An adaptive control method and device for warp knitting process based on a warp knitting process knowledge base

By using an adaptive control method based on a warp knitting process knowledge base, the problems of inconsistent terminology and scattered parameters in warp knitting production were solved, enabling rapid style changes and consistency across multiple machines, thereby improving production efficiency and product quality stability.

CN123067346APending Publication Date: 2026-11-10SINCETECH FUJIAN TECH CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202610998410.2
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-07-06
Publication Date
2026-11-10

AI Technical Summary

Technical Problem

The lack of a unified terminology system in warp knitting production and the fragmented process parameters make it impossible to accumulate and transfer experience and knowledge. Changing styles takes a long time, and there is poor consistency among multiple machines. The lack of adaptability affects efficiency and consistency.

Method used

An adaptive control method based on a warp knitting process knowledge base generates standardized fields through semantic understanding and standardized parsing, generates process parameter packages by combining precise retrieval and similarity reasoning, and compensates and corrects by collecting machine data in real time, forming a closed-loop control link.

Benefits of technology

It enables rapid model changeover and machine adjustment, improves production efficiency and consistency across multiple machines, reduces changeover time, increases experience reuse rate and product quality stability, and achieves an intelligent upgrade from experience-dependent to data-driven.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123067346A_ABST
    Figure CN123067346A_ABST
Patent Text Reader

Abstract

This application discloses a warp knitting process adaptive control method and device based on a warp knitting process knowledge base. By semantically understanding and standardizing the product demand information and mapping it into standardized fields, it solves the problems of inconsistent warp knitting terminology and difficulty in accumulating experience. It accurately retrieves information in the knowledge base using standardized fields as joint primary keys. When no match is found, it generates adaptation parameters through similarity reasoning, transforming process parameters from scattered process sheets into a structured knowledge base, which greatly improves the reuse rate of experience and shortens the time for machine adjustment when changing styles. By collecting the actual output value of the machine in real time and calculating the deviation from the standard target value, it generates compensation coefficients based on the historical data of the machine for closed-loop correction. It achieves personalized compensation for the mechanical individual differences of different machines, effectively ensuring the consistency of processes across multiple machines and reducing product fluctuations.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This application relates to the field of warp knitting manufacturing technology, and in particular to a warp knitting process adaptive control method and device based on a warp knitting process knowledge base. Background Technology

[0002] Warp knitting is an important knitting technique widely used in clothing, home textiles, and industrial textiles. During warp knitting production, process parameters (such as warp feed rate, tension, machine speed, and loop length) directly determine the fabric's weight, width, density, and appearance quality. Different fabric types, weight specifications, and widths require different combinations of process parameters, and adjustments and debugging of the warp knitting machine are necessary when changing production styles.

[0003] Currently, warp knitting production mainly relies on the experience of process engineers for process formulation and machine calibration. Some companies have established a simple process sheet recording system, which records the process parameters previously used according to product model for reference in subsequent production. When it is necessary to produce a new product, process engineers manually determine parameters such as warp feed, tension, and draw density based on experience or by reviewing historical process sheets, and then set them one by one on the machine for trial production, making repeated adjustments based on the trial production results.

[0004] The warp knitting industry has long lacked a standardized terminology system, resulting in inconsistent process terminology. For example, there are serious issues with the misuse of synonyms and variants in equipment parameters, process actions, fabric structures, defect descriptions, and fabric specifications, hindering the effective accumulation and transfer of experiential knowledge. Furthermore, existing technology often uses fragmented process sheets, lacking a structured and standardized parameter library for different fabrics, weights, and widths. Different specifications lack correlation, making retrieval and reasoning impossible. New product production relies heavily on trial and error based on experience, leading to low experience reuse rates. The lack of standardized parameter libraries and rapid retrieval methods necessitates manual record checking, calculations, and item-by-item settings, requiring multiple trial productions and corrections, taking hours or even days and severely impacting efficiency and equipment utilization. Moreover, when the same process parameters are executed on machines with different mechanical states, the lack of adaptive compensation results in inconsistent weights, widths, and densities, leading to large fluctuations and poor batch consistency in parallel production across multiple machines.

[0005] In summary, current warp knitting production lacks a technical solution that can unify terminology, standardize the storage of process knowledge, support rapid retrieval and reasoning, and adaptively adjust parameters based on actual machine feedback. Problems such as the inability to accumulate knowledge, the difficulty in reusing experience, low changeover efficiency, and poor consistency among multiple machines have persisted for a long time. There is an urgent need to provide a warp knitting process adaptive control method that can solve the above technical problems. Summary of the Invention

[0006] To address the problems in the prior art, this application provides a warp knitting process adaptive control method and apparatus based on a warp knitting process knowledge base, in order to solve one or more technical problems existing in the prior art.

[0007] In a first aspect, this application provides an adaptive control method for warp knitting processes based on a warp knitting process knowledge base, the method comprising: Receive product demand information, perform semantic understanding and standardized parsing on the product demand information, and generate standardized fields, the standardized fields including at least one of fabric type, weight, and width; Using the standardized field as a composite primary key, a precise retrieval is performed in the pre-constructed warp knitting process knowledge base. Each knowledge record in the warp knitting process knowledge base uses the preset standardized field as a composite primary key, and each knowledge record stores a standard process parameter package and a standard target value. If the precise search matches the knowledge record, the corresponding standard process parameter package and the corresponding standard target value are read; if the precise search does not match the knowledge record, a standard process parameter package and standard target value adapted to the product requirement information are generated based on the standardized fields and the knowledge record. The acquired standard process parameter package is sent to the target warp knitting machine for production; The actual output value of the target warp knitting machine is collected in real time, the deviation between the actual output value and the standard target value is calculated, a compensation coefficient is generated based on the deviation and the historical stable production data of the target warp knitting machine, and the process parameters of the target warp knitting machine are corrected and reissued.

[0008] In some embodiments, generating a standard process parameter package and standard target values ​​adapted to the product requirement information based on the standardized fields and the knowledge records includes: The standardized fields are converted into multidimensional feature vectors, and the similarity between the multidimensional feature vectors and the knowledge records is calculated. The knowledge records are sorted based on the similarity and the top N knowledge records are selected, where N≥1 and N is an integer; Based on the selected knowledge records, a standard process parameter package adapted to the product requirement information is generated by combining interpolation algorithms and rule-based reasoning, and corresponding standard target values ​​are generated simultaneously through reasoning.

[0009] In some embodiments, the step of generating a standard process parameter package adapted to the product requirement information based on the selected knowledge records, combined with interpolation algorithms and rule-based reasoning, includes: Based on at least one of the standardized fields, including fabric type, weight, and width information, the association rules pre-stored in the knowledge graph are triggered. The association rules include at least one of the association rules between weight and stretch density, and the association rules between weave type and warp control method. Using the selected knowledge records as sample data, and taking at least one of the weight and width as independent variables, linear interpolation calculations are performed to solve for the parameter reference value of at least one of the warp feed ratio, drawing density, and tension range; The parameter baseline value is corrected according to rules based on at least one of the fabric type and yarn characteristics, and at the same time, auxiliary process parameters including at least one of the yarn threading method and machine speed range are derived, and a standard process parameter package including at least one of the warp feed amount, tension, drafting density, machine speed and yarn threading method is output.

[0010] In some embodiments, the synchronous inference to generate the corresponding standard target value includes: Based on the standard target value corresponding to the selected knowledge record, and combined with at least one of the fabric type, weight, and width, a standard target value that fits the product requirement information is generated by weighted interpolation and preset process rules. The standard target value includes at least one of standard weight, standard width, and standard fabric density.

[0011] In some embodiments, the semantic understanding and standardized parsing processing of the product demand information to generate standardized fields includes: Based on a pre-built warp knitting terminology mapping table, standardized fields are generated according to the product requirement information. The warp knitting terminology mapping table includes standard terms obtained by normalizing and mapping synonyms and synonyms based on at least one of equipment parameters, process actions, fabric structure, defect descriptions, and fabric specifications.

[0012] In some embodiments, the step of using the standardized field as a composite primary key to perform precise retrieval in a pre-built warp knitting process knowledge base includes: Using the fabric type, weight, and width as the core joint search fields, the precise search is determined to have hit the knowledge record only when the fabric type, weight, and width all completely match the knowledge record recorded in the warp knitting process knowledge base.

[0013] In some embodiments, the real-time acquisition of the actual output value of the target warp knitting machine, calculation of the offset between the actual output value and the standard target value, generation of a compensation coefficient based on the offset and the historical stable production data of the target warp knitting machine, and re-issuance of the process parameters in the standard process parameter package after correction include: After the target warp knitting machine stabilizes in production, real-time data on actual weight, actual width, and actual fabric density are collected. The actual weight, actual width, and actual fabric density are respectively calculated as sample offset data based on their offset from the corresponding standard target values. The sample offset data, which consists of no fewer than a preset number of consecutive sets, is subjected to mean filtering to output effective offset data. Based on the effective offset data, the historical stable production data, the mechanical state characteristics of the target warp knitting machine and the product process sensitivity, a multi-parameter coupled compensation coefficient is generated. The compensation coefficient is used to correct the current operating process parameters of the target warp knitting machine, and the corrected process parameters are then reissued to the target warp knitting machine.

[0014] In some embodiments, the method further includes: The revised process parameters and their corresponding standardized fields are entered as new knowledge records into the warp knitting process knowledge base.

[0015] In some embodiments, the step of recording the modified process parameters and corresponding standardized fields as new knowledge records into the warp knitting process knowledge base includes: Obtain M sets of actual output data during continuous and stable production on the target warp knitting machine, where M≥10 and M is an integer; Calculate the average deviation and standard deviation between the actual output data of the M groups and the standard target value; When the average deviation is less than a first preset threshold and the standard deviation is less than a second preset threshold, the corrected process parameters and the corresponding standardized fields are entered as new knowledge records into the warp knitting process knowledge base.

[0016] In a second aspect, this application provides a warp knitting process adaptive control device based on a warp knitting process knowledge base, the device being used to implement the warp knitting process adaptive control method based on a warp knitting process knowledge base as described in any of the first aspects.

[0017] In a third aspect, embodiments of this application disclose a computer device, including a memory and a processor, wherein the memory stores a computer program that can run on the processor, and when the computer program is executed by the processor, it implements the warp knitting process adaptive control method based on a warp knitting process knowledge base as described in any of the first aspects.

[0018] In a fourth aspect, embodiments of this application disclose a computer-readable storage medium storing a computer program, which, when executed, implements the warp knitting process adaptive control method based on a warp knitting process knowledge base as described in any of the first aspects.

[0019] According to the specific embodiments provided in this application, the following technical effects are disclosed: In this embodiment, firstly, by semantically understanding and standardizing product demand information, product demands are mapped into standardized fields, fundamentally solving the industry problems of inconsistent warp knitting process terminology and the inability to accumulate experience. Secondly, using standardized fields as joint primary keys, precise searches are performed in the knowledge base, and when no match is found, adaptation parameters are generated through similarity reasoning. This transforms the process parameters for different fabrics, weights, and widths from scattered process sheets into a structured, searchable knowledge base, significantly improving experience reuse and shortening the time for changing styles and adjusting machines. Thirdly, by collecting the actual output values ​​of the machines in real time and calculating the offset from the standard target values, compensation coefficients are generated based on historical machine data for closed-loop correction. This enables personalized compensation for individual differences in the machinery of different machines, effectively ensuring process consistency across multiple machines and reducing product fluctuations. Finally, an adaptive control chain of "retrieval execution → feedback collection → offset compensation → parameter correction" is formed, realizing the intelligent upgrade of warp knitting production from experience-dependent to data-driven, systematically solving multiple technical problems such as inconsistent terminology, lack of standardized parameters, slow style changes, poor consistency across multiple machines, and lack of adaptive capabilities.

[0020] Additional aspects and advantages of this application will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description

[0021] The disclosure of this invention will become more readily understood with reference to the accompanying drawings. It will be readily understood by those skilled in the art that these drawings are for illustrative purposes only and are not intended to limit the scope of protection of this invention. Furthermore, similar numbers in the drawings are used to denote similar components, wherein: Figure 1 This is an architecture diagram of the warp knitting process knowledge base provided in some embodiments of this application; Figure 2 This is a flowchart of an adaptive control method for warp knitting process based on a warp knitting process knowledge base provided in some embodiments of this application; Figure 3 This is a schematic diagram of the structure of a computer device provided in some embodiments of this application. Detailed Implementation

[0022] Some embodiments of the present invention will now be described with reference to the accompanying drawings. Those skilled in the art should understand that these embodiments are merely illustrative of the technical principles of the present invention and are not intended to limit the scope of protection of the present invention.

[0023] As described in the background section, in existing warp knitting production, there has long been a serious problem of interchangeable and misused terminology across five dimensions: equipment parameters, process actions, fabric structure, defect descriptions, and fabric specifications. This hinders the effective accumulation and transfer of experiential knowledge. Furthermore, the lack of standardized, structured parameter libraries for different fabrics, weights, and widths necessitates manual review of historical records, parameter conversion, and machine setup during style changes. Multiple trial productions and adjustments are required to achieve acceptable quality, often taking hours or even days, severely impacting production efficiency and equipment utilization. Moreover, when the same process parameters are executed on machines with different mechanical states, the lack of adaptive compensation results in inconsistent weights, widths, and densities, leading to significant product fluctuations and poor batch consistency during parallel production. None of these methods adequately meet the urgent needs of intelligent upgrades in warp knitting production for high efficiency, high consistency, and rapid style changes.

[0024] To address one or more of the aforementioned issues, this application proposes an adaptive control method and device for warp knitting processes based on a warp knitting process knowledge base. By using semantic understanding and standardized parsing, product demand information is mapped to standardized fields, fundamentally solving the problems of inconsistent warp knitting terminology and difficulty in accumulating experience. Precise retrieval is performed in a pre-built knowledge base using standardized fields as joint primary keys. If no match is found, a matching parameter package is generated through similarity reasoning, enabling rapid machine adjustment for style changes. Furthermore, by collecting real-time data on actual machine output and calculating the offset from the standard target value, compensation coefficients are generated based on historical machine data for closed-loop correction. Personalized compensation is achieved for individual differences in the machinery of different machines, effectively ensuring process consistency across multiple machines. This method forms an adaptive control chain of "retrieval execution → feedback collection → offset compensation → parameter correction," reducing the time for style adjustment from hours or even days to seconds or minutes. This significantly improves production efficiency, experience reuse rate, and product quality stability, realizing an intelligent upgrade from experience-dependent to data-driven warp knitting production. It provides a highly efficient collaborative solution for warp knitting manufacturing, from process design to machine execution.

[0025] The solution of this application will now be described in detail with reference to the accompanying drawings and various embodiments.

[0026] Example 1 The warp knitting process adaptive control method provided in this application embodiment is implemented based on a pre-built warp knitting process knowledge base, referring to... Figure 1As shown, this warp knitting process knowledge base mainly includes a semantic understanding layer, a structured knowledge base core, and knowledge base functional interfaces. The semantic understanding layer includes functional modules such as terminology standardization, parameter normalization, and rule formalization. In some embodiments, the terminology standardization module has a built-in warp knitting domain terminology mapping table covering dimensions such as equipment parameters, process actions, fabric structure, defect descriptions, and fabric specifications. The terminology standardization module receives product requirement information in natural language form (such as "1800 mm yarn length," "yarn arrangement method," "plain weave," etc.), and uses string matching and semantic mapping algorithms to uniformly convert synonymous terms in each dimension into standard terms. For example, "yarn feed," "yarn length," and "yarn feed amount" are uniformly mapped to "warp feed amount"; "warp threading," "yarn arrangement," and "yarn loading" are uniformly mapped to "yarn threading"; and "plain weave," "flat combing," and "flat combing structure" are uniformly mapped to "warp-plain weave." Through terminology standardization, misunderstandings caused by inconsistent terminology are eliminated, laying a unified semantic foundation for subsequent retrieval and reasoning. Parameter normalization and rule formalization functional modules are also included. The parameter normalization module is responsible for converting process parameters with different units, dimensions, and expressions into a unified dimension and numerical range. For example, machine speed parameters may be input in different units such as "revolutions per minute," "needles per second," and "meters per minute." This module converts them into standard units (such as "revolutions per minute") according to preset unit conversion rules. Tension parameters may be expressed in different units such as "cN," "g," and "N," which are also converted into standard units (such as "cN"). In addition, this module performs range verification and outlier detection on the parameters to ensure the rationality and validity of the input parameters. The rule formalization module transforms the empirical knowledge and process rules in the warp knitting field into computer-executable logical expressions. This module retrieves pre-stored association rules from knowledge graphs or rule bases, such as "weight and tension density are negatively correlated," "the warp feed ratio of mesh structures needs to be controlled in segments," and "warp and satin structures need to be matched with specific yarn threading methods," and formalizes them into condition-action pairs (IF-THEN rules) or mathematical constraint expressions. The formalized rules can be called by subsequent reasoning modules during parameter interpolation and correction, realizing the structured expression and automated application of domain knowledge.

[0027] The structured knowledge base is the core data element of this application's technical solution, used for the structured storage and unified management of various types of knowledge involved in the entire warp knitting production process. This knowledge base uses standardized fields (at least including fabric type, weight, and width) as joint primary keys, and each knowledge record corresponds to a complete package of process parameters and standard target values. Further reference... Figure 1 As shown, the structured knowledge base mainly includes the following sub-bases: The process parameter library stores core process parameters required for warp knitting production, including standard stitch count, tension range, speed range, temperature range, warp feed rate, draft density, and loop length. Each process parameter record is associated with the corresponding fabric type, weight, width, and weave type, and indicates the applicable machine model range. The process parameter library supports precise searching and similarity searching by standardized fields, serving as the primary data source for quickly obtaining recommended parameter packages when changing production styles.

[0028] The fabric attribute library stores the inherent attribute characteristics of different fabrics, including weight specifications, width standards, yarn specifications, fabric classification, and physical properties (elasticity coefficient, shrinkage rate), etc. This sub-library provides fabric-dimensional constraints for parameter interpolation and rule correction. For example, different yarn specifications correspond to different recommended tension ranges, and different fabric classifications correspond to different draft density benchmarks.

[0029] The pattern rule library stores the structural rules and technological constraints related to warp knitting patterns. This sub-library records the weave structure of different patterns (warp plain weave, warp satin weave, mesh weave, jacquard weave, etc.) and their corresponding density distribution, color rules, functional zoning, warp feed control methods, yarn threading path rules, comb bar movement patterns, and connection constraints between different weaves. The pattern rule library provides the basis for reasoning about the weave dimension in similarity searches, for example, the warp feed ratio of mesh weaves needs to be segmented and controlled according to the mesh size and distribution density.

[0030] The experience case library is used to structure and manage various practical cases generated during warp knitting production. Each experience record includes at least the following dimensions of information: Successful cases, which record the effective combination of process parameters and production effects under specific fabric types, weights, widths, and weave conditions, serving as high-confidence samples for similar searches and parameter recommendations; Failed cases, which record the combination of process parameters that led to product quality defects and the reasons for failure, used to avoid known invalid or undesirable parameter ranges when recommending parameters; Machine adjustment records, which store the complete adjustment process of the machine from initial parameter settings to stable production during each style change, including the adjustment direction, adjustment step size, and corresponding quality feedback of each parameter, providing operational path references for subsequent similar style changes; Optimization trajectory, which records the iterative optimization history of process parameters in multiple production runs of the same product, showing the evolution of parameters from initial values ​​to optimal values ​​and the comparison of the effects of each optimization, providing data support for continuous model learning and parameter accuracy improvement.

[0031] The defect knowledge base is used for structured storage and management of various defects in the warp knitting production process. Each defect record includes at least the following dimensions of information: Defect type, using standardized terminology to uniformly name defects, covering common categories such as yarn breakage, needle marks, uneven density, oil stains, horizontal stripes, and vertical stripes, eliminating differences in naming the same defect between different shifts and machines; Cause analysis, recording the possible causes of the defect and their weights, for example, uneven density is related to multiple factors such as fluctuations in warp feed, slippage of the tension roller, and unstable yarn tension; Solutions, storing verified and effective process adjustment measures for the defect, such as specific operation instructions for increasing warp feed, reducing machine speed, cleaning the yarn path, or changing the yarn batch; Preventive measures, recording process specifications or monitoring parameter thresholds to avoid the recurrence of the defect, such as setting an alarm when the warp feed deviation exceeds ±2%.

[0032] The knowledge base interface serves as an interactive bridge between the structured knowledge base and upper-layer application modules, providing unified data access and manipulation capabilities. Further reference... Figure 1As shown, the knowledge base functional interface mainly includes three major functional modules: intelligent retrieval interface, semantic reasoning interface, and knowledge expansion interface. The intelligent retrieval interface receives standardized fields output by the semantic understanding layer and performs retrieval operations within the structured knowledge base. This interface supports two retrieval modes: precise retrieval mode, which performs a complete match using the standardized fields as the joint primary key, directly returning the standard process parameter package and standard target value from the corresponding knowledge record upon a match, suitable for the rapid reuse of existing mature products; and similarity retrieval mode, where, when the precise retrieval fails, the standardized fields are converted into multi-dimensional feature vectors, and cosine similarity is calculated with the feature vectors of all records in the knowledge base. The Top-N high-similarity records are selected and returned for subsequent interpolation and rule reasoning. The intelligent retrieval interface is the core support for rapid model changeover and machine tuning. The semantic reasoning interface is used to call pre-stored association rules and experiential knowledge in the knowledge base to perform reasoning operations to generate or correct process parameters. This interface primarily supports the following reasoning functions: rule triggering, which triggers corresponding association rules (such as the relationship between basis weight and draft density, and the relationship between fabric type and warp feed control method) based on information such as fabric type and weave characteristics in standardized fields; parameter interpolation, which uses multiple similar records found in similarity retrieval as samples and performs linear interpolation with the target basis weight or width as the independent variable to calculate the baseline values ​​of parameters such as warp feed ratio, draft density, and tension range; and rule correction, which corrects the interpolation results based on rules such as weave type and yarn characteristics, and derives auxiliary process parameters such as yarn threading method and machine speed range. The semantic reasoning interface transforms the implicit knowledge of domain experts into executable reasoning logic, enabling the automated application of experiential knowledge. The knowledge expansion interface supports the dynamic updating and continuous iteration of the knowledge base. It primarily includes: adding new records by receiving production-verified process parameter packages and their corresponding standardized fields, storing them as new knowledge records in the knowledge base; updating existing records by correcting and optimizing parameter values, target values, or confidence levels in existing knowledge records; and expanding the rule base by formally entering new association rules or empirical patterns summarized from production into the rule base. The knowledge expansion interface supports both manual review and automatic system entry (the latter requires meeting preset entry verification conditions, such as multiple consecutive sets of actual output data deviating from the standard target value within a threshold range). Through the knowledge expansion interface, the knowledge base can be continuously enriched and self-optimized during the production process, possessing continuous learning capabilities.

[0033] In some embodiments, the construction process of the warp knitting process knowledge base includes: First, collecting process data of historically successfully produced warp knitting products, and converting key features such as fabric type, weight, and width in each historical process record into standardized fields using the same semantic understanding and standardized parsing methods; then, using the combination of these standardized fields as a unique joint primary key, associating and storing the complete process parameter package and standard target value corresponding to that record. The standard process parameter package includes at least adjustable warp knitting production parameters such as warp feed, warp tension, machine speed, loop length, and draft density; the standard target value includes at least the expected fabric weight, width, fabric density, and appearance quality indicators under the process parameter package. Each knowledge record is stored in the knowledge base in the form of a triple of "joint primary key—process parameter package—standard target value," ensuring that the process data between different product specifications are independent, structurally unified, and unambiguous.

[0034] Figure 2 This is a flowchart of the warp knitting process adaptive control method based on a warp knitting process knowledge base provided in the embodiments of this application, referred to... Figure 2 As shown, the method includes the following steps: S100: Receive product demand information, perform semantic understanding and standardized parsing processing on the product demand information, and generate standardized fields, wherein the standardized fields include at least one of fabric type, weight, and width.

[0035] Specifically, the system first receives product requirement information from users or the upper-level production management system. This information can be in the form of natural language text, structured form data, or a combination of both. For example, process engineers might input descriptions like "need to produce a stretch warp-knitted fabric with a weight of 180 g / m² and a width of 160 cm" or "weave a spandex mesh fabric with a weight of 210 g / m² and a width of 150 cm" through a human-computer interaction interface. Because the warp knitting industry suffers from significant synonymy and confusion regarding fabric type, weight, and width (e.g., "stretch fabric" vs. "spandex fabric," "weight" vs. "unit area mass," "width" vs. "width width"), directly using the original requirements for retrieval or reasoning will lead to numerous omissions or errors.

[0036] To this end, this application embodiment performs semantic understanding and standardized parsing processing on the received product requirement information. Specifically, the semantic understanding module of the pre-built warp knitting process knowledge base is invoked to perform word segmentation, entity recognition, and intent extraction on the input information using natural language processing technology, identifying candidate phrases related to fabric type, weight, and width. Subsequently, the standardized parsing module matches the candidate phrases with the terminology standardization knowledge base: for the fabric type field, "elastic fabric," "spandex-containing fabric," and "elastic warp-knitted fabric" are uniformly mapped to the standard value "spandex warp-knitted fabric"; for the weight field, the numerical value and unit are extracted, and "weight 180," "180 gsm," and "180 grams per square meter" are uniformly parsed as "180 g / m²"; for the width field, "width 160," "width 1.6 meters," and "160 cm" are uniformly parsed as "160 cm." If a field value is missing in the input information, it can be completed through a preset default value or rule-based reasoning, or the user can be prompted to supplement the input. Understandably, this warp knitting process knowledge base pre-collects common synonyms, near-synonyms, abbreviations, and non-standard expressions in the industry, and establishes a mapping relationship between them and standard field values.

[0037] Ultimately, semantic understanding and standardized parsing generate a set of structured, unambiguous standardized fields, including at least fabric type, weight, and width. These standardized fields serve as the unified primary key for subsequent process parameter retrieval and reasoning, eliminating the obstacles to experience accumulation caused by inconsistent warp knitting terminology and synonyms at the source. This lays the data foundation for building a reusable and searchable structured process knowledge base.

[0038] S200: Using the standardized field as the composite primary key, perform precise retrieval in the pre-constructed warp knitting process knowledge base. Each knowledge record in the warp knitting process knowledge base uses the preset standardized field as the composite primary key, and each knowledge record stores a standard process parameter package and a standard target value.

[0039] To improve the retrieval efficiency and reuse rate of process parameters, this application embodiment uses the standardized fields as the composite primary key for precise retrieval in the warp knitting process knowledge base. The specific structure of the warp knitting process knowledge base has been described in detail in the foregoing embodiments. Its core is that each knowledge record uses a preset standardized field (at least including fabric type, weight, and width, and may further include extended fields such as weave type and yarn specifications) as a unique composite primary key, and each record corresponds to storing a complete set of standard process parameters and the expected standard target values ​​to be achieved under that parameter package.

[0040] Specifically, once the system obtains standardized fields of product requirement information (e.g., fabric type is spandex warp-knitted fabric, weight is 180 g / m², width is 160 cm, etc.), it calls the intelligent retrieval interface in the knowledge base function interface, passing this set of standardized fields as query parameters. The intelligent retrieval interface first initiates precise retrieval mode: in the core storage area of ​​the structured knowledge base (including process parameter library, fabric attribute library, pattern rule library, experience case library, etc.), it performs an equality matching query using all fields of the composite primary key. Since the data records in the knowledge base have been indexed and stored according to the same standardized fields during construction, precise retrieval can directly locate unique knowledge records with completely identical composite primary keys.

[0041] It should be noted that precise retrieval only performs a complete match on all fields of the composite primary key, without performing fuzzy matching or similarity calculation. This design has the following technical advantages: First, when a successful process record for the same specification product already exists in the knowledge base, precise retrieval can unambiguously and without trial and error reproduce the process parameter package, directly avoiding the process of manually reviewing historical process sheets and repeated trial and error adjustments, significantly shortening the time for changing models and adjusting machines; Second, using standardized fields as the composite primary key eliminates mismatches caused by inconsistent terminology or differences in expression, ensuring that each retrieval result has a definite uniqueness; Third, precise retrieval has extremely low computational overhead and can be completed in milliseconds, meeting the real-time requirements of high-frequency model changes on the production floor.

[0042] S300: If the precise search matches the knowledge record, then the corresponding standard process parameter package and the corresponding standard target value are read; if the precise search does not match the knowledge record, then a standard process parameter package and standard target value adapted to the product requirement information are generated based on the standardized fields and the knowledge record.

[0043] Specifically, during the precise retrieval process, if the current standardized field completely matches the composite primary key of a knowledge record in the warp knitting process knowledge base, it is considered a hit. In this case, the pre-stored standard process parameter package and corresponding standard target value in the knowledge record are directly read as the benchmark parameter set for subsequent machine production. If the precise retrieval fails, meaning there is no composite primary key record in the knowledge base that is exactly the same as the current standardized field, then based on the standardized field and the existing similar knowledge records in the knowledge base, a standard process parameter package and standard target value adapted to the current product requirement information are generated through a combination of similarity retrieval, parameter interpolation, and rule reasoning. This enables intelligent derivation of process parameters for product specifications without historical records.

[0044] S400: The acquired standard process parameter package is sent to the target warp knitting machine for production.

[0045] Specifically, firstly, based on the target warp knitting machine's model and communication protocol, the standard process parameter package is formatted and converted into a control instruction set recognizable by the target warp knitting machine. These control instructions include at least the warp feed rate setting, warp tension setting, machine speed setting, loop length setting, and draft density setting. Subsequently, the control instructions are sent to the target warp knitting machine's programmable logic controller (PLC) or embedded control system via an industrial fieldbus (such as EtherCAT, Profinet, etc.) or a workshop wireless network. Upon receiving the instructions, the controller or control system automatically writes the parameters and confirms the status of each actuator (warp feed motor, draft motor, spindle motor, etc.) and starts production. Throughout the production process, the machine's execution status is monitored in real time to ensure stable operation of the process parameters according to the set values. Through the above-mentioned sending and execution steps, fully automated parameter transmission from demand analysis to machine production is achieved, avoiding the time-consuming and error-prone problems of manual item-by-item setting, and significantly improving the efficiency of changing models and adjusting the machine.

[0046] S500: Real-time acquisition of the actual output value of the target warp knitting machine, calculation of the offset between the actual output value and the standard target value, generation of compensation coefficient based on the offset and the historical stable production data of the target warp knitting machine, correction of the current operating process parameters of the target warp knitting machine and re-issuance.

[0047] Specifically, during the operation of the target warp knitting machine, sensors and online detection devices deployed on the machine (including an online weight detector, width measurement sensor, fabric density visual inspection module, tension sensor, and speed encoder) collect the actual output value of the target warp knitting machine in real time and compare it with the standard target value to calculate the deviation between the two. Based on this deviation and the historical stable production data of the target warp knitting machine, a corresponding compensation coefficient is generated. This compensation coefficient is used to correct the currently operating process parameters, and the corrected parameters are then re-issued to the target warp knitting machine, forming a closed-loop control link of "collection—calculation—compensation—correction—issuance," ensuring that the machine's output continuously approaches the standard target value.

[0048] In some embodiments, the method for generating a standard process parameter package and standard target values ​​adapted to the product requirement information based on the standardized fields and the knowledge records includes: The standardized fields are converted into multidimensional feature vectors, and the similarity between the multidimensional feature vectors and the knowledge records is calculated. The knowledge records are sorted based on the similarity and the top N knowledge records are selected, where N≥1 and N is an integer; Based on the selected knowledge records, a standard process parameter package adapted to the product requirement information is generated by combining interpolation algorithms and rule-based reasoning, and corresponding standard target values ​​are generated simultaneously through reasoning.

[0049] When an exact search using standardized fields as the joint primary key fails, the similarity inference generation process begins. First, the standardized fields (at least including fabric type, weight, and width) are converted into multi-dimensional feature vectors. For continuous numerical fields, such as weight and width, their values ​​can be used as the corresponding dimension components of the vector and normalized. For discrete categorical fields, such as fabric type, one-hot encoding or word embedding can be used to map them into vector segments. Then, each knowledge record in the warp knitting process knowledge base is traversed, extracting the corresponding preset standardized fields from the joint primary key of each knowledge record. The feature vector of this knowledge record is generated according to the same feature transformation rules, and the similarity between the multi-dimensional feature vector of the current product requirement information and the feature vectors of each knowledge record is calculated. After the similarity calculation, the knowledge records are sorted from high to low similarity, and the top N knowledge records are selected, where N ≥ 1 and N is an integer (usually 3 to 5). Based on this, the semantic inference interface is called, combining interpolation algorithms and rule-based inference to generate a standard process parameter package adapted to the current product requirements. While generating the process parameter package, the corresponding standard target values ​​are simultaneously inferred, namely the expected fabric weight, width, fabric density and appearance quality indicators. The inference method can adopt the same logic as parameter interpolation, and the target values ​​in historical successful cases are used as a reference for interpolation and rule correction.

[0050] By organically combining the above-mentioned feature vectorization, similarity calculation, filtering, interpolation and rule reasoning, the embodiments of this application can fully utilize the successful experiences and domain knowledge of similar specifications in the knowledge base to intelligently derive standard process parameter packages and standard target values ​​that are adapted to the needs of new products, even without completely identical historical records.

[0051] It should be noted that this application does not limit the specific calculation method for similarity. Those skilled in the art can choose an appropriate similarity measurement method based on the data characteristics and real-time requirements of the warp knitting process knowledge base. As an example, and not a limitation, cosine similarity can be used for evaluation, that is, by calculating the cosine value of the angle between two feature vectors to measure their directional consistency. This method is suitable for high-dimensional sparse vector scenarios and has high computational efficiency. In addition, Euclidean distance, Manhattan distance, Pearson correlation coefficient, or approximate similarity retrieval based on locality-sensitive hashing can also be used. Any measurement method that can characterize the degree of similarity between feature vectors corresponding to standardized fields falls within the protection scope of this application.

[0052] In some embodiments, the step of generating a standard process parameter package adapted to the product requirement information based on the selected knowledge records, combined with interpolation algorithms and rule-based reasoning, includes: Based on at least one of the standardized fields, including fabric type, weight, and width information, the association rules pre-stored in the knowledge graph are triggered. The association rules include at least one of the association rules between weight and stretch density, and the association rules between weave type and warp control method. Using the selected knowledge records as sample data, and taking at least one of the weight and width as independent variables, linear interpolation calculations are performed to solve for the parameter reference value of at least one of the warp feed ratio, drawing density, and tension range; The parameter baseline value is modified according to rules based on at least one of the fabric type and yarn characteristics, and at the same time, auxiliary process parameters including at least one of the yarn threading method and machine speed range are derived, and a standard process parameter package including at least one of the warp feed amount, tension, drafting density, machine speed and yarn threading method is output.

[0053] Specifically, firstly, based on at least one of the fabric type, weight, and width information carried in the standardized fields, pre-stored association rules within the knowledge graph are triggered. These association rules include at least one of the following: association rules between weight and tension density (e.g., higher weight corresponds to increased tension density to maintain fabric tightness) and association rules between weave type and warp feed control method (e.g., mesh weaves require segmented warp feed control, and warp-satin weaves require a specific warp feed ratio). Subsequently, using the selected N knowledge records as sample data, and at least one of the weight and width information from the current product requirement as independent variables, linear interpolation is performed on the corresponding parameters in the sample data. The interpolation process is performed separately for core process parameters such as warp feed ratio, tension density, and tension range, to obtain the parameter baseline values ​​under the current target weight or width. After obtaining the baseline value, the system further modifies it according to the fabric type (e.g., spandex warp-knitted fabric, nylon warp-knitted mesh) and yarn characteristics (e.g., yarn count, elastic modulus) based on current product requirements. For example, the warp feed is appropriately increased when the spandex content is high, while the tension baseline is reduced for fine denier yarns. Simultaneously, auxiliary process parameters are automatically derived based on triggering rules and correction results, including at least the yarn threading method (e.g., full threading / loose threading ratio) and machine speed range. Finally, all the above parameters are integrated to output a complete standard process parameter package. This package includes at least warp feed, tension, draft density, machine speed, and yarn threading method, providing a set of directly executable parameter instructions for subsequent machine production.

[0054] In some embodiments, the synchronous inference to generate the corresponding standard target value includes: Based on the standard target value corresponding to the selected knowledge record, and combined with at least one of the fabric type, weight, and width, a standard target value that fits the product requirement information is generated by weighted interpolation and preset process rules. The standard target value includes at least one of standard weight, standard width, and standard fabric density.

[0055] Specifically, using the standard target values ​​corresponding to N knowledge records selected through similarity screening as the base sample, and combining the fabric type, weight, and width information from the current product requirement information, a combination of weighted interpolation and preset process rules is used to fit and generate the target value. First, different interpolation weights are assigned according to the similarity between each knowledge record and the current requirement, with records with higher similarity receiving greater weights. Then, using the weight and width of the current requirement as independent variables, weighted linear interpolation is performed on the standard weight, standard width, and standard fabric density in the sample records to obtain preliminary target value estimates. Based on this, preset process rules are invoked for correction. For example, when the fabric type is elastic, the standard width is appropriately scaled down based on the elasticity coefficient; when the fabric density requirement is high, the standard weight is fine-tuned based on the weave type. Finally, a complete set of standard target values ​​adapted to the current product requirement information is fitted and generated, including at least standard weight, standard width, and standard fabric density. These target values ​​will serve as the benchmark for offset calculation in subsequent closed-loop control, guiding the adaptive correction of machine process parameters.

[0056] In some embodiments, the semantic understanding and standardized parsing processing of the product demand information to generate standardized fields includes: Based on a pre-built warp knitting terminology mapping table, standardized fields are generated according to the product requirement information. The warp knitting terminology mapping table includes standard terms obtained by normalizing and mapping synonyms and synonyms based on at least one of equipment parameters, process actions, fabric structure, defect descriptions, and fabric specifications.

[0057] Specifically, when performing semantic understanding and standardized parsing of product demand information, standardized fields can be generated based on a pre-built warp knitting terminology mapping table. This mapping table covers the core terminology dimensions involved in the entire warp knitting production process, including at least one of the following: equipment parameter terminology mapping, process action terminology mapping, fabric structure terminology mapping, defect description terminology mapping, and fabric specification terminology mapping. For each dimension, the mapping table pre-collects commonly used synonyms and alternative expressions within the industry and normalizes them to a unified standard terminology. For example: in the equipment parameter dimension, different terms such as "yarn feed", "thread length", and "thread feed" are uniformly mapped to "warp feed"; in the process action dimension, "warp threading", "yarn arrangement", and "yarn loading" are uniformly mapped to "yarn threading"; in the fabric structure dimension, "plain weave", "combed weave", and "combed weave" are uniformly mapped to "warp plain weave"; in the defect description dimension, "horizontal bar", "horizontal stripe", and "stop mark" are uniformly mapped to "stop horizontal stripe"; in the fabric specification dimension, "gram weight", "gram weight per square meter", and "gsm" are uniformly mapped to "gram weight", and "width", "width", and "fabric cover" are uniformly mapped to "width".

[0058] Upon receiving product requirement information in natural language form, string matching and semantic mapping algorithms can be used to match each candidate phrase in the input information with synonyms and synonyms in the aforementioned mapping table. Once a match is found, it is replaced with the corresponding standard term. Through this processing based on a terminology mapping table in the warp knitting field, product requirement information from different sources and with different expression habits can be uniformly converted into unambiguous standardized fields. This fundamentally solves the long-standing problem of mixed terminology and synonyms in the warp knitting industry, laying a semantically unified field foundation for subsequent accurate retrieval and similarity reasoning.

[0059] In some embodiments, the step of using the standardized field as a composite primary key to perform precise retrieval in a pre-built warp knitting process knowledge base includes: Using the fabric type, weight, and width as the core joint search fields, the precise search is determined to have hit the knowledge record only when the fabric type, weight, and width all completely match the knowledge record recorded in the warp knitting process knowledge base.

[0060] Specifically, during precise retrieval, the fabric type, weight, and width generated after semantic understanding and standardized parsing can be used as the core joint retrieval fields. The precise retrieval employs an exact match rule: a knowledge record is considered matched only if all three fields—fabric type, weight, and width—required by the current product are completely identical to the corresponding fields stored in a knowledge record in the warp knitting process knowledge base. In other words, if any field differs (e.g., different fabric types, inconsistent weight values, or inconsistent width values), even if the other two fields are the same, it is not considered a match, and the process proceeds to subsequent similarity reasoning. By using fabric type, weight, and width as the core joint retrieval fields and performing a strict exact match determination, this embodiment of the application can unambiguously locate successful process records that are completely identical to the current product specifications, achieving accurate reuse of historical mature experience and avoiding the risk of misusing inapplicable process parameters due to partial field matching.

[0061] In some embodiments, the real-time acquisition of the actual output value of the target warp knitting machine, calculation of the offset between the actual output value and the standard target value, generation of a compensation coefficient based on the offset and the historical stable production data of the target warp knitting machine, and re-issuance of the process parameters in the standard process parameter package after correction include: After the target warp knitting machine stabilizes in production, real-time data on actual weight, actual width, and actual fabric density are collected. The actual weight, actual width, and actual fabric density are respectively calculated as sample offset data based on their offset from the corresponding standard target values. The sample offset data, which consists of no fewer than a preset number of consecutive sets, is subjected to mean filtering to output effective offset data. Based on the effective offset data, the historical stable production data, the mechanical state characteristics of the target warp knitting machine and the product process sensitivity, a multi-parameter coupled compensation coefficient is generated. The compensation coefficient is used to correct the current operating process parameters of the target warp knitting machine, and the corrected process parameters are then reissued to the target warp knitting machine.

[0062] After the target warp knitting machine starts production according to the standard process parameter package and the production status stabilizes, a real-time data acquisition and closed-loop correction process is initiated. Specifically, the actual weight, actual width, and actual fabric density data during the production process can be collected in real time using an online weight detector, width measurement sensor, and fabric density visual inspection module deployed on the target warp knitting machine. The collected data is uploaded to the control system batch by batch at preset time intervals or production length intervals.

[0063] Subsequently, the weight offset between the actual weight and the standard weight, the width offset between the actual width and the standard width, and the density offset between the actual fabric density and the standard fabric density are calculated separately, and these three are used together as the sample offset data for the current batch. To avoid miscompensation due to single abnormal fluctuations (such as instantaneous yarn splicing or environmental disturbances), mean filtering can be applied to at least a preset number of sample offset data sets (e.g., five or ten consecutive sets) to filter out high-frequency noise and occasional spikes before outputting valid offset data. Through mean filtering, the system can more stably and reliably reflect the systematic deviation trend of machine production.

[0064] After obtaining the effective offset data, it is combined with the historical stable production data of the target warp knitting machine. Simultaneously, the machine's mechanical characteristics (such as spindle bearing wear, warp feed motor response characteristics, and tension roller surface friction coefficient) and the current product's process sensitivity (i.e., the responsiveness of different process parameters to quality indicators) are comprehensively considered. A multi-parameter coupled compensation coefficient is generated through a preset compensation model. This compensation coefficient is not an independent correction for a single process parameter, but rather a coefficient vector that jointly adjusts multiple related parameters such as warp feed, tension density, and tension to address the complex relationships of mutual coupling and influence between different parameters in warp knitting production. For example, when the basis weight is too high and the width is too narrow, the compensation coefficient may simultaneously increase the tension density and decrease the warp feed to synergistically correct deviations in both dimensions.

[0065] After generating the compensation coefficient, the process parameters of the target warp knitting machine are jointly corrected using this coefficient. The corrected process parameters are then redeployed to the programmable logic controller or embedded control system of the target warp knitting machine via an industrial fieldbus. The machine controller automatically updates the setpoints of each actuator and continues production. The aforementioned closed-loop process of "acquisition—filtering—offset calculation—multi-parameter coupling compensation—correction distribution" is continuously executed in a loop during production until the offset between the actual output value and the standard target value stabilizes within a preset tolerance threshold. Through this mechanism, the embodiments of this application can achieve dynamic and accurate adaptive compensation for individual mechanical differences of different machines, long-term wear and tear, and environmental changes, effectively ensuring the consistency and stability of product quality when multiple machines are produced in parallel.

[0066] In some embodiments, the method further includes: The revised process parameters and their corresponding standardized fields are entered as new knowledge records into the warp knitting process knowledge base.

[0067] Specifically, after the aforementioned closed-loop adaptive correction process is completed, the verified and effective process parameters and corresponding standardized fields obtained from the target warp knitting machine can be entered as new knowledge records into the warp knitting process knowledge base. Standardized fields must include at least fabric type, weight, and width; process parameters must include at least the corrected warp feed, tension, draft density, machine speed, and yarn threading method, forming a complete parameter package. The entry operation can be performed through the knowledge extension interface in the knowledge base's functional interface. Through this dynamic update mechanism, the warp knitting process knowledge base can continuously absorb verified and effective process experience during actual production, continuously enrich the range of product specifications covered, and gradually improve the sample quality of similar reasoning and the accuracy of parameter recommendations, thereby achieving the knowledge base's self-iteration and continuous learning capabilities.

[0068] In some embodiments, the step of recording the modified process parameters and corresponding standardized fields as new knowledge records into the warp knitting process knowledge base includes: Obtain M sets of actual output data during continuous and stable production on the target warp knitting machine, where M≥10 and M is an integer; Calculate the average deviation and standard deviation between the actual output data of the M groups and the standard target value; When the average deviation is less than a first preset threshold and the standard deviation is less than a second preset threshold, the corrected process parameters and the corresponding standardized fields are entered as new knowledge records into the warp knitting process knowledge base.

[0069] Before entering the revised process parameters and corresponding standardized fields into the warp knitting process knowledge base, they can be validated to ensure the reliability and generalization ability of the newly added knowledge records. Specifically, M sets of actual output data during continuous and stable production on the target warp knitting machine can be obtained, where M is a preset integer and M≥10 (e.g., M is 15 or 20). The actual output data includes at least actual basis weight, actual fabric width, and actual fabric density. Subsequently, the average deviation and standard deviation between these M sets of actual output data and the corresponding standard target values ​​are calculated. The average deviation reflects the system performance deviation of the revised process parameters in actual production, while the standard deviation reflects the stability and consistency of the production process. When the average deviation is less than a first preset threshold (e.g., the average basis weight deviation is less than ±2% of the standard basis weight) and the standard deviation is less than a second preset threshold (e.g., the standard deviation of basis weight is less than 1% of the standard basis weight), it indicates that the revised process parameters can produce products that meet the expected quality targets in a long-term and stable manner. Based on this, the parameter combination is judged to have high confidence and can be used as effective experience accumulation. Once the above conditions are met, the corrected process parameters and their corresponding standardized fields are automatically added as new knowledge records and entered into the warp knitting process knowledge base in the form of a triple of "combined primary key - standard process parameter package - standard target value". Through this admission verification mechanism, the embodiments of this application can avoid mistakenly storing process parameters that are accidentally successful or lack stability into the knowledge base, thus ensuring the data quality of the knowledge base and the reliability of subsequent inference results.

[0070] Example 2 Corresponding to Embodiment 1 above, this application also provides a warp knitting process adaptive control device based on a warp knitting process knowledge base. This device is used to implement the warp knitting process adaptive control method based on a warp knitting process knowledge base as described in any one of Embodiment 1. In this embodiment, content that is the same as or similar to that in Embodiment 1 above can be referred to the above description and will not be repeated hereafter.

[0071] Example 3 Corresponding to Embodiment 1 above, this application also provides a computer device, including: a processor and a memory, wherein the memory stores a computer program that can run on the processor, and when the computer program is executed by the processor, it executes the warp knitting process adaptive control method based on the warp knitting process knowledge base provided in any of the above embodiments.

[0072] in, Figure 3An exemplary computer device 1500 is shown, which may specifically include a processor 1510, a video display adapter 1511, a disk drive 1512, an input / output interface 1513, a network interface 1514, and a memory 1520. The processor 1510, video display adapter 1511, disk drive 1512, input / output interface 1513, network interface 1514, and memory 1520 can communicate with each other via a communication bus 1530.

[0073] The processor 1510 can be implemented using a general-purpose CPU (Central Processing Unit), microprocessor, application-specific integrated circuit (ASIC), or one or more integrated circuits, and is used to execute relevant programs to implement the technical solution provided by the present invention.

[0074] The memory 1520 can be implemented in the form of ROM (Read Only Memory), RAM (Random Access Memory), static storage device, dynamic storage device, etc. The memory 1520 can store the operating system 1521 for controlling the operation of the electronic device, and the basic input / output system (BIOS) for controlling the low-level operations of the electronic device. Additionally, it can store a web browser 1523, a data storage management system 1524, and a device identification information processing system 1525, etc. The aforementioned device identification information processing system 1525 can be the application program that specifically implements the aforementioned steps in this embodiment of the invention. In summary, when implementing the technical solution provided by this invention through software or firmware, the relevant program code is stored in the memory 1520 and is called and executed by the processor 1510.

[0075] Input / output interface 1513 is used to connect input / output modules to realize information input and output. Input / output modules can be configured as components in the device (not shown in the figure) or externally connected to the device to provide corresponding functions. Input devices may include keyboards, mice, touch screens, microphones, various sensors, etc., and output devices may include displays, speakers, vibrators, indicator lights, etc.

[0076] Network interface 1514 is used to connect a communication module (not shown in the figure) to enable communication between this device and other devices. The communication module can communicate via wired means (such as USB, Ethernet cable, etc.) or wireless means (such as mobile network, WIFI, Bluetooth, etc.).

[0077] The bus includes a pathway for transmitting information between various components of the device, such as processor 1510, video display adapter 1511, disk drive 1512, input / output interface 1513, network interface 1514, and memory 1520.

[0078] In addition, the electronic device can also obtain information on specific claim conditions from the virtual resource object claim condition information database for condition judgment, and so on.

[0079] It should be noted that although the above-described device only shows the processor 1510, video display adapter 1511, disk drive 1512, input / output interface 1513, network interface 1514, memory 1520, bus, etc., in specific implementations, the device may also include other components necessary for normal operation. Furthermore, those skilled in the art will understand that the above-described device may only include the components necessary for implementing the present invention, and not necessarily all the components shown in the figures.

[0080] Example 4 Corresponding to Embodiment 1 above, this application also provides a computer-readable storage medium. In this embodiment, the content that is the same as or similar to that in Embodiment 1 above can be referred to the above description and will not be repeated hereafter.

[0081] The computer-readable storage medium stores a computer program thereon, which, when executed by a processor, implements the warp knitting process adaptive control method based on a warp knitting process knowledge base as described above.

[0082] In some implementations of this application, when the computer program is executed by the processor, it can also implement the steps corresponding to the method described in Embodiment 1. For details, please refer to the detailed description in Embodiment 1, which will not be repeated here.

[0083] As can be seen from the above description of the embodiments, those skilled in the art can clearly understand that the present invention can be implemented by means of software plus necessary general-purpose hardware platforms. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in various embodiments or some parts of the embodiments of the present invention.

[0084] The various embodiments in this specification are described in a progressive manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, for system or system embodiments, since they are basically similar to method embodiments, the description is relatively simple, and relevant parts can be referred to the descriptions in the method embodiments. The systems and system embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without creative effort.

[0085] It should be understood that various parts of this application can be implemented using hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented using software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0086] It should be noted that, unless otherwise specified, the relevant parts of the different embodiments described above can be referred to and learned from each other. This application will not repeat the same parts.

[0087] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of this application. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0088] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this application, "multiple" means at least two, such as two, three, etc., unless otherwise explicitly specified.

[0089] Although embodiments of this application have been shown and described above, it is to be understood that the various embodiments described herein are presented only to aid in understanding and teaching the claimed features. These embodiments are provided only as representative examples of embodiments and are not exhaustive and / or exclusive. It should be understood that the advantages, embodiments, examples, functions, features, structures and / or other aspects described herein should not be considered as limitations on the scope of the invention as defined by the claims or on the equivalents of the claims, and other embodiments may be utilized and modifications may be made without departing from the scope of the claimed invention. Various embodiments of the invention may suitably include, consist of, or substantially consist of suitable combinations of the disclosed elements, components, features, parts, steps, means, etc., rather than those specifically described herein. Furthermore, this disclosure may include other inventions not currently claimed but which may be claimed in the future.

Claims

1. A warp knitting process adaptive control method based on a warp knitting process knowledge base, characterized in that, The method includes: Receive product demand information, perform semantic understanding and standardized parsing on the product demand information, and generate standardized fields, the standardized fields including at least one of fabric type, weight, and width; Using the standardized field as a composite primary key, a precise retrieval is performed in the pre-constructed warp knitting process knowledge base. Each knowledge record in the warp knitting process knowledge base uses the preset standardized field as a composite primary key, and each knowledge record stores a standard process parameter package and a standard target value. If the precise search matches the knowledge record, the corresponding standard process parameter package and the corresponding standard target value are read; if the precise search does not match the knowledge record, a standard process parameter package and standard target value adapted to the product requirement information are generated based on the standardized fields and the knowledge record. The acquired standard process parameter package is sent to the target warp knitting machine for production; The actual output value of the target warp knitting machine is collected in real time, the deviation between the actual output value and the standard target value is calculated, a compensation coefficient is generated based on the deviation and the historical stable production data of the target warp knitting machine, and the process parameters of the target warp knitting machine are corrected and reissued.

2. The warp knitting process adaptive control method based on a warp knitting process knowledge base according to claim 1, characterized in that, The step of generating a standard process parameter package and standard target values ​​adapted to the product requirement information based on the standardized fields and the knowledge records includes: The standardized fields are converted into multidimensional feature vectors, and the similarity between the multidimensional feature vectors and the knowledge records is calculated. The knowledge records are sorted based on the similarity and the top N knowledge records are selected, where N≥1 and N is an integer; Based on the selected knowledge records, a standard process parameter package adapted to the product requirement information is generated by combining interpolation algorithms and rule-based reasoning, and corresponding standard target values ​​are generated simultaneously through reasoning.

3. The warp knitting process adaptive control method based on a warp knitting process knowledge base according to claim 2, characterized in that, The standard process parameter package adapted to the product requirement information is generated based on the selected knowledge records, combined with interpolation algorithms and rule-based reasoning, including: Based on at least one of the standardized fields, including fabric type, weight, and width information, the association rules pre-stored in the knowledge graph are triggered. The association rules include at least one of the association rules between weight and stretch density, and the association rules between weave type and warp control method. Using the selected knowledge records as sample data, and taking at least one of the weight and width as independent variables, linear interpolation calculations are performed to solve for the parameter reference value of at least one of the warp feed ratio, drawing density, and tension range; The parameter baseline value is corrected according to rules based on at least one of the fabric type and yarn characteristics, and at the same time, auxiliary process parameters including at least one of the yarn threading method and machine speed range are derived, and a standard process parameter package including at least one of the warp feed amount, tension, drafting density, machine speed and yarn threading method is output.

4. The warp knitting process adaptive control method based on a warp knitting process knowledge base according to claim 2, characterized in that, The synchronous inference generates the corresponding standard target value, including: Based on the standard target value corresponding to the selected knowledge record, and combined with at least one of the fabric type, weight, and width, a standard target value that fits the product requirement information is generated by weighted interpolation and preset process rules. The standard target value includes at least one of standard weight, standard width, and standard fabric density.

5. The warp knitting process adaptive control method based on a warp knitting process knowledge base according to any one of claims 1 to 4, characterized in that, The semantic understanding and standardized parsing of the product demand information to generate standardized fields includes: Based on a pre-built warp knitting terminology mapping table, standardized fields are generated according to the product requirement information. The warp knitting terminology mapping table includes standard terms obtained by normalizing and mapping synonyms and synonyms based on at least one of equipment parameters, process actions, fabric structure, defect descriptions, and fabric specifications.

6. The warp knitting process adaptive control method based on a warp knitting process knowledge base according to any one of claims 1 to 4, characterized in that, The step of using the standardized field as a composite primary key to perform precise retrieval in a pre-built warp knitting process knowledge base includes: Using the fabric type, weight, and width as the core joint search fields, the precise search is determined to have hit the knowledge record only when the fabric type, weight, and width all completely match the knowledge record recorded in the warp knitting process knowledge base.

7. The warp knitting process adaptive control method based on a warp knitting process knowledge base according to any one of claims 1 to 4, characterized in that, The process involves real-time acquisition of the actual output value of the target warp knitting machine, calculation of the offset between the actual output value and the standard target value, generation of a compensation coefficient based on the offset and the historical stable production data of the target warp knitting machine, and re-issuance of the process parameters in the standard process parameter package after correction. This includes: After the target warp knitting machine stabilizes in production, real-time data on actual weight, actual width, and actual fabric density are collected. The actual weight, actual width, and actual fabric density are respectively calculated as sample offset data based on their offset from the corresponding standard target values. The sample offset data, which consists of no fewer than a preset number of consecutive sets, is subjected to mean filtering to output effective offset data. Based on the effective offset data, the historical stable production data, the mechanical state characteristics of the target warp knitting machine and the product process sensitivity, a multi-parameter coupled compensation coefficient is generated. The compensation coefficient is used to correct the current operating process parameters of the target warp knitting machine, and the corrected process parameters are then reissued to the target warp knitting machine.

8. The warp knitting process adaptive control method based on a warp knitting process knowledge base according to any one of claims 1 to 4, characterized in that, The method further includes: The revised process parameters and their corresponding standardized fields are entered as new knowledge records into the warp knitting process knowledge base.

9. The warp knitting process adaptive control method based on a warp knitting process knowledge base according to claim 8, characterized in that, The step of entering the revised process parameters and corresponding standardized fields as new knowledge records into the warp knitting process knowledge base includes: Obtain M sets of actual output data during continuous and stable production on the target warp knitting machine, where M≥10 and M is an integer; Calculate the average deviation and standard deviation between the actual output data of the M groups and the standard target value; When the average deviation is less than a first preset threshold and the standard deviation is less than a second preset threshold, the corrected process parameters and the corresponding standardized fields are entered as new knowledge records into the warp knitting process knowledge base.

10. A warp knitting process adaptive control device based on a warp knitting process knowledge base, characterized in that, The device is used to implement the warp knitting process adaptive control method based on the warp knitting process knowledge base as described in any one of claims 1-9.