Spandex wrap yarn production intelligent debugging auxiliary decision-making system based on data fusion

Abstract

The invention relates to the technical field of decision verification, in particular to a spandex wrap yarn production intelligent debugging auxiliary decision system based on data fusion, comprising a multi-scale fusion unit used for collecting microscopic data, mesoscopic data and macroscopic data in production to obtain multi-scale data; obtaining a fusion weight based on production constraint requirements, and generating multi-scale fusion data; the multi-target collaborative decision-making unit is used for extracting a multi-scale fusion feature based on the multi-scale fusion data, establishing a multi-scale target function comprising the multi-scale fusion feature and obtaining a first debugging scheme; the scheme optimization unit is used for constructing a multi-time scale analysis mechanism of short-term online parameter fine tuning, medium-term process mode recognition and long-term knowledge graph evolution, and analyzing the first debugging scheme to generate a second debugging scheme; and the optimal scheme verification unit is used for establishing a digital simulation model to verify and adjust the risk assessment of the second debugging scheme so as to generate an optimal debugging scheme.

Description

Technical Field

[0001] The present invention relates to the technical field of decision verification, and in particular to an intelligent machine adjustment auxiliary decision system for spandex covered yarn production based on data fusion. Background Art

[0002] Spandex coated yarn is an important elastic textile material. Its production process involves the composite processing of spandex filaments and coated fibers. It is a complex system engineering problem with multi-parameter coupling, and computer-aided design and simulation optimization are required to determine the optimal combination of process parameters.

[0003] Existing computer-aided design (CAD) methods are primarily based on single mathematical models or simplified physical models, guiding parameter selection by establishing mappings between parameters and quality indicators. However, these methods are inadequate when dealing with multi-parameter coupling problems, failing to fully consider the interactions and constraints between parameters. For mechanical parameter variable design, existing CAD methods lack effective multivariate coupling analysis capabilities.

[0004] Spandex-coated yarn production involves multiple mechanical parameters, including tension, speed, temperature, and angle. These parameters are subject to complex nonlinear relationships and constraints. Traditional design methods, often employing empirical or simplified linear models, fail to accurately describe the coupling mechanisms between these parameters, leading to deviations from the optimal design solution. Furthermore, existing design tools lack the ability to quantitatively analyze parameter sensitivity, making it difficult to identify key design variables and optimization directions.

[0005] Therefore, an intelligent machine adjustment auxiliary decision-making system for spandex covered yarn production based on data fusion is proposed. Summary of the Invention

[0006] The purpose of the present invention is to provide an intelligent machine adjustment auxiliary decision-making system for spandex covered yarn production based on data fusion, including a multi-scale fusion unit for collecting micro data, meso data and macro data in production to obtain multi-scale data; obtaining fusion weights based on production constraint requirements to generate multi-scale fusion data; a multi-objective collaborative decision-making unit for extracting multi-scale fusion features based on multi-scale fusion data, establishing a multi-scale objective function including multi-scale fusion features, and obtaining a first machine adjustment plan; a plan optimization unit for constructing a multi-time scale analysis mechanism for short-term online parameter fine-tuning, mid-term process pattern recognition and long-term knowledge graph evolution, analyzing the first machine adjustment plan, and generating a second machine adjustment plan; an optimal plan verification unit for establishing a digital simulation model to verify and adjust the risk assessment of the second machine adjustment plan, and generating an optimal machine adjustment plan.

[0007] To achieve the above object, the present invention provides the following technical solutions:

[0008] The intelligent machine adjustment auxiliary decision-making system for spandex covered yarn production based on data fusion includes:

[0009] The multi-scale fusion unit is used to collect micro-data, meso-data and macro-data in production to obtain multi-scale data; based on production constraints, fusion weights are obtained to generate multi-scale fusion data;

[0010] A multi-objective collaborative decision-making unit extracts multi-scale fusion features based on multi-scale fusion data, establishes a multi-scale objective function including the multi-scale fusion features, and obtains a first machine adjustment plan; the multi-scale objective function includes a sub-objective function for optimizing molecular chain orientation, a sub-objective function for homogenizing fiber stress distribution, and a sub-objective function for maximizing yarn comprehensive performance;

[0011] The solution optimization unit is used to build a multi-time scale analysis mechanism for short-term online parameter fine-tuning, mid-term process pattern recognition, and long-term knowledge graph evolution. It analyzes the first machine adjustment plan and generates a second machine adjustment plan.

[0012] The optimal solution verification unit is used to establish a digital simulation model to verify and adjust the risk assessment of the second machine adjustment plan and generate the optimal machine adjustment plan.

[0013] Preferably, the microscopic data include the position, speed and orientation angle data of the molecular chain segment movement; the mesoscopic data include the fiber elongation, curvature and stress distribution data; the macroscopic data include the yarn strength, elongation and uniformity data;

[0014] According to the production process parameters, the influence weight coefficient of multi-scale data on product quality is calculated; based on the influence weight coefficient and the data reliability assessment results, the fusion weight of micro data, meso data and macro data is determined; the multi-scale data is weightedly fused according to the corresponding fusion weight to generate multi-scale fusion data.

[0015] Preferably, the multi-scale fusion features are obtained from the multi-scale fusion data by a feature extraction algorithm, including molecular chain orientation features, fiber strain uniformity features and yarn comprehensive performance features;

[0016] The multi-scale objective function is established as a function with the goals of optimizing molecular chain orientation, homogenizing fiber stress distribution, and maximizing yarn comprehensive performance;

[0017] The process of obtaining the first machine adjustment plan is as follows: using the multi-scale fusion feature as the input variable, the multi-scale objective function as the optimization target, and under the constraints of process parameters, generating the first machine adjustment plan including the coating speed, tension setting and temperature control parameters.

[0018] Preferably, the optimization process of the multi-scale objective function includes: obtaining the mapping relationship between the molecular chain orientation characteristics and the coating speed and tension setting, and constructing a molecular chain orientation optimization sub-objective function; obtaining the association model between the fiber strain uniformity characteristics and the coating speed and temperature control parameters, and constructing a fiber stress distribution homogenization sub-objective function; establishing a comprehensive evaluation model between the yarn comprehensive performance characteristics and the coating speed, tension setting and temperature control parameters, and constructing a yarn comprehensive performance maximization sub-objective function;

[0019] The three sub-objective functions are combined into a multi-scale objective function; an iterative optimization is performed to obtain an optimal solution set under the conditions of satisfying equipment capacity constraints, process safety constraints, and product quality constraints; and the optimal parameter combination is selected from the optimal solution set as the first machine adjustment plan based on production priority.

[0020] Preferably, the multi-time scale analysis mechanism includes:

[0021] Short-term online parameter fine-tuning mechanism: within the time scale of seconds, based on real-time quality detection feedback, key process parameters are adjusted to achieve online optimization of parameters; medium-term process pattern recognition mechanism: within the time scale of minutes, the matching degree between the current production status and the historical process pattern is identified, a process pattern database is established and pattern features are updated in real time; long-term knowledge graph evolution mechanism: within the time scale of hours, based on accumulated production data and process experience, the knowledge graph is used to update the association rules between process parameters and product quality to form a continuously evolving knowledge system; multi-time scale analysis mechanism realizes information sharing and coordinated optimization through data interfaces.

[0022] Preferably, the first machine adjustment plan is input into the multi-time scale analysis mechanism for comprehensive evaluation; the key parameters in the first machine adjustment plan are adjusted in real time through the short-term online parameter fine-tuning mechanism to obtain a set of fine-tuning parameters; the first machine adjustment plan is matched and analyzed with the historical optimal process mode through the mid-term process pattern recognition mechanism to identify potential process optimization space; the relevant process knowledge and experience rules are called through the long-term knowledge graph evolution mechanism to perform knowledge-driven improvement on the first machine adjustment plan; the first machine adjustment plan is iteratively optimized by combining the fine-tuning parameter set, process optimization space and knowledge-driven improvement to generate a second machine adjustment plan.

[0023] Preferably, the digital simulation model of the optimal solution verification unit includes:

[0024] The physical simulation layer is used to simulate the physical deformation process of spandex yarn and outer fiber, analyze the influence of process parameters in the second machine adjustment scheme on the physical properties of the material, and output material stress, strain and deformation data;

[0025] The process simulation layer simulates the complete process flow of covered yarn production based on material stress, strain, and deformation data, analyzes the impact of process parameter adjustments on production stability and product quality, and outputs process feasibility and quality prediction data;

[0026] The system simulation layer simulates the operating status of the production system based on process feasibility and quality prediction data, evaluates the second machine adjustment plan and execution risks, and outputs system performance evaluation and risk warning data;

[0027] Information flow is achieved through the data transmission interface, and the upper layer outputs data as constraints and input parameters for the lower layer, achieving multi-level simulation verification and risk assessment to generate the optimal machine adjustment plan.

[0028] Compared with the prior art, the present invention has the following beneficial effects:

[0029] 1. The present invention introduces three types of multi-scale data: microscopic, mesoscopic, and macroscopic, including the motion parameters of molecular segments, the mechanical morphological characteristics of fibers, and the physical properties of the yarn as a whole, and performs weighted processing through a fusion weight mechanism to generate multi-scale fusion data. By integrating multiple levels of factors that affect product quality, the present invention can more comprehensively reflect the correlation between raw materials, process parameters, and product performance, thereby improving the parameter sensitivity analysis capabilities during the machine adjustment process. Multi-scale data fusion processing not only enhances the adaptability of the model to complex process changes, but also improves the scientific nature and accuracy of the machine adjustment plan, effectively reduces the trial and error cost and the machine adjustment cycle, and significantly improves the quality consistency and stability of the production process.

[0030] 2. The present invention constructs a ternary multi-scale objective function with the goals of optimizing the molecular chain orientation, homogenizing the fiber stress distribution and maximizing the comprehensive performance of the yarn, and performs process parameter optimization on the basis of this data. This mechanism breaks through the existing machine adjustment method that only focuses on the optimization method of a local single target (such as tensile strength or output). It can comprehensively consider the orderliness of the microstructure and the balance of the mesoscopic stress distribution while ensuring the overall performance of the yarn, thereby achieving collaborative optimization between raw material utilization efficiency, process stability and product quality. Through the joint modeling of multi-objective functions and production constraints, the present invention can extract the machine adjustment plan that meets the actual production priority from the optimization solution set, so that the process adjustment process has both a global perspective and takes into account personalized needs. The introduction of this multi-objective collaborative optimization mechanism provides a more systematic, balanced and efficient machine adjustment path for the production of spandex coated yarn.

[0031] 3. The present invention introduces a multi-time scale analysis mechanism with three stages of “short-term-medium-term-long-term”, which deeply optimizes the preliminary machine adjustment plan through online parameter fine-tuning at the second level, process pattern recognition at the minute level, and knowledge graph evolution at the hour level. At the same time, by establishing a hierarchical digital simulation model, the full-process simulation verification and risk assessment of the machine adjustment plan before execution are realized, effectively avoiding uncertainties and process failures that may occur during the implementation of the plan. The present invention can realize dynamic adjustment and closed-loop optimization of the machine adjustment process, which not only improves the response speed of the machine adjustment and production flexibility, but also enhances the system's adaptability to different production tasks and environmental changes. The collaborative design of multiple time scales and multi-level verification has significantly improved the feasibility, safety and robustness of the machine adjustment plan, providing a strong technical guarantee for intelligent spinning production. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 This is a structural diagram of the intelligent machine adjustment auxiliary decision-making system for spandex covered yarn production based on data fusion provided by the present invention;

[0033] Figure 2 The present invention provides a flow chart of intelligent machine adjustment auxiliary decision-making for spandex covered yarn production based on data fusion;

[0034] Figure 3 A schematic diagram of intelligent machine adjustment auxiliary decision-making provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0035] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0036] Example 1:

[0037] The present invention provides an intelligent machine adjustment auxiliary decision system for spandex covered yarn production based on data fusion, please refer to Figure 1 and Figure 2 , the technical solution is as follows:

[0038] The multi-scale fusion unit is used to collect micro-data, meso-data and macro-data in production to obtain multi-scale data; based on production constraints, fusion weights are obtained to generate multi-scale fusion data;

[0039] A multi-objective collaborative decision-making unit extracts multi-scale fusion features based on multi-scale fusion data, establishes a multi-scale objective function including the multi-scale fusion features, and obtains a first machine adjustment plan; the multi-scale objective function includes a sub-objective function for optimizing molecular chain orientation, a sub-objective function for homogenizing fiber stress distribution, and a sub-objective function for maximizing yarn comprehensive performance;

[0040] The solution optimization unit is used to build a multi-time scale analysis mechanism for short-term online parameter fine-tuning, mid-term process pattern recognition, and long-term knowledge graph evolution. It analyzes the first machine adjustment plan and generates a second machine adjustment plan.

[0041] The optimal solution verification unit is used to establish a digital simulation model to verify and adjust the risk assessment of the second machine adjustment plan and generate the optimal machine adjustment plan.

[0042] Furthermore, the microscopic data includes the position, speed and orientation angle data of the movement of molecular chain segments; the mesoscopic data includes the fiber elongation, curvature and stress distribution data; the macroscopic data includes the yarn strength, elongation and uniformity data; and the molecular chain is the structure between spandex-coated yarns.

[0043] According to the production process parameters, the influence weight coefficient of multi-scale data on product quality is calculated; based on the influence weight coefficient and the data reliability assessment results, the fusion weight of micro data, meso data and macro data is determined; the multi-scale data is weightedly fused according to the corresponding fusion weight to generate multi-scale fusion data.

[0044] In this embodiment, the present invention realizes accurate data perception and comprehensive evaluation of the entire spandex covered yarn production process by introducing a fusion mechanism of microscopic, mesoscopic and macroscopic multi-scale data. Among them, the microscopic level can reflect the influence of material structure evolution on fiber performance; the mesoscopic level helps to reveal the local stress state and deformation law of process parameters in the fiber forming process; the macroscopic level comprehensively characterizes the overall performance of the final product. By constructing a fusion weight model, the degree of influence of data at different scales on product quality and its data reliability are comprehensively considered, and the weighted fusion of multi-scale data is systematically realized, making the fused data more representative and instructive. It effectively opens up the cross-scale correlation path from the original structural level to the final performance, provides comprehensive and accurate data support for subsequent machine adjustment decisions, and significantly improves the scientificity and adaptability of the machine adjustment plan, thereby improving production efficiency and reducing trial and error costs.

[0045] The data integrity assessment results evaluate the completeness and missing rate of multi-scale data collection; the data accuracy assessment results evaluate the measurement accuracy and systematic error of the data by comparing with the standard reference value; the data consistency assessment results evaluate the consistency and repeatability of data collected by the same type of sensors; the data timeliness assessment results evaluate the data collection time delay and data update frequency; the data stability assessment results evaluate the volatility and trend stability of the data in the time series; the comprehensive reliability score is the data reliability assessment result obtained by comprehensively weighting the above assessment results.

[0046] Furthermore, the multi-scale fusion features are obtained from the multi-scale fusion data by a feature extraction algorithm, including molecular chain orientation features, fiber strain uniformity features and yarn comprehensive performance features;

[0047] The multi-scale objective function is established as a function with the goals of optimizing molecular chain orientation, homogenizing fiber stress distribution, and maximizing yarn comprehensive performance;

[0048] The process of obtaining the first machine adjustment plan is as follows: using the multi-scale fusion feature as the input variable, the multi-scale objective function as the optimization target, and under the constraints of process parameters, generating the first machine adjustment plan including the coating speed, tension setting and temperature control parameters.

[0049] In this embodiment, the present invention realizes the systematic extraction and quantitative expression of key performance indicators at different scales by constructing a multi-scale fusion feature system covering molecular chain orientation, fiber strain uniformity and yarn comprehensive performance, which can more comprehensively reflect the multi-level impact of machine adjustment parameter changes on yarn quality. On this basis, a multi-scale optimization model with a multi-scale objective function as the core is proposed, which integrates the orderliness of the molecular structure level, the balance of the fiber mechanical state and the overall performance of the finished yarn into the machine adjustment target, and achieves a coordinated balance of the coupling relationship between various performance targets. By inputting the multi-scale fusion feature into the objective function model and performing optimization calculations based on the parameter constraints of the actual production process, a first machine adjustment scheme containing key parameters such as coating speed, tension setting and temperature control can be automatically generated. This scheme not only improves the intelligence level of the machine adjustment process and reduces the reliance on human experience intervention, but also significantly enhances the adaptability and robustness of the machine adjustment scheme to different performance targets, which helps to achieve stable control and optimization of the quality of coated yarn products.

[0050] Furthermore, the optimization process of the multi-scale objective function includes: obtaining the mapping relationship between the molecular chain orientation characteristics and the coating speed and tension setting, and constructing a molecular chain orientation optimization sub-objective function; obtaining the correlation model between the fiber strain uniformity characteristics and the coating speed and temperature control parameters, and constructing a fiber stress distribution homogenization sub-objective function; establishing a comprehensive evaluation model between the yarn comprehensive performance characteristics and the coating speed, tension setting and temperature control parameters, and constructing a yarn comprehensive performance maximization sub-objective function;

[0051] The three sub-objective functions are combined into a multi-scale objective function; an iterative optimization is performed to obtain an optimal solution set under the conditions of satisfying equipment capacity constraints, process safety constraints, and product quality constraints; and the optimal parameter combination is selected from the optimal solution set as the first machine adjustment plan based on production priority.

[0052] In this embodiment, the present invention fully exploits the inherent correlation between different performance characteristics and process parameters by decomposing the multiscale objective function into three sub-objective functions: optimizing molecular chain orientation, homogenizing fiber stress distribution, and maximizing yarn overall performance. By constructing a mapping relationship or evaluation model between the characteristic parameters and the coating speed, tension setting, and temperature control parameters, each sub-objective function is made analyzable and optimizable, thereby accurately characterizing the impact of the machine adjustment parameters on product performance. Furthermore, the multiscale objective function is iteratively optimized under various production constraints, not only ensuring the feasibility and safety of the optimization results, but also providing a multi-solution space for machine adjustment plans. Based on the current production priority, the system selects the most comprehensive parameter combination from the optimal solution set as the first machine adjustment plan, effectively improving the accuracy and responsiveness of the machine adjustment process. This method can achieve dynamic trade-offs between different objectives and optimize overall performance while ensuring equipment operation safety and product quality compliance, significantly enhancing the intelligence, personalization, and flexibility of the machine adjustment strategy in the spandex coated yarn production process.

[0053] Molecular chain orientation optimization sub-objective function: The molecular chain orientation characteristics are used as optimization variables, and the coating speed and tension are set as control parameters. The goal is to make the molecular chain reach the optimal orientation state during the coating process. By optimizing the combination of coating speed and tension parameters, the consistency and stability of the molecular chain orientation are maximized, ensuring that the spandex yarn maintains a good molecular structure arrangement during the coating process.

[0054] Fiber stress distribution homogenization sub-objective function: Taking the fiber strain uniformity characteristics as the optimization target, and the coating speed and temperature control parameters as the adjustment variables, the goal is to achieve maximum uniformity of the stress distribution of the fiber during the coating process. By coordinating the matching relationship between the coating speed and temperature control, the unevenness of the stress distribution inside the fiber is minimized, and fiber damage caused by local stress concentration is avoided.

[0055] Sub-objective function for maximizing comprehensive yarn performance: Taking the comprehensive performance characteristics of the yarn as the evaluation criteria, the synergistic effect of the three parameters of covering speed, tension setting and temperature control is comprehensively considered. The goal is to maximize the overall performance of the yarn, including the comprehensive optimization of strength, elongation and uniformity, to ensure that the final product meets the expected comprehensive quality requirements.

[0056] The multi-scale objective function organically combines the three sub-objective functions through a multi-objective optimization method and is constructed in a weighted summation manner. The molecular chain orientation optimization sub-objective function, the fiber stress distribution homogenization sub-objective function and the yarn comprehensive performance maximization sub-objective function are respectively assigned corresponding weight coefficients, and the weight coefficients are dynamically determined according to production priorities and product quality requirements; the comprehensive objective function seeks to achieve the optimal parameter combination for the overall system performance while ensuring the balance between the sub-objectives, thereby realizing the coordinated optimization of microscopic molecular structure, microscopic fiber performance and macroscopic yarn quality.

[0057] Furthermore, the multi-time scale analysis mechanism includes:

[0058] Short-term online parameter fine-tuning mechanism: within the time scale of seconds, based on real-time quality detection feedback, key process parameters are adjusted to achieve online optimization of parameters; medium-term process pattern recognition mechanism: within the time scale of minutes, the matching degree between the current production status and the historical process pattern is identified, a process pattern database is established and pattern features are updated in real time; long-term knowledge graph evolution mechanism: within the time scale of hours, based on accumulated production data and process experience, the knowledge graph is used to update the association rules between process parameters and product quality to form a continuously evolving knowledge system; multi-time scale analysis mechanism realizes information sharing and coordinated optimization through data interfaces.

[0059] In this embodiment, the present invention realizes the coordinated optimization and dynamic evolution of the machine adjustment strategy in different time dimensions by constructing a multi-time scale analysis mechanism covering three levels: short-term, medium-term and long-term. Within the time scale of seconds, the short-term online parameter fine-tuning mechanism is based on real-time quality detection results, which can quickly perceive product status changes and instantly fine-tune key process parameters, effectively suppressing quality fluctuations and improving response efficiency; within the time scale of minutes, the medium-term process pattern recognition mechanism can match and analyze the current production status with historical typical process patterns, which not only improves the rationality of process parameter settings, but also promotes the structured management and rapid call of process knowledge; within the time scale of hours, the long-term knowledge graph evolution mechanism relies on continuously accumulated production data and experience, and models and updates the complex relationship between process parameters and product quality in a graph manner, so that the machine adjustment decision has learning ability and long-term adaptability. The three realize information sharing and coordinated optimization through data interfaces, and build a complete closed-loop control system from real-time adjustment to long-term evolution, which not only improves the intelligence level of the machine adjustment process and the stability of the production system, but also provides solid data and knowledge support for the high-quality continuous production of spandex coated yarn.

[0060] Furthermore, the first machine adjustment plan is input into the multi-time scale analysis mechanism for comprehensive evaluation; the key parameters in the first machine adjustment plan are adjusted in real time through the short-term online parameter fine-tuning mechanism to obtain a set of fine-tuning parameters; the first machine adjustment plan is matched and analyzed with the historical optimal process mode through the mid-term process pattern recognition mechanism to identify potential process optimization space; the relevant process knowledge and experience rules are called through the long-term knowledge graph evolution mechanism to perform knowledge-driven improvement on the first machine adjustment plan; the first machine adjustment plan is iteratively optimized by combining the fine-tuning parameter set, process optimization space and knowledge-driven improvement to generate a second machine adjustment plan.

[0061] In this embodiment, the present invention integrates a multi-timescale analysis mechanism into the first tuning solution for comprehensive evaluation and optimization, constructing a dynamic tuning iterative path from real-time feedback to experience accumulation. A short-term online parameter fine-tuning mechanism enables real-time adjustments to key parameters in the first tuning solution within seconds, significantly improving the sensitivity and accuracy of parameter settings. A mid-term process pattern recognition mechanism matches and analyzes the first tuning solution with historically optimal process patterns, deeply exploring the potential optimization space for the current tuning solution in practical applications and improving the solution's process compatibility and robustness. A long-term knowledge graph evolution mechanism further incorporates production knowledge and empirical rules, utilizing graph reasoning and data-driven methods to structurally reinforce and modify the first tuning solution at the knowledge level, enhancing its interpretability and foresight. Ultimately, by integrating the fine-tuning parameter set, process optimization space, and knowledge-driven improvement content, the first tuning solution is comprehensively iteratively optimized, generating a second tuning solution with more comprehensive performance and intelligent adaptability. This method significantly improves the scientific nature, flexibility, and intelligence of tuning decisions, providing a more stable, efficient, and sustainably optimized tuning solution for spandex covered yarn production.

[0062] Furthermore, the digital simulation model of the optimal solution verification unit includes:

[0063] The physical simulation layer is used to simulate the physical deformation process of spandex yarn and outer fiber, analyze the influence of process parameters in the second machine adjustment scheme on the physical properties of the material, and output material stress, strain and deformation data;

[0064] The process simulation layer simulates the complete process flow of covered yarn production based on material stress, strain, and deformation data, analyzes the impact of process parameter adjustments on production stability and product quality, and outputs process feasibility and quality prediction data;

[0065] The system simulation layer simulates the operating status of the production system based on process feasibility and quality prediction data, evaluates the second machine adjustment plan and execution risks, and outputs system performance evaluation and risk warning data;

[0066] Information flow is achieved through the data transmission interface, and the upper layer outputs data as constraints and input parameters for the lower layer, achieving multi-level simulation verification and risk assessment to generate the optimal machine adjustment plan.

[0067] In this embodiment, the present invention systematically verifies the second machine adjustment scheme by constructing a multi-level digital simulation model, significantly improving the feasibility assessment accuracy and risk prediction ability of the machine adjustment scheme. The physical simulation layer is based on the physical deformation process of spandex yarn and external fiber under the influence of machine adjustment parameters, which can intuitively reflect the stress, strain and deformation behavior of the material under actual working conditions, and provide high-fidelity input data for subsequent process analysis; the process simulation layer simulates the complete covered yarn production process based on the physical simulation results, accurately evaluates the specific impact of process parameter changes on production stability and product quality, and outputs process feasibility and quality prediction data to ensure the adaptability and stability of the scheme in actual application; the system simulation layer further simulates the operating status of the overall production system based on process simulation data, systematically analyzes the potential risks and performance of the second machine adjustment scheme during execution, and outputs comprehensive system performance evaluation and early warning information. Each simulation layer realizes information linkage and hierarchical constraints through a data transmission interface, building a complete closed-loop simulation chain from material behavior to system operation. This mechanism not only enhances the predictive and stable control capabilities of the machine adjustment plan, but also provides a scientific and reliable decision-making basis for the generation of the final optimal machine adjustment plan, greatly improving the intelligence and refinement level of the spandex covered yarn production process.

[0068] The process feasibility data includes: equipment operation stability data, which evaluates the operating status and stability performance of each production equipment under the second machine adjustment plan; process parameter matching data, which evaluates the coordination and compatibility between the various process parameters in the machine adjustment plan; production efficiency prediction data, which evaluates the changes in production speed and efficiency after the implementation of the machine adjustment plan; process safety assessment data, which evaluates the safety risk level of the machine adjustment process and new parameter settings; resource consumption assessment data, which evaluates the impact of the machine adjustment plan on the consumption of raw materials, energy and human resources; process implementation difficulty data, which evaluates the operational complexity and implementation feasibility of the machine adjustment plan.

[0069] The quality prediction data include: product strength performance prediction data, which predicts the strength indicators and change trends of the yarn produced under the second machine adjustment plan; product elongation performance prediction data, which predicts the changes in yarn elongation and elastic recovery performance; product uniformity prediction data, which predicts the uniformity of yarn diameter, density uniformity and performance consistency; product appearance quality prediction data, which predicts the surface smoothness, color uniformity and coating integrity of the yarn; product comprehensive grade prediction data, which predicts the product quality grade based on a comprehensive evaluation of various quality indicators; quality stability prediction data, which predicts the stability and controllability of product quality between batches and over time; unqualified risk prediction data, which predicts the types of quality defects that may occur and the probability of their occurrence.

[0070] This invention introduces a multi-scale fusion mechanism of microscopic, mesoscopic and macroscopic data, constructs a multi-objective collaborative decision-making model, a multi-time scale analysis mechanism and a digital simulation verification unit, and achieves accurate data perception, intelligent machine adjustment decision-making and risk assessment for the entire spandex coated yarn production process. The system can effectively connect the cross-scale correlation from material structure to final product performance, collaboratively optimize the molecular chain orientation, fiber stress distribution uniformity and yarn comprehensive performance, and refine the machine adjustment plan through continuous learning and iteration. For details, please refer to Figure 3 This significantly improves the scientific nature, adaptability, intelligence, and robustness of the machine adjustment plan, enhances product quality stability and production efficiency, reduces trial-and-error costs and reliance on manual experience, and provides comprehensive technical support and decision-making basis for the high-quality, efficient, stable, and sustainably optimized production of spandex covered yarn.

[0071] Example 2:

[0072] This invention provides an application example of a data-fusion-based intelligent machine adjustment decision-making system for spandex-coated yarn production in a specific production environment. The actual application scenario is the automated 40D spandex-coated yarn production workshop of a large textile enterprise, which primarily produces spandex-coated yarn for high-end elastic fabrics. The specific implementation of the technical solution is as follows:

[0073] In the actual production environment of this textile enterprise, the workshop temperature is controlled between 22 and 25 degrees Celsius, and the humidity is maintained at around 65%. The multi-scale fusion unit collects comprehensive data from the 40D spandex covered yarn production process through a distributed sensor network.

[0074] In the core equipment area of ​​the production workshop, the system utilizes molecular dynamics sensors, installed at key locations within the coating machine. The sensor network collects 1,000 data points per second. Microscopic data includes the positional coordinates of molecular chain segments in three-dimensional space, recorded as x, y, and z coordinates; velocity vectors expressed as x, y, and z velocities; and orientation angle θ, which is measured with an accuracy of 0.1 degrees. The standard deviation of the molecular chain orientation angles of high-quality 40D spandex must be kept within a threshold range, otherwise the elastic recovery properties of the final product will be affected.

[0075] The production workshop is equipped with tension sensors, optical deformation detectors, and stress distribution scanners. These devices continuously monitor the mechanical state at the fiber level.

[0076] During actual measurement, the elongation of the outer coating fiber must be strictly controlled within a narrow range. If the elongation is too low, the coating will loosen; if it is too high, it will cause fiber breakage. The coefficient of unevenness in the stress distribution within the fiber must be less than 12%. Exceeding this value will cause quality problems in subsequent processing. The system uses a high-frequency sampling interval of 1 millisecond to ensure accurate capture of the dynamic changes in the fiber, which is at the leading level in the industry. The end of the production line is equipped with a yarn strength tester, elongation detection device, and uniformity analyzer.

[0077] A statistical analysis of production data from the past six months established a model for calculating impact weight coefficients. Taking into account the data reliability assessment results, the reliability coefficient of micro-data was 0.92, primarily affected by ambient temperature and humidity; the reliability coefficient of meso-data was 0.95, indicating good stability; and the reliability coefficient of macro-data was 0.98, indicating the highest accuracy.

[0078] Multi-scale fused data is generated using a weighted average algorithm. The specific calculation process is: the fused data equals 0.32 times the microscopic data, 0.38 times the mesoscopic data, and 0.30 times the macroscopic data. This fusion method fully considers the characteristics and reliability of data at different scales, ensuring the scientific and accurate fusion results.

[0079] Based on the generated multi-scale fusion data, key features are extracted and an optimization objective function is constructed. Specific implementation of multi-scale fusion feature extraction: The system uses a feature extraction algorithm that combines principal component analysis (PCA) and wavelet transform to extract three types of key features from massive fusion data.

[0080] The molecular chain orientation characteristics are obtained by calculating the statistical distribution parameters of the orientation angle. These parameters reflect the degree of arrangement order of the spandex molecules during the coating process and directly affect the elastic recovery properties of the yarn.

[0081] The fiber strain uniformity characteristic is quantified by the degree of discreteness of the stress distribution. The higher the uniformity index, the more uniform the stress distribution inside the fiber and the more stable the product quality.

[0082] The comprehensive performance characteristics of yarn are characterized by a weighted comprehensive score of strength, elongation and evenness, and the comprehensive performance index reflects the overall quality level of the product.

[0083] The first sub-objective function is the molecular chain orientation optimization function, whose mathematical goal is to find the maximum value, which is composed of the orientation consistency and the orientation angle stability. The goal of this function is to achieve the best alignment of the molecular chains during the coating process.

[0084] The second sub-objective function is the fiber stress distribution homogenization function, which aims to find the minimum value, which is composed of the stress non-uniformity coefficient plus the strain gradient. This function aims to reduce the stress concentration inside the fiber.

[0085] The third sub-objective function is the yarn comprehensive performance maximization function, which aims to find the maximum value, which is composed of strength index, elongation performance and uniformity. This function ensures that the final product achieves the best comprehensive performance.

[0086] Implementation of the first machine adjustment plan: When process parameter adjustments are required, the system automatically initiates the optimization calculation process. Using multi-scale fusion features as input variables, the system performs optimization calculations within the constraints of equipment capacity. Process safety constraints include tension setting ranges and temperature control ranges, which are essential for ensuring production safety.

[0087] The system uses a genetic algorithm to optimize the solution. This algorithm simulates the process of biological evolution, gradually finding the optimal solution through selection, crossover, and mutation. After 500 iterations, the system generated the first adjustment plan: wrapping speed, tension setting, and temperature control parameters.

[0088] The continuously running solution optimization system has built a multi-time scale analysis mechanism. It consists of three subsystems, each responsible for optimization analysis at a different time scale.

[0089] When the system suddenly detects a slight fluctuation in the yarn strength index, the short-term fine-tuning system immediately initiates the emergency response procedure within a 2-second time window.

[0090] When the quality monitoring system detects deviations from target yarn strength, it promptly identifies and addresses them. The system automatically calculates the optimal adjustment plan, adjusting both the wrapping speed and temperature. This coordinated adjustment ensures a dynamic balance between parameters. A set of fine-tuning parameters is obtained through cumulative optimization over 10 consecutive fine-tuning cycles.

[0091] The mid-term pattern recognition system begins to work, matching the first machine adjustment plan with the historical optimal patterns stored in the process pattern database. Using Euclidean distance calculation and similarity evaluation algorithms, the system searches the database for the most matching process pattern. After calculation, the system identifies the historical pattern with the best match.

[0092] The long-term knowledge evolution system starts a deep analysis program and calls the process knowledge graph established by the enterprise, which contains knowledge nodes and association rules.

[0093] The knowledge graph stores numerous correlations between process parameters and product quality, such as the relationship between coating speed and fiber stress distribution and the relationship between temperature control and molecular chain orientation. Based on the production data accumulated over the past month, the system updates the weight coefficients of these association rules, making the knowledge graph more accurate and practical.

[0094] The system comprehensively considers the fine-tuning parameter set, process optimization space, and knowledge-driven improvement information, and uses an intelligent fusion algorithm to optimize the solution. After complex weighted fusion calculations, a second machine adjustment solution was generated.

[0095] A three-level digital simulation model was established to fully verify the second machine adjustment plan. Inside a quiet simulation computing room, a high-performance computer cluster began running a physical simulation program. The system used finite element analysis to create a digital model of the 40D spandex yarn and outer covering fiber.

[0096] The simulation simulated the physical deformation of the spandex yarn and the coated fiber under specific conditions of coating speed and tension. The results showed that under these parameter settings, the stress distribution of the spandex yarn was more uniform. The calculated value of the fiber strain gradient indicated that deformation within the fiber was more uniform.

[0097] Based on the material stress, strain and deformation data output by physical simulation, the process simulation system begins to simulate the complete covered yarn production process.

[0098] The simulation system creates a digital model of the covering machine, including key components such as the spinning system, covering system, drafting system, and winding system. In this virtual environment, the system simulates the continuous production process and analyzes the specific impact of process parameter adjustments on production stability and product quality.

[0099] Building on the process simulation, the system simulation layer further simulated the entire production system's operating status under the second machine adjustment plan. The simulation system simulated production scenarios across different shifts and loads. Through layered verification and data analysis across the three simulation models, the system fully confirmed the feasibility and superiority of the second machine adjustment plan.

[0100] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. The intelligent machine adjustment auxiliary decision-making system for spandex covered yarn production based on data fusion is characterized by: include: Multi-scale fusion unit, used to collect micro-data, meso-data and macro-data in production to obtain multi-scale data; Obtain fusion weights based on production constraints and generate multi-scale fusion data; The multi-objective collaborative decision-making unit extracts multi-scale fusion features based on the multi-scale fusion data, establishes a multi-scale objective function including the multi-scale fusion features, and obtains the first engine adjustment plan; The solution optimization unit is used to build a multi-time scale analysis mechanism for short-term online parameter fine-tuning, mid-term process pattern recognition, and long-term knowledge graph evolution. It analyzes the first machine adjustment plan and generates a second machine adjustment plan. The optimal solution verification unit is used to establish a digital simulation model to verify and adjust the risk assessment of the second machine adjustment plan and generate the optimal machine adjustment plan.

2. The intelligent machine adjustment auxiliary decision-making system for spandex covered yarn production based on data fusion according to claim 1 is characterized by: The microscopic data include the position, speed and orientation angle data of the molecular chain segment movement; the mesoscopic data include the fiber elongation, curvature and stress distribution data; the macroscopic data include the yarn strength, elongation and uniformity data; According to the production process parameters, the influence weight coefficient of multi-scale data on product quality is calculated; based on the influence weight coefficient and the data reliability assessment results, the fusion weight of micro data, meso data and macro data is determined; the multi-scale data is weightedly fused according to the corresponding fusion weight to generate multi-scale fusion data.

3. The intelligent machine adjustment auxiliary decision-making system for spandex covered yarn production based on data fusion according to claim 1 is characterized by: The multi-scale fusion features are obtained from the multi-scale fusion data through a feature extraction algorithm, including molecular chain orientation features, fiber strain uniformity features and yarn comprehensive performance features; The multi-scale objective function is established as a function with the goals of optimizing molecular chain orientation, homogenizing fiber stress distribution, and maximizing yarn comprehensive performance; The process of obtaining the first machine adjustment plan is as follows: using the multi-scale fusion feature as the input variable, the multi-scale objective function as the optimization target, and under the constraints of process parameters, generating the first machine adjustment plan including the coating speed, tension setting and temperature control parameters.

4. The intelligent machine adjustment auxiliary decision-making system for spandex covered yarn production based on data fusion according to claim 3 is characterized by: The optimization process of the multi-scale objective function includes: obtaining the mapping relationship between the molecular chain orientation characteristics and the coating speed and tension setting, and constructing a molecular chain orientation optimization sub-objective function; obtaining the correlation model between the fiber strain uniformity characteristics and the coating speed and temperature control parameters, and constructing a fiber stress distribution homogenization sub-objective function; establishing a comprehensive evaluation model between the yarn comprehensive performance characteristics and the coating speed, tension setting and temperature control parameters, and constructing a yarn comprehensive performance maximization sub-objective function; The three sub-objective functions are combined into a multi-scale objective function; an iterative optimization is performed to obtain an optimal solution set under the conditions of satisfying equipment capacity constraints, process safety constraints, and product quality constraints; and the optimal parameter combination is selected from the optimal solution set as the first machine adjustment plan based on production priority.

5. The intelligent machine adjustment auxiliary decision-making system for spandex covered yarn production based on data fusion according to claim 1 is characterized by: The multi-time scale analysis mechanism includes: Short-term online parameter fine-tuning mechanism: within the time scale of seconds, based on real-time quality detection feedback, key process parameters are adjusted to achieve online optimization of parameters; medium-term process pattern recognition mechanism: within the time scale of minutes, the matching degree between the current production status and the historical process pattern is identified, a process pattern database is established and pattern features are updated in real time; long-term knowledge graph evolution mechanism: within the time scale of hours, based on accumulated production data and process experience, the knowledge graph is used to update the association rules between process parameters and product quality to form a continuously evolving knowledge system; multi-time scale analysis mechanism realizes information sharing and coordinated optimization through data interfaces.

6. The intelligent machine adjustment auxiliary decision-making system for spandex covered yarn production based on data fusion according to claim 5 is characterized by: The first generator adjustment plan is input into the multi-time scale analysis mechanism for comprehensive evaluation; the key parameters in the first generator adjustment plan are adjusted in real time through the short-term online parameter fine-tuning mechanism to obtain a set of fine-tuning parameters; Through the mid-term process pattern recognition mechanism, the first machine adjustment plan is matched and analyzed with the historical optimal process pattern to identify potential process optimization space; Through the long-term knowledge graph evolution mechanism, relevant process knowledge and experience rules are called upon to perform knowledge-driven improvements on the first machine adjustment plan. The first machine adjustment plan is iteratively optimized based on the comprehensive fine-tuning parameter set, process optimization space and knowledge-driven improvement to generate the second machine adjustment plan.

7. The intelligent machine adjustment auxiliary decision-making system for spandex covered yarn production based on data fusion according to claim 1 is characterized by: The digital simulation model of the optimal solution verification unit includes: The physical simulation layer is used to simulate the physical deformation process of spandex yarn and outer fiber, analyze the influence of process parameters in the second machine adjustment scheme on the physical properties of the material, and output material stress, strain and deformation data; The process simulation layer simulates the complete process flow of covered yarn production based on material stress, strain, and deformation data, analyzes the impact of process parameter adjustments on production stability and product quality, and outputs process feasibility and quality prediction data; The system simulation layer simulates the operating status of the production system based on process feasibility and quality prediction data, evaluates the second machine adjustment plan and execution risks, and outputs system performance evaluation and risk warning data; Information flow is achieved through the data transmission interface, and the upper layer outputs data as constraints and input parameters for the lower layer, achieving multi-level simulation verification and risk assessment to generate the optimal machine adjustment plan.

Images