiple associated variable value constraints to generate the lining structure optimization guidance space.

[0070] Specifically, if the initial lining risk detection matrix does not meet the lining risk constraints, the risk coefficients in the initial lining risk detection matrix are compared with the set risk thresholds. For risk coefficients that do not meet the threshold conditions, their position index in the matrix is ​​extracted, and they are constructed and recorded in the lining risk anomaly area.

[0071] The attributes of the lining risk anomaly areas are identified by extracting the associated risk indicator types based on the simulation scenario and risk type corresponding to their matrix index, forming a set of lining risk anomaly factors. The lining risk anomaly factors are used to characterize the source category of the risk anomaly area. For example, if an anomaly location comes from "fatigue risk coefficient exceeding the limit" under application scenario A, then the anomaly factor is "fatigue risk-A".

[0072] For each lining risk anomaly factor, a set of lining structure variables is extracted from the initial lining structure design. These variables are adjustable parameters describing the lining structure, including but not limited to the number of lining layers, their arrangement, adhesive strength, elastic modulus, and thickness. A model is then constructed to correlate the risk factor with these structural variables. Information gain, Pearson correlation coefficient, and grey relational analysis are used to evaluate their correlation. Variables that meet a predetermined correlation threshold (e.g., a correlation coefficient greater than 0.7) with the lining risk anomaly factor are selected to form a risk-related structural variable set.

[0073] Based on the historical design database or knowledge graph of the target garment, the value records of the risk-related structural variables are retrieved to determine the reasonable value range of each variable. Multiple value constraints for related variables are established, such as: the number of lining layers should be [2, 4], and the adhesive strength should not be less than 0.85 MPa. Based on these multiple value constraints, experimental design methods (such as orthogonal experiments, Latin hypercube sampling) or optimization algorithms (such as genetic algorithms, particle swarm optimization) are used to search for and generate parameter combinations, forming a lining structure optimization guidance space. Each combination in this lining structure optimization guidance space satisfies historical feasibility and has the potential to mitigate identified risks, providing a basis for candidate structural schemes for subsequent optimization.

[0074] Furthermore, step S500 includes: Step S510: Based on the lining risk detection channel and the lining risk constraint, perform optimization analysis on the lining structure adjustment scheme group according to the multiple simulation application fields, and generate multiple single-field lining structure optimization groups.

[0075] Step S520: Perform intersection analysis based on the multiple single-field lining structure optimization groups to obtain a multi-field joint lining structure optimization group.

[0076] Step S530: Based on the multidimensional risk indicators of the lining, weights are allocated to construct a comprehensive risk analysis function for the lining. The multidimensional risk indicators of the lining include the lining deformation risk, the lining fatigue risk, and the lining recovery risk.

[0077] Step S540: Based on the comprehensive risk analysis function of the lining fabric, perform comprehensive risk minimization optimization on the multi-field joint lining fabric structure optimization group to generate the lining fabric structure optimization strategy.

[0078] Specifically, based on the lining risk detection channel and lining risk constraints, for the lining structure adjustment scheme group generated in step S400, simulation and risk assessment are performed in each simulation application field to obtain structural adjustment performance feedback under that scenario. Then, according to the set optimization criteria (such as minimizing risk coefficients, material usage balance, structural rationality, etc.), a heuristic search algorithm (such as genetic algorithm, particle swarm optimization, simulated annealing, etc.) is used to generate a single-field lining structure optimization group for that simulation scenario. This single-field lining structure optimization group is a set of lining structure adjustment schemes that meet the lining risk constraints, selected through optimization analysis in a single simulation application field.

[0079] Intersection analysis was performed on the single-field lining structure optimization groups obtained from all simulation application fields. The lining structure adjustment schemes that performed well in most or all simulation application fields were retained, and the solution set members that performed well only in specific scenarios but had serious adaptability defects were eliminated, thus forming a multi-field joint lining structure optimization group.

[0080] Based on the actual usage requirements and design objectives of the lining, the weights of each risk indicator are determined. Then, a weighted summation method is used to construct a comprehensive risk analysis function for the lining, achieving a unified evaluation of multi-dimensional risk indicators. Preferably, the comprehensive risk analysis function for the lining is: R total =ω d *R d +ω f *R f +ω r *R r Where: R total The overall risk value for the lining structure scheme; ω d R is the weighting factor for the risk coefficient of lining deformation. d ω represents the risk factor for lining deformation. f R is the weighting factor for the fatigue risk coefficient of the lining fabric. f ω represents the fatigue risk factor of the lining fabric. r R is the weighting factor for the risk coefficient of lining recovery. r ω represents the risk factor for lining restoration. d +ω f +ω r =1. For example, if the risk of deformation has a significant impact on the appearance of the clothing, a higher weight should be assigned. Weight allocation can be done using methods such as expert experience, the Analytic Hierarchy Process (AHP), or the entropy weight method. The above weighting factors can be set according to the target clothing type, usage scenario, and design preferences, using methods such as expert experience, the Analytic Hierarchy Process (AHP), or the entropy weight method. For example, for functional sportswear, a higher weight should be given to fatigue risk (e.g., ω). d =0.2, ω f=0.5, ω r =0.3), while for formal wear, the focus is more on the risk of deformation (e.g., ω). d =0.6, ω f =0.2, ω r =0.2).

[0081] Using the comprehensive risk analysis function of the lining fabric as the fitness function, optimization algorithms (such as genetic algorithms and particle swarm optimization algorithms) are employed to evaluate and screen all structural adjustment schemes in the multi-field joint lining fabric structure optimization group. The principle of minimizing the comprehensive risk of the lining fabric is adopted to find the scheme that minimizes the comprehensive risk. The final determined scheme is the lining fabric structure optimization strategy.

[0082] Furthermore, step S510 includes: Step S511: Extract the first simulation application field based on the multiple simulation application fields.

[0083] Step S512: Extract the Pth lining structure adjustment scheme according to the group of lining structure adjustment schemes, where P is a positive integer.

[0084] Step S513: Apply the simulation to the Pth lining structure adjustment scheme according to the first simulation application field to obtain the simulation data of the Pth lining application.

[0085] Step S514: Input the simulated data of the Pth lining fabric into the lining fabric risk detection channel to obtain the risk detection result of the Pth lining fabric.

[0086] Step S515: If the risk detection result of the Pth lining fabric meets the lining fabric risk constraint, add the Pth lining fabric structure adjustment scheme to the first single-field lining fabric structure optimization group.

[0087] Step S516: Based on the lining risk detection channel and the lining risk constraint, continue to iteratively optimize the lining structure adjustment scheme group according to the first simulation application field, construct the first single-field lining structure optimization group, and add the first single-field lining structure optimization group to the multiple single-field lining structure optimization groups.

[0088] Specifically, the first simulation application field is extracted from multiple simulation application fields, that is, the simulation scenario currently being optimized and analyzed is selected from the multiple simulation environments generated in step S250. The first simulation application field can be any simulation application field, including specific wearing conditions such as "continuous wearing under high humidity and high temperature environment", "stretching action under low temperature static state", and "intermittent use under high intensity bending behavior".

[0089] Extract the P-th lining structure adjustment scheme from the group of lining structure adjustment schemes, that is, extract each scheme sequentially from the group of lining structure adjustment schemes for analysis. Here, P is a positive integer used to identify the scheme number being processed. The lining structure adjustment scheme is a combination of structural variations generated in step S400, with different material parameters, layer arrangement, bonding strategies, or geometric shapes, etc.

[0090] The adjustment scheme of the P-th lining structure is loaded into the first simulation application field for virtual simulation operation. Using a finite element simulation engine or a data-driven simulation system, the dynamic response behavior of the lining structure under the first simulation application field condition is modeled and solved, thereby obtaining the simulation data of the P-th lining application, including strain distribution, energy dissipation, recovery path, fatigue crack propagation trajectory, etc.

[0091] The simulation data of the Pth lining fabric is input into the lining fabric risk detection channel. Using the trained lining fabric deformation risk detection model, lining fabric fatigue risk detection model, and lining fabric recovery risk detection model, the risk coefficient of the Pth lining fabric structure adjustment scheme under this first simulation application field is calculated respectively. Finally, the risk detection result of the Pth lining fabric is output. The risk detection result of the Pth lining fabric includes: RP={R P,d R P,f R P,r}where R P,d R P,f R P,r Let F represent the lining deformation risk coefficient, lining fatigue risk coefficient, and lining recovery risk coefficient of the Pth lining structure adjustment scheme, respectively.

[0092] The risk detection result of the Pth lining fabric is compared with the preset lining fabric risk constraints. If the three risk coefficients are all lower than their corresponding risk constraint thresholds, that is, the risk detection result of the Pth lining fabric satisfies the lining fabric risk constraints, then the Pth lining fabric structure adjustment scheme is added to the first single-field lining fabric structure optimization group as the preliminary feasible structural solution under the first simulation application field.

[0093] Based on the lining risk detection channel and risk constraints, the lining structure adjustment scheme group is iteratively optimized in the current simulation application field. The next lining structure adjustment scheme is extracted from the lining structure adjustment scheme group for simulation and evaluation. The structural solution space is continuously expanded to complete the construction of the first single-field lining structure optimization group. This first single-field lining structure optimization group is added to multiple single-field lining structure optimization groups for subsequent step S520 to perform intersection analysis.

[0094] In summary, the method for optimizing the lining composite structure using twin simulation provided in this invention has the following beneficial effects: By constructing an initial lining model and integrating various scenario factors such as environment, time, mechanics, and behavior in garment use, this study explores typical lining application scenarios and conducts risk perception and optimization guidance in multiple simulated application environments. First, a machine learning model is used to perform multi-dimensional detection of deformation, fatigue, and recovery risks of the lining under various application scenarios, constructing a unified risk detection matrix to achieve accurate evaluation of the lining structural performance. Then, after detecting areas that do not meet expected risk constraints, a parameter-constrained guidance space is established by identifying the correlation between risk factors and structural variables, significantly compressing the dimensionality of optimization variables and improving solution efficiency. Next, for each characteristic scenario, multiple sets of structural adjustment schemes are simulated and iterated, constructing multiple single-field structural solution sets. These are then integrated into a joint optimization group through intersection analysis to ensure the universality and robustness of the optimization results across multiple scenarios. Finally, by constructing a comprehensive lining risk analysis function that integrates deformation, fatigue, and recovery indicators, the multi-field structural solution sets are driven to perform global minimization optimization to obtain an lining structure optimization strategy that satisfies the minimum risk criterion.

[0095] Overall, the embodiments of the present invention achieve full-process structural performance improvement from simulation modeling and risk identification to guided optimization, effectively improving the adaptability, safety and stability of lining structure design under different usage scenarios, solving the problems of insufficient response to actual scenarios, poor risk identification ability and low optimization efficiency of existing optimization methods, and significantly enhancing the optimization efficiency of lining composite structures and the stability of lining structures.

[0096] Example 2, as Figure 3 As shown, based on the same inventive concept as in Embodiment 1, this embodiment of the invention provides a system for optimizing lining composite structures using twin simulation, the system comprising: The lining structure simulation module 10 is used to perform twin simulation based on the initial lining structure scheme of the target garment and construct an initial lining model.

[0097] The scene feature mining module 20 is used to mine the lining application scene features of the target garment based on the application scene factor set, and to construct multiple simulated application scenes.

[0098] The risk detection module 30 is used to introduce the lining risk detection channel, combine the multiple simulation application fields to perform risk detection on the initial lining model, and construct the initial lining risk detection matrix.

[0099] The adjustment scheme generation module 40 is used to construct a lining structure optimization guidance space if the initial lining risk detection matrix does not meet the lining risk constraints, and adjust the initial lining structure scheme according to the lining structure optimization guidance space to generate a group of lining structure adjustment schemes.

[0100] The lining structure optimization module 50 is used to perform multi-field joint optimization of the lining structure adjustment scheme group based on the lining risk detection channel and the lining risk constraints, according to the multiple simulation application fields, to obtain the lining structure optimization strategy.

[0101] Furthermore, the application scenario factor set includes multiple application scenario factors, including environmental factors, time factors, mechanical factors, and behavioral factors.

[0102] Furthermore, the lining risk constraints include lining deformation risk constraints, lining fatigue risk constraints, and lining recovery risk constraints.

[0103] Furthermore, in this embodiment of the invention, the scene feature mining module 20 is also used to perform the following steps: Application scenarios for lining are retrieved based on the target garment to obtain an application scenario retrieval set; trend analysis is performed on the application scenario retrieval set based on the application scenario factor set to construct multiple scenario factor parameter domains; random value combinations are performed on the multiple scenario factor parameter domains to obtain an initial application scenario set; pairwise difference evaluation is performed on the initial application scenario set to obtain a scenario difference evaluation set, and clustering and fusion are performed on the initial application scenario set based on the scenario difference evaluation set to obtain multiple characteristic application scenarios; multiple simulation application fields are constructed based on the multiple characteristic application scenarios.

[0104] Furthermore, in this embodiment of the invention, the risk detection module 30 is also used to perform the following steps: Based on the multiple simulation application scenarios, the initial lining model is simulated and applied to obtain multiple sets of lining application simulation data; the multiple sets of lining application simulation data are input into the lining risk detection channel to obtain multiple lining risk detection results; based on the multiple lining risk detection results, multiple application risk detection matrices are constructed, and the multiple application risk detection matrices are integrated to generate the initial lining risk detection matrix.

[0105] Furthermore, in this embodiment of the invention, the risk detection module 30 is also used to perform the following steps: The lining risk detection channel includes a lining deformation risk detection model, a lining fatigue risk detection model, and a lining recovery risk detection model. Based on the multiple sets of lining application simulation data, a first set of lining application simulation data is extracted, and features are extracted from the first set of lining application simulation data to obtain first lining deformation simulation data, first lining fatigue simulation data, and first lining recovery simulation data. The first lining deformation simulation data is input into the lining deformation risk detection model to obtain a first lining deformation risk coefficient. The first lining fatigue simulation data is input into the lining fatigue risk detection model to obtain a first lining fatigue risk coefficient. The first lining recovery simulation data is input into the lining recovery risk detection model to obtain a first lining recovery risk coefficient. Combining the first lining deformation risk coefficient and the first lining fatigue risk coefficient, a first lining risk detection result is generated, and the first lining risk detection result is added to the multiple lining risk detection results.

[0106] Furthermore, in this embodiment of the invention, the adjustment scheme generation module 40 is also used to perform the following steps: If the initial lining risk detection matrix does not meet the lining risk constraints, an lining risk anomaly region is generated; attribute identification is performed on the lining risk anomaly region to obtain lining risk anomaly factors; the lining risk anomaly factors are used to perform correlation evaluation on the lining structure variable set of the initial lining structure scheme to obtain a risk correlation structure variable set that meets a predetermined correlation degree; the value records of the risk correlation structure variable set are retrieved according to the target garment to establish multiple correlation variable value constraints; the parameter values ​​of the risk correlation structure variable set are combined according to the multiple correlation variable value constraints to generate the lining structure optimization guidance space.

[0107] Furthermore, in this embodiment of the invention, the lining structure optimization module 50 is also used to perform the following steps: Based on the lining risk detection channel and the lining risk constraints, optimization analysis is performed on the lining structure adjustment scheme group according to the multiple simulation application fields, generating multiple single-field lining structure optimization groups; intersection analysis is performed on the multiple single-field lining structure optimization groups to obtain a multi-field joint lining structure optimization group; weight allocation is performed according to the lining multi-dimensional risk indicators to construct a lining comprehensive risk analysis function, whereby the lining multi-dimensional risk indicators include lining deformation risk, lining fatigue risk, and lining recovery risk; the lining comprehensive risk is minimized in the multi-field joint lining structure optimization group based on the lining comprehensive risk analysis function to generate the lining structure optimization strategy.

[0108] Furthermore, in this embodiment of the invention, the lining structure optimization module 50 is also used to perform the following steps: The first simulation application field is extracted based on the multiple simulation application fields; the Pth lining structure adjustment scheme is extracted based on the lining structure adjustment scheme group, where P is a positive integer; the Pth lining structure adjustment scheme is simulated and applied based on the first simulation application field to obtain the Pth lining application simulation data; the Pth lining application simulation data is input into the lining risk detection channel to obtain the Pth lining risk detection result; if the Pth lining risk detection result satisfies the lining risk constraint, the Pth lining structure adjustment scheme is added to the first single-field lining structure optimization group; based on the lining risk detection channel and the lining risk constraint, the lining structure adjustment scheme group is iteratively optimized based on the first simulation application field to construct the first single-field lining structure optimization group, and the first single-field lining structure optimization group is added to the multiple single-field lining structure optimization groups.

[0109] Through the foregoing detailed description of the method for optimizing the lining composite structure using twin simulation, those skilled in the art can clearly understand that the lining composite structure optimization system using twin simulation in this embodiment, compared with the system disclosed in Embodiment 2, has corresponding functional modules and beneficial effects as it corresponds to the method disclosed in Embodiment 1. For relevant details, please refer to the description in the method section.

[0110] The above description of the disclosed embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for optimizing lining composite structures using twin simulation, characterized in that, The method includes: Based on the initial design of the lining structure of the target garment, a twin simulation is performed to construct an initial lining model; Based on the application scenario factor set, the target garment's lining application scenario features are mined to construct multiple simulated application scenarios; A risk detection channel for the lining fabric is introduced, and the initial lining fabric model is subjected to risk detection in combination with the multiple simulation application fields to construct an initial lining fabric risk detection matrix. If the initial lining risk detection matrix does not meet the lining risk constraints, a lining structure optimization guidance space is constructed, and the initial lining structure scheme is adjusted according to the lining structure optimization guidance space to generate a group of lining structure adjustment schemes. Based on the lining risk detection channel and the lining risk constraints, the lining structure adjustment scheme group is jointly optimized in multiple simulation application fields to obtain the lining structure optimization strategy.

2. The method for optimizing the lining composite structure using twin simulation as described in claim 1, characterized in that, Based on the application scenario factor set, the lining application scenario features of the target garment are mined, and multiple simulation application scenarios are constructed, including: Based on the target garment, an application scenario search for the lining is performed to obtain an application scenario search set; Based on the application scenario factor set, perform trend analysis on the application scenario retrieval set to construct multiple scenario factor parameter domains; An initial set of application scenarios is obtained by randomly combining the parameters of the multiple scenario factors. The initial application scenario set is subjected to pairwise difference evaluation to obtain a scenario difference evaluation set, and the initial application scenario set is clustered and fused according to the scenario difference evaluation set to obtain multiple feature application scenarios. Based on the aforementioned multiple characteristic application scenarios, the aforementioned multiple simulation application fields are constructed.

3. The method for optimizing the lining composite structure using twin simulation as described in claim 1, characterized in that, A risk detection channel for the lining fabric is introduced, and the initial lining fabric model is subjected to risk detection in conjunction with the multiple simulation application scenarios to construct an initial lining fabric risk detection matrix, including: Based on the multiple simulation application scenarios, the initial lining model is simulated and applied to obtain multiple sets of lining application simulation data. The simulated data of the multiple sets of lining fabrics are input into the lining fabric risk detection channel to obtain multiple lining fabric risk detection results. Based on the multiple lining risk detection results, multiple application risk detection matrices are constructed, and the multiple application risk detection matrices are integrated to generate the initial lining risk detection matrix.

4. The method for optimizing the lining composite structure using twin simulation as described in claim 3, characterized in that, The simulated data of the multiple sets of lining fabrics are input into the lining fabric risk detection channel to obtain multiple lining fabric risk detection results, including: The lining risk detection channel includes a lining deformation risk detection model, a lining fatigue risk detection model, and a lining recovery risk detection model. Based on the multiple sets of lining application simulation data, the first set of lining application simulation data is extracted, and feature extraction is performed on the first set of lining application simulation data to obtain the first lining deformation simulation data, the first lining fatigue simulation data, and the first lining recovery simulation data. Input the first lining deformation simulation data into the lining deformation risk detection model to obtain the first lining deformation risk coefficient; The fatigue simulation data of the first lining fabric is input into the fatigue risk detection model of the lining fabric to obtain the fatigue risk coefficient of the first lining fabric. The first lining fabric recovery simulation data is input into the lining fabric recovery risk detection model to obtain the first lining fabric recovery risk coefficient. The first lining fabric deformation risk coefficient and the first lining fabric fatigue risk coefficient are combined to generate the first lining fabric risk detection result. The first lining fabric risk detection result is then added to the plurality of lining fabric risk detection results.

5. The method for optimizing the lining composite structure using twin simulation as described in claim 1, characterized in that, If the initial lining risk detection matrix does not meet the lining risk constraints, a lining structure optimization guidance space is constructed, including: If the initial lining risk detection matrix does not meet the lining risk constraints, an abnormal lining risk zone is generated. The risk anomaly area of ​​the lining fabric is identified by attribute identification to obtain the risk anomaly factor of the lining fabric; Based on the lining risk anomaly factor, the lining structure variable set of the initial lining structure scheme is evaluated to obtain a risk-related structure variable set that meets the predetermined correlation degree. Based on the target clothing, the value records of the risk-related structural variable set are retrieved, and multiple value constraints of the related variables are established. Based on the constraints of the multiple associated variable values, the risk-related structural variable set is combined with parameter values ​​to generate the lining structure optimization guidance space.

6. The method for optimizing the lining composite structure using twin simulation as described in claim 1, characterized in that, Based on the lining risk detection channel and the lining risk constraints, a multi-field joint optimization is performed on the lining structure adjustment scheme group according to the multiple simulation application fields to obtain the lining structure optimization strategy, including: Based on the lining risk detection channel and the lining risk constraint, the optimization analysis of the lining structure adjustment scheme group is performed according to the multiple simulation application fields to generate multiple single-field lining structure optimization groups. Based on the intersection analysis of the multiple single-field lining structure optimization groups, a multi-field joint lining structure optimization group is obtained. Based on the weight allocation of the lining multidimensional risk indicators, a comprehensive risk analysis function for the lining is constructed. The lining multidimensional risk indicators include lining deformation risk, lining fatigue risk, and lining recovery risk. Based on the comprehensive risk analysis function of the lining fabric, the comprehensive risk of the lining fabric is minimized in the multi-field joint lining fabric structure optimization group, and the lining fabric structure optimization strategy is generated.

7. The method for optimizing the lining composite structure using twin simulation as described in claim 6, characterized in that, Based on the lining risk detection channel and the lining risk constraints, optimization analysis is performed on the lining structure adjustment scheme group according to the multiple simulation application fields, generating multiple single-field lining structure optimization groups, including: The first simulation application field is extracted based on the multiple simulation application fields; Based on the group of lining structure adjustment schemes, the Pth lining structure adjustment scheme is extracted, where P is a positive integer; The Pth lining structure adjustment scheme is simulated and applied according to the first simulation application field to obtain the Pth lining application simulation data. The simulated data of the Pth lining fabric is input into the lining fabric risk detection channel to obtain the risk detection result of the Pth lining fabric. If the risk detection result of the Pth lining fabric meets the lining fabric risk constraint, the Pth lining fabric structure adjustment scheme will be added to the first single-field lining fabric structure optimization group. Based on the lining risk detection channel and the lining risk constraints, the lining structure adjustment scheme group is iteratively optimized according to the first simulation application field to construct the first single-field lining structure optimization group, and the first single-field lining structure optimization group is added to the multiple single-field lining structure optimization groups.

8. The method for optimizing the lining composite structure using twin simulation as described in claim 1, characterized in that, The application scenario factor set includes multiple application scenario factors, including environmental factors, time factors, mechanical factors, and behavioral factors.

9. The method for optimizing the lining composite structure using twin simulation as described in claim 1, characterized in that, The lining risk constraints include lining deformation risk constraints, lining fatigue risk constraints, and lining recovery risk constraints.

10. A system for optimizing lining composite structures using twin simulation, characterized in that, The system is used to execute the method for optimizing the lining composite structure using twin simulation as described in any one of claims 1-9, including: The lining structure simulation module is used to perform twin simulations based on the initial lining structure scheme of the target garment and build an initial lining model. The scene feature mining module is used to mine the lining application scene features of the target garment based on the application scene factor set, and to construct multiple simulated application scenes. The risk detection module is used to introduce the lining risk detection channel, combine the multiple simulation application fields to perform risk detection on the initial lining model, and construct the initial lining risk detection matrix. The adjustment scheme generation module is used to construct a lining structure optimization guidance space if the initial lining risk detection matrix does not meet the lining risk constraints, and adjust the initial lining structure scheme according to the lining structure optimization guidance space to generate a group of lining structure adjustment schemes. The lining structure optimization module is used to perform multi-field joint optimization of the lining structure adjustment scheme group based on the lining risk detection channel and the lining risk constraints, according to the multiple simulation application fields, to obtain the lining structure optimization strategy.

Images