Knowledge graph-based fly weaving process reasoning method and system

Abstract

The application discloses a fly-weaving weaving process reasoning method and system based on a knowledge graph. The method comprises the following steps: constructing a fly-weaving weaving knowledge graph based on multi-dimensional historical fly-weaving weaving data and mining association rules of parameters and defects; constructing a bidirectional association reasoning model of defects and parameters based on the fly-weaving weaving knowledge graph and the association rules; acquiring real-time fly-weaving weaving data or target weaving material parameters; performing feature fusion on the real-time fly-weaving weaving data or the target weaving material parameters to obtain global fusion features; and inputting the global fusion features into the bidirectional association reasoning model to obtain a reasoning result. The application can solve the problems of strong dependence on fly-weaving weaving experience, no standard parameters, difficult prediction of defects and difficult reuse of knowledge.

Description

Technical Field

[0001] This application relates to the field of artificial intelligence technology, and in particular to a reasoning method and system for flyweaving process based on knowledge graphs. Background Technology

[0002] Flyknitting is an intelligent weaving technology that uses CNC flat knitting machines and functional yarns as raw materials to achieve one-piece molding of shoe uppers. Compared with traditional shoe upper processing technology, it has advantages such as lightweight, good breathability, good fit and strong plasticity. It is now widely used in the preparation of various shoe materials and is the core technology for the functional upgrading of shoe materials.

[0003] The existing fly knitting process relies on manual experience, and the process parameters, defect rules, and pattern constraints are scattered and lack a unified system; multi-source data cannot be integrated, resulting in low process debugging efficiency and high defect rate; knowledge is difficult to reuse between different machine models, raw materials, and structures, and the intelligent decision-making capability is insufficient. Summary of the Invention

[0004] This application aims to address at least one of the technical problems existing in the prior art. To this end, this application proposes a knowledge graph-based reasoning method and system for flyknitting processes to solve the aforementioned technical problem.

[0005] In a first aspect, this application provides a knowledge graph-based reasoning method for flyknitting processes, the method comprising: A knowledge graph of flyweaving was constructed based on multi-dimensional historical flyweaving data, and the association rules between parameters and defects were mined. Based on the aforementioned weaving knowledge graph and the aforementioned association rules, a bidirectional association reasoning model for defects and parameters is constructed. Obtain real-time flyweaving data or target fabric parameters; The real-time flying weaving data or target fabric parameters are fused to obtain global fused features; The global fusion features are input into the bidirectional correlation reasoning model to obtain the reasoning results; wherein, the reasoning results include at least one of defect prediction, parameter optimization and automatic process recommendation.

[0006] In at least some embodiments of this application, The flyknitting data includes process parameters, raw material properties, pattern structure, and defect characteristics.

[0007] In at least some embodiments of this application, The construction of a knowledge graph for knitting based on multi-dimensional historical knitting data and the mining of association rules between parameters and defects include: Using key objects in the flying knitting data as entities and relationships such as cause, association, constraint, and mitigation, a flying knitting knowledge graph is constructed by extracting triples from multi-dimensional historical flying knitting data. Association rule mining algorithms are used to extract frequent itemsets from multi-dimensional historical flyweaving data, calculate the support and confidence of each parameter combination with defects, and screen out association rules between parameters with high confidence and defects.

[0008] In at least some embodiments of this application, The construction of a bidirectional association reasoning model for defects and parameters based on the knitting knowledge graph and the association rules includes: assigning confidence scores to the association rules and storing them in the knitting knowledge graph to form a bidirectional association reasoning model for defects and parameters.

[0009] In at least some embodiments of this application, The step of fusing features from the real-time flyweaving data or target fabric parameters to obtain global fused features includes: Based on the pre-trained process semantic coding model using corpus data specific to the field of fly weaving, a unified semantic coding is performed on real-time fly weaving data to obtain a unified semantic feature vector; The entities and relationships in the weaving knowledge graph are vectorized by knowledge embedding to obtain knowledge embedding vectors; A multi-head attention mechanism is used to adaptively and weightfully fuse the unified semantic feature vector and the knowledge embedding vector to obtain a global fused feature.

[0010] In at least some embodiments of this application, The step of vectorizing entities and relationships in the knitting knowledge graph through knowledge embedding to obtain knowledge embedding vectors includes: Map the entities and relationships in the aforementioned weaving knowledge graph into low-dimensional dense vectors; Make similar processes and similar defects close in distance in vector space; The topological structure and semantic associations in the aforementioned weaving knowledge graph are preserved to obtain the knowledge embedding vector.

[0011] In at least some embodiments of this application, The multi-head attention mechanism is used to adaptively and weightedly fuse the unified semantic feature vector and the knowledge embedding vector to obtain a global fused feature, including: The unified semantic feature vector and the knowledge embedding vector are concatenated along the channel dimension to form a multi-dimensional hybrid feature. The importance weights of different dimensions of features are automatically learned through a multi-head attention mechanism. Key features are given high weights and noisy features are given low weights. Multi-dimensional features are adaptively weighted and fused to output global fused features.

[0012] In at least some embodiments of this application, The step of inputting the global fusion features into the bidirectional correlation reasoning model to obtain the reasoning result includes: If the global fusion features are derived from real-time flyweaving data, the bidirectional correlation reasoning model outputs defect prediction and parameter optimization. If the global fusion features are derived from the target fabric parameters, the bidirectional correlation inference model will automatically recommend the output process.

[0013] In at least some embodiments of this application, The method further includes: Online verification is performed based on the real-time knitting data and quality inspection results. The verification results are then fed back to the knitting knowledge graph and the bidirectional association reasoning model for closed-loop optimization.

[0014] In a second aspect, this application provides a knowledge graph-based reasoning system for flyknitting processes, comprising: The graph construction module is used to build a knowledge graph of flying knitting based on multi-dimensional historical flying knitting data and to mine the association rules between parameters and defects. The model building module is used to construct a bidirectional correlation reasoning model of defects and parameters based on the weaving knowledge graph and the association rules. The acquisition module is used to acquire real-time flyweaving data or target fabric parameters; The fusion module is used to perform feature fusion on the real-time flying weaving data or target fabric parameters to obtain global fusion features; The inference module is used to input the global fusion features into the bidirectional correlation inference model to obtain inference results; wherein, the inference results include at least one of defect prediction, parameter optimization and automatic process recommendation.

[0015] In a third aspect, this application provides an electronic device including a memory, one or more processors, and one or more application programs, wherein the one or more application programs are stored in the memory and are configured to, when invoked by the one or more processors, cause the one or more processors to perform the method as described in any one aspect.

[0016] In a fourth aspect, this application provides a computer-readable storage medium storing a plurality of program codes adapted to be loaded and run by a processor to perform the method as described in any one of the first aspects.

[0017] The above-described technical solutions of this application have at least one or more of the following beneficial effects: In implementing the technical solution of this application, a multi-dimensional knowledge graph of flyknitting is constructed to achieve unified integration and structured expression of all elements of data, including process parameters, raw material properties, pattern structure, equipment status, and defect information. This transforms scattered process experience into standardized, searchable, and reasonable digital knowledge, significantly improving the systematization, standardization, and reusability of flyknitting process knowledge. Based on the knowledge graph and parameters... The bidirectional correlation reasoning mechanism established by defect association rule mining can automatically complete intelligent defect prediction, rapid root cause location, and precise optimization of process parameters, significantly reducing reliance on human experience, effectively shortening the machine adjustment and trial weaving cycle, and improving the efficiency of style and material changes. At the same time, this application can perceive the production status in real time and adaptively recommend the optimal process solution, stabilize the consistency of the weaving process, reduce common defects such as missed stitches, skipped stitches, and uneven density, significantly reduce the defect rate, and improve the overall yield and production efficiency of fly knitting.

[0018] 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 this application. Attached Figure Description

[0019] The disclosure of this application 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 application. Furthermore, similar numbers in the drawings are used to denote similar components, wherein: Figure 1 This is a schematic flowchart illustrating the main steps of a knowledge graph-based flyknitting process reasoning method according to an embodiment of this application. Figure 2 This is a schematic flowchart of the main steps of a knowledge fusion process according to an embodiment of this application. Detailed Implementation

[0020] Some embodiments of this application are described below with reference to the accompanying drawings. Those skilled in the art should understand that these embodiments are merely illustrative of the technical principles of this application and are not intended to limit the scope of protection of this application.

[0021] Reference Figure 1 , Figure 1 This is a schematic flowchart illustrating the main steps of a knowledge graph-based reasoning method for flyknitting processes according to an embodiment of this application. The method mainly includes steps S101-S105: Step S101: Construct a knowledge graph of fly-weaving based on multi-dimensional historical fly-weaving data and mine the association rules between parameters and defects; Step S102: Construct a bidirectional association reasoning model of defects and parameters based on the knowledge graph of flyweaving and association rules; Step S103: Obtain real-time flyweaving data or target fabric parameters; Step S104: Perform feature fusion on real-time flyweaving data or target fabric parameters to obtain global fusion features; Step S105: Input the global fusion features into the bidirectional correlation reasoning model to obtain the reasoning results; wherein, the reasoning results include at least one of defect prediction, parameter optimization and automatic process recommendation.

[0022] The data for flyknitting includes: process parameters, raw material properties, pattern structure, and defect characteristics. These are various types of data from different dimensions of the flyknitting process, comprehensively reflecting the flyknitting production process and considering and determining production efficiency and quality from multiple perspectives. Specifically, parameters include needle pitch, yarn tension, knitting speed, and temperature control; raw material properties include yarn specifications, composition, elastic modulus, color fastness, coefficient of friction, and elongation; pattern structure includes weave structure, pattern design, density distribution, color configuration, area zoning, and pattern complexity; and defect characteristics include skipped stitches, uneven density, uneven fabric surface, and color difference. Flyknitting data can also include machine status data, such as machine type, needle position, operating status, and equipment parameters; and historical results data, such as process debugging records, quality inspection results, maintenance records, and good / defective product labels.

[0023] Among them, defect prediction refers to the input of current parameters, raw materials, or patterns, etc., and the output of defect probability and risk level by the bidirectional correlation reasoning model; parameter optimization refers to the bidirectional correlation reasoning model being able to reverse reason and locate suspicious parameters when defects occur, and can provide correction directions and values ​​(such as tension). 5%, yarn feed +3%); automatic process recommendation refers to inputting the target pattern or raw material, etc., and the two-way correlation reasoning model searches similar cases to deduce the optimal parameter combination, such as the combination of needle pitch, yarn feed, and tension.

[0024] Based on the above steps S101-S105, the method of this embodiment constructs a multi-dimensional knowledge graph about fly-knitting, realizing the unified integration and structured expression of all elements of data such as process parameters, raw material properties, pattern structure, equipment status, and defect information. This transforms scattered process experience into standardized, queryable, and reasonable digital knowledge, significantly improving the systematization, standardization, and reusability of fly-knitting process knowledge. Based on the knowledge graph and parameters... The bidirectional correlation reasoning mechanism established by defect association rule mining can automatically complete intelligent defect prediction, rapid root cause location, and precise optimization of process parameters, significantly reducing reliance on human experience, effectively shortening the machine adjustment and trial weaving cycle, and improving the efficiency of changing styles and materials. At the same time, this application can perceive the production status in real time and adaptively recommend the optimal process solution, stabilize the consistency of the weaving process, reduce common defects such as missed stitches, skipped stitches, and uneven density, significantly reduce the defect rate, and improve the overall yield and production efficiency of flyknitting. This application realizes the transformation of flyknitting from experience-driven to data-driven and intelligent decision-making, supports large-scale intelligent production in scenarios with multiple machine models, multiple raw materials, and multiple patterns, and promotes the upgrading of flyknitting manufacturing towards high efficiency, stability, and intelligence.

[0025] In one embodiment, step S101 specifically includes: Using key objects in the flying knitting data as entities and relationships such as cause, association, constraint, and mitigation, a flying knitting knowledge graph is constructed by extracting triples from multi-dimensional historical flying knitting data. This can be understood as abstracting key objects in the flyweaving process into entities, such as parameter entities, raw material entities, pattern structure entities, machine equipment entities, defect entities, and historical result entities; defining the logical relationships between entities: Cause relationships refer to the relationship between parameters / raw materials / patterns leading to defects, such as excessive tension causing skipped stitches, insufficient yarn feed causing thin density, improper needle spacing causing missed stitches, and mismatched raw material elasticity causing uneven fabric surface; Association relationships refer to the association between parameters, parameters and raw materials, and parameters and patterns; Constraint relationships refer to the restrictions of patterns on parameter ranges, and restrictions of raw materials on parameter ranges; Relief relationships refer to the fact that adjusting a parameter will eliminate a defect, such as reducing tension to relieve skipped stitches; a knowledge graph is constructed in the form of "entity-relationship-entity" triples, such as: excessive tension - leading to - skipped stitches, insufficient yarn feed - leading to - thin density, improper needle spacing - leading to - missed stitches, etc., forming a directed graph structure that supports path query and reasoning; Association rule mining algorithms are used to extract frequent itemsets from multi-dimensional historical flyweaving data, calculate the support and confidence of each parameter combination with defects, and screen out association rules between parameters with high confidence and defects.

[0026] Specifically, multi-dimensional historical flyknitting data can be preprocessed. Preprocessing operations include, but are not limited to, cleaning, deduplication, normalization, outlier removal, discretization or binning of continuous parameters, and standardized naming of defects. A parameter combination plus a defect result is considered a transaction, for example: {high tension, normal yarn feed, normal needle pitch, normal raw material} + {skipped stitches}; {insufficient yarn feed, normal tension, complex pattern} + {thin density}. Association rule mining algorithms (such as Apriori or FP-Growth) are used to automatically mine frequent itemsets from a large number of transactions, i.e., which parameter combinations frequently occur with which defects. The mining target is parameter condition set → defect, denoted as: X → Y. Support and confidence are calculated. Support represents the frequency of this rule in historical data; Support = number of transactions containing X and Y ÷ total number of transactions. Confidence represents the probability that Y occurs when X occurs; Confidence = number of transactions containing X and Y ÷ number of transactions containing X. Strong association rules are filtered out, and only those with support and confidence scores greater than or equal to a certain threshold are retained as the final standardized rules for output. Examples include: high tension → skipped stitches, insufficient yarn feed → thin density, improper stitch length → missed stitches, etc. These standardized rules are then stored in a unified format in the knowledge graph as the basis for inference. Furthermore, as new production data is continuously added, the rules are periodically re-mined and updated, with confidence scores automatically adjusted as data increases, removing low-confidence rules and adding high-confidence rules.

[0027] In a specific embodiment, step S102 specifically includes: assigning confidence scores to association rules and storing them in the knitting knowledge graph to form a bidirectional association reasoning model of defects and parameters. It can be understood that bidirectional means that parameters can indicate what defects might occur, and defects can be used to determine which parameters might cause them. Assigning confidence scores to association rules, such as high tension → skipped stitches (confidence score 0.92), allows the association reasoning model to select the rule with the highest confidence score as the basis for reasoning, and also provides the correction direction and value for parameter adjustment based on the confidence score.

[0028] In one embodiment, reference Figure 2 Step S104 specifically includes: Step S401: Based on the pre-trained process semantic coding model using corpus data specific to the flying knitting field, perform unified semantic coding on real-time flying knitting data to obtain a unified semantic feature vector; Step S402: Vectorize the entities and relationships in the weaving knowledge graph through knowledge embedding to obtain knowledge embedding vectors; Step S403: Adaptively weightedly fuse the unified semantic feature vector and the knowledge embedding vector using a multi-head attention mechanism to obtain the global fused feature.

[0029] Specifically, the pre-trained model is based on a corpus of language data specific to the flyknitting field. This corpus includes, but is not limited to, historical process parameter databases, defect case databases, pattern structure databases, raw material databases, and machine log databases. Using masked language modeling (MLM) and contrastive learning, a dedicated process semantic encoding model for flyknitting is fine-tuned on massive amounts of real flyknitting data. This model is used for process terminology normalization and parameter encoding. Defect correlation prediction, flower pattern Process matching and defect cause classification. The process semantic coding model converts the input discrete process data (needle type, raw material, pattern) into process tokens; normalizes and bins the continuous parameters (tension, speed, temperature); the pre-trained model extracts deep semantic features through a Transformer encoder; and performs feature alignment and dimensional unification on different modal information, outputting a fixed-dimensional, computable, and fusionable process semantic feature vector to achieve a unified representation of heterogeneous data.

[0030] Specifically, the knowledge embedding process involves mapping entities and relationships in the weaving knowledge graph to low-dimensional dense vectors; ensuring that similar processes and defects are close in distance in the vector space; and preserving the topological structure and semantic associations in the weaving knowledge graph to obtain knowledge embedding vectors, providing computable input for subsequent attention fusion and inference models.

[0031] Specifically, adaptive weighted fusion includes: concatenating unified semantic feature vectors and knowledge embedding vectors along the channel dimension to form multi-dimensional hybrid features; automatically learning the importance weights of features in different dimensions through a multi-head attention mechanism, assigning high weights to key features and low weights to noisy features, and adaptively weighting and fusing multi-dimensional features to output globally fused features. In other words, the multi-head attention mechanism calculation process involves the model automatically learning the importance of each dimension's features: calculating attention weights for process parameters, raw materials, patterns, defects, and knowledge graphs separately; automatically assigning high weights to key features affecting defects (such as tension and yarn feed) and low weights to secondary features; and concatenating and linearly transforming the attention outputs from different heads to achieve deep fusion of features such as process parameters, raw material attributes, pattern structure, and defect features, forming globally fused features that can be directly used for inference.

[0032] In one embodiment, step S105 specifically includes: If the global fusion features are derived from real-time flyweaving data, the bidirectional correlation inference model outputs defect prediction and parameter optimization. For example, real-time data collected during production includes: needle pitch, yarn feed, yarn tension, knitting speed, machine vibration, temperature, needle position status, etc. Defect localization requires combining raw material and pattern data: because skipped stitches, misplaced stitches, and uneven density may be caused by improper raw material elasticity, complex pattern structure, excessive yarn friction, and sudden changes in regional density. Therefore, the collected data, raw material data, and pattern data are input into the inference model together. Through multi-factor joint localization, the model can quickly distinguish whether the problem is a parameter issue, a raw material issue, or a pattern structure issue, and provide a precise correction solution. If the global fusion features originate from the target fabric parameters, the bidirectional association inference model automatically recommends the process. For example, in the production of flyknit shoe uppers, the inputs are target pattern data (weave structure, density distribution, regional zoning) and yarn raw material data (specifications, composition, elastic modulus, linear density). First, the pattern and raw material information are processed through data alignment and standardization. Then, they are converted into feature vectors by a pre-trained semantic encoding model. Process features are then fused through knowledge embedding and attention mechanisms. Subsequently, the process proceeds to the defect... The bidirectional correlation reasoning model of parameters, based on knowledge graph matching of similar historical cases, forward reasoning of risk-free optimal process range, and finally outputs a set of optimal process parameters such as needle pitch, yarn feed, yarn tension, and knitting speed, which greatly reduces manual trial knitting. When switching production between different flying knitting machines, it can also realize the rapid reuse of process knowledge and adaptive parameter matching.

[0033] In one embodiment, the method further includes: performing online verification based on real-time knitting data and quality inspection results, feeding the verification results back to the knitting knowledge graph and bidirectional association reasoning model, supporting incremental updates and cross-device knowledge transfer, and forming data collection. Feature encoding Knowledge Integration Reasoning and decision making Online verification Closed-loop optimization through feedback iteration.

[0034] Based on the above implementation methods, this application first standardizes and structures the collected process parameters, raw material properties, pattern structure, and defect-related data, transforming discrete process information into process tokens and combining them with numerical normalization. A pre-trained knitting process semantic encoding model is then used to complete unified feature representation, resulting in multi-dimensional semantic vectors. Next, entities and relationships in the knowledge graph are transformed into vector representations through knowledge embedding, and together with semantic features, they undergo adaptive weighted fusion through a multi-head attention mechanism to highlight key process-related features. Subsequently, based on the fused features, and relying on the pre-constructed knowledge graph and parameter-defect association rules mined from historical data, bidirectional path reasoning is performed: forward, defect risks are predicted based on current process conditions; backward, root cause parameters are located based on existing defects, and parameter adjustment directions and optimization suggestions are given based on rule confidence. Ultimately, intelligent defect prediction, automatic optimization of process parameters, and standard solution recommendation are achieved, forming an iteratively optimizable intelligent reasoning closed loop.

[0035] It should be noted that although the steps in the above embodiments are described in a specific order, those skilled in the art will understand that in order to achieve the effect of this application, different steps do not necessarily have to be executed in such an order. They can be executed simultaneously (in parallel) or in other orders, and these variations are all within the scope of protection of this application.

[0036] Furthermore, this embodiment provides a knowledge graph-based reasoning system for flyknitting processes, including: The graph construction module is used to build a knowledge graph of flying knitting based on multi-dimensional historical flying knitting data and to mine the association rules between parameters and defects. The model building module is used to construct a bidirectional correlation reasoning model of defects and parameters based on the flyweaving knowledge graph and association rules; The acquisition module is used to acquire real-time flyweaving data or target fabric parameters; The fusion module is used to perform feature fusion on real-time flyweaving data or target fabric parameters to obtain global fusion features; The inference module is used to input global fusion features into the bidirectional correlation inference model to obtain inference results; wherein, the inference results include at least one of defect prediction, parameter optimization and automatic process recommendation.

[0037] For specific implementation methods and effects, please refer to the above-mentioned specific implementation method of the knowledge graph-based flyweaving process reasoning method. Repeated parts will not be repeated.

[0038] Furthermore, this application also provides an electronic device, including a memory, one or more processors, and one or more application programs, wherein the one or more application programs are stored in the memory, and the one or more application programs are configured to cause the one or more processors to perform the method described in any of the preceding technical solutions when invoked by the one or more processors.

[0039] The electronic device in this application mainly includes a memory and a processor. The memory can be configured to store a program for executing the methods of the above method embodiments, and the processor can be configured to execute the program in the memory. The program includes, but is not limited to, a program for executing the methods of the above method embodiments. For ease of explanation, only the parts related to the embodiments of this application are shown. For specific technical details not disclosed, please refer to the method section of the embodiments of this application.

[0040] In the embodiments of this application, the electronic device may be a control device comprising various electronic components. In some possible implementations, the electronic device may include multiple storage devices and multiple processors. The program executing the method of the above method embodiments may be divided into multiple subroutines, each subroutine may be loaded and run by a processor to execute different steps of the method of the above method embodiments. Specifically, each subroutine may be stored in different memories, and each processor may be configured to execute programs in one or more memories to jointly implement the method of the above method embodiments, that is, each processor executes different steps of the method of the above method embodiments to jointly implement the method of the above method embodiments.

[0041] The aforementioned electronic equipment is used to perform Figure 1 The method embodiments shown are similar in technical principle, technical problem solved and technical effect produced. Those skilled in the art can clearly understand that, for the sake of convenience and brevity, the specific working process of the electronic device and related descriptions can be referred to the content described in the method embodiments, and will not be repeated here.

[0042] The aforementioned multiple processors can be processors deployed on the same device. For example, the aforementioned computer device can be a high-performance device composed of multiple processors, and the aforementioned multiple processors can be processors configured on that high-performance device. Alternatively, the aforementioned multiple processors can also be processors deployed on different devices. For example, the aforementioned computer device can be a server cluster, and the aforementioned multiple processors can be processors on different servers within the server cluster.

[0043] Those skilled in the art will understand that all or part of the processes in the method of the above-described embodiment can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable file, or some intermediate form. The computer-readable storage medium can include any entity or device capable of carrying the computer program code, a medium, a USB flash drive, a portable hard drive, a magnetic disk, an optical disk, a computer memory, a read-only memory, a random access memory, an electrical carrier signal, a telecommunication signal, and a software distribution medium, etc. It should be noted that the content included in the computer-readable storage medium can be appropriately added or removed according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, the computer-readable storage medium does not include electrical carrier signals and telecommunication signals.

[0044] Furthermore, this application also provides a computer-readable storage medium. In one embodiment of the computer-readable storage medium according to this application, the computer-readable storage medium can be configured to store a program that performs the methods of the above-described method embodiments, the program of which can be loaded and run by a processor to implement the above-described methods. For ease of explanation, only the parts related to the embodiments of this application are shown; for specific technical details not disclosed, please refer to the method section of the embodiments of this application. The computer-readable storage medium can be a storage device device comprising various electronic devices. Optionally, in the embodiments of this application, the computer-readable storage medium is a non-transitory computer-readable storage medium.

[0045] Furthermore, it should be understood that since the various modules are only provided to illustrate the functional modules of the device described in this application, the physical devices corresponding to these modules may be the processor itself, or a part of the processor's software, hardware, or a combination of both. Therefore, the number of modules shown in the figures is merely illustrative. Those skilled in the art will understand that the various modules in the system can be adaptively split or merged. Such splitting or merging of specific modules will not cause the technical solution to deviate from the principles of this application; therefore, the technical solutions after splitting or merging will fall within the protection scope of this application.

[0046] 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.

[0047] 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.

[0048] 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.

[0049] In this application, unless otherwise expressly specified and limited, the terms "installation," "connection," "joining," and "fixing," etc., should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral part; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; they can refer to the internal communication of two components or the interaction between two components, unless otherwise expressly limited. Those skilled in the art can understand the specific meaning of the above terms in this application according to the specific circumstances.

[0050] Although embodiments of this application have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting this application. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments within the scope of this application.

Claims

1. A knowledge graph-based reasoning method for flyweaving processes, characterized in that, The method includes: A knowledge graph of flyweaving was constructed based on multi-dimensional historical flyweaving data, and the association rules between parameters and defects were mined. Based on the aforementioned weaving knowledge graph and the aforementioned association rules, a bidirectional association reasoning model for defects and parameters is constructed. Obtain real-time flyweaving data or target fabric parameters; The real-time flying weaving data or target fabric parameters are fused to obtain global fused features; The global fusion features are input into the bidirectional correlation reasoning model to obtain the reasoning results; wherein, the reasoning results include at least one of defect prediction, parameter optimization and automatic process recommendation.

2. The knowledge graph-based reasoning method for flyweaving processes according to claim 1, characterized in that, The flyknitting data includes process parameters, raw material properties, pattern structure, and defect characteristics.

3. The knowledge graph-based reasoning method for flyweaving processes according to claim 1, characterized in that, The construction of a knowledge graph for knitting based on multi-dimensional historical knitting data and the mining of association rules between parameters and defects include: Using key objects in the flying knitting data as entities and relationships such as cause, association, constraint, and mitigation, a flying knitting knowledge graph is constructed by extracting triples from multi-dimensional historical flying knitting data. Association rule mining algorithms are used to extract frequent itemsets from multi-dimensional historical flyweaving data, calculate the support and confidence of each parameter combination with defects, and screen out association rules between parameters with high confidence and defects.

4. The knowledge graph-based reasoning method for flyweaving processes according to claim 1, characterized in that, The construction of a bidirectional association reasoning model for defects and parameters based on the knitting knowledge graph and the association rules includes: assigning confidence scores to the association rules and storing them in the knitting knowledge graph to form a bidirectional association reasoning model for defects and parameters.

5. The knowledge graph-based reasoning method for flyweaving processes according to claim 1, characterized in that, The step of fusing features from the real-time flyweaving data or target fabric parameters to obtain global fused features includes: Based on the pre-trained process semantic coding model using corpus data specific to the field of fly weaving, a unified semantic coding is performed on real-time fly weaving data to obtain a unified semantic feature vector; The entities and relationships in the weaving knowledge graph are vectorized by knowledge embedding to obtain knowledge embedding vectors; A multi-head attention mechanism is used to adaptively and weightfully fuse the unified semantic feature vector and the knowledge embedding vector to obtain a global fused feature.

6. The knowledge graph-based reasoning method for flyweaving processes according to claim 5, characterized in that, The step of vectorizing entities and relationships in the knitting knowledge graph through knowledge embedding to obtain knowledge embedding vectors includes: Map the entities and relationships in the aforementioned weaving knowledge graph into low-dimensional dense vectors; Make similar processes and similar defects close in distance in vector space; The topological structure and semantic associations in the aforementioned weaving knowledge graph are preserved to obtain the knowledge embedding vector.

7. The knowledge graph-based reasoning method for flyweaving processes according to claim 5, characterized in that, The multi-head attention mechanism is used to adaptively and weightedly fuse the unified semantic feature vector and the knowledge embedding vector to obtain a global fused feature, including: The unified semantic feature vector and the knowledge embedding vector are concatenated along the channel dimension to form a multi-dimensional hybrid feature. The importance weights of different dimensions of features are automatically learned through a multi-head attention mechanism. Key features are given high weights and noisy features are given low weights. Multi-dimensional features are adaptively weighted and fused to output global fused features.

8. The knowledge graph-based reasoning method for flyweaving processes according to claim 1, characterized in that, The step of inputting the global fusion features into the bidirectional correlation reasoning model to obtain the reasoning result includes: If the global fusion features are derived from real-time flyweaving data, the bidirectional correlation reasoning model outputs defect prediction and parameter optimization. If the global fusion features are derived from the target fabric parameters, the bidirectional correlation inference model will automatically recommend the output process.

9. The knowledge graph-based reasoning method for flyweaving processes according to any one of claims 1-8, characterized in that, The method further includes: Online verification is performed based on the real-time knitting data and quality inspection results. The verification results are then fed back to the knitting knowledge graph and the bidirectional association reasoning model for closed-loop optimization.

10. A knowledge graph-based reasoning system for knitting processes, characterized in that, include: The graph construction module is used to build a knowledge graph of flying knitting based on multi-dimensional historical flying knitting data and to mine the association rules between parameters and defects. The model building module is used to construct a bidirectional correlation reasoning model of defects and parameters based on the weaving knowledge graph and the association rules. The acquisition module is used to acquire real-time flyweaving data or target fabric parameters; The fusion module is used to perform feature fusion on the real-time flying weaving data or target fabric parameters to obtain global fusion features; The inference module is used to input the global fusion features into the bidirectional correlation inference model to obtain inference results; wherein, the inference results include at least one of defect prediction, parameter optimization, and automatic process recommendation.

Images