Lace fabric incoming material quality evaluation method based on data analysis

Abstract

The invention discloses an incoming lace fabric quality evaluation method based on data analysis, and belongs to the technical field of quality management, and the method specifically comprises the steps: receiving a process parameter change list, querying a pre-constructed lace fabric process knowledge graph, and recognizing the quality attribute of an incoming lace fabric influenced by the changed process parameters; generating a predictive detection scheme based on the identified quality attribute type of the lace fabric; retrieving the stock information of all batches of lace fabrics in the current warehouse, and determining the batch range of the lace fabrics needing to be detected; starting predictive detection operation on the determined lace fabric batch, and executing quality data acquisition according to detection parameters in a predictive detection scheme; integrating the collected quality data with the lace fabric batch information to generate a quality evaluation record; according to the method, the fundamental transformation of the supplied material quality evaluation of the lace fabric from passive response to active prediction is realized.

Description

Technical Field

[0001] This invention relates to the field of quality management technology, specifically to a data analysis-based method for assessing the quality of incoming lace fabrics. Background Technology

[0002] In the textile, apparel, and home decoration industries, the quality of incoming lace fabrics directly determines the appearance, durability, and market acceptance of downstream finished products. With the penetration of intelligent manufacturing technology into the textile field, the production process parameters of lace fabrics need to be dynamically adjusted according to order demands, and the number of batches in warehouse inventory is increasing. The traditional quality assessment model that relies on manual experience can no longer meet the needs of mass production for real-time and accurate assessment. The industry urgently needs to build a data-driven, systematic incoming material quality assessment system. Currently, a basic process framework for assessing the quality of incoming lace fabrics has been established: downstream production systems record changes in process parameters, warehouse management systems store batch inventory data, and the testing process uses standardized equipment to collect data such as surface images and physical properties. Some companies have also established quality standard libraries to achieve preliminary comparisons between test data and standard thresholds. These methods can complete basic quality screening in scenarios with stable process parameters and small batches of inventory, ensuring the smoothness of the production process to a certain extent.

[0003] However, existing assessment schemes are essentially passive and reactive, making proactive prevention difficult and exhibiting key limitations: First, when faced with downstream process adjustments, the system can only passively receive change information and cannot proactively assess the potential impact of such changes on raw material quality. It must rely on manual decisions regarding testing, which not only risks overlooking critical quality risks due to insufficient human experience but also leads to a failure to keep pace with process changes in quality prediction. Second, most testing schemes are fixed and cannot flexibly adjust the level of detail and sampling quantity based on the impact of process changes on quality. This results in either inadequate testing for changes with significant impact or unnecessary testing for changes with minor impact, failing to adapt to the actual needs of different process change scenarios. Summary of the Invention

[0004] The purpose of this invention is to provide a data analysis-based method for evaluating the quality of incoming lace fabrics, thereby solving the following technical problems: The existing quality assessment scheme for incoming lace fabrics is essentially a passive response. When downstream processes are adjusted, the system only passively receives change information and requires manual decision-making on whether to conduct tests. This easily leads to the omission of key quality risk points and delayed quality prediction. Furthermore, the testing scheme is fixed and cannot flexibly adjust the level of detail and sampling quantity based on the impact of process changes on quality.

[0005] The objective of this invention can be achieved through the following technical solutions: A data analysis-based method for assessing the quality of incoming lace fabric includes the following steps: S1. Receive a process parameter change list, which includes the changed process parameter items and their parameter values. S2. Based on the list of process parameter changes, query the pre-constructed lace fabric process knowledge graph to identify the incoming lace fabric quality attributes affected by the changed process parameters. S3. Based on the identified lace fabric quality attribute types, generate a predictive detection scheme, which includes a set of quality attributes to be detected and corresponding detection parameters. S4. Based on the predictive testing scheme, retrieve the inventory information of all batches of lace fabric in the current warehouse to determine the range of lace fabric batches that need to be tested. S5. Initiate predictive testing for the identified batch of lace fabric and collect quality data according to the testing parameters in the predictive testing plan. S6. Integrate the collected quality data with the batch information of the lace fabric to generate a quality assessment record.

[0006] As a further aspect of the present invention: in step S1, the specific process of receiving the process parameter change list is as follows: After the downstream production management system completes the process parameter configuration update, it automatically generates a process parameter change notification file. The process parameter change notification file adopts a structured text format, which includes a change timestamp, the operator's identifier, and a list of changed process parameter item names. For each changed process parameter item, the parameter value before the change and the parameter value after the change are recorded. The system receives the process parameter change notification file through the file transfer service, parses the file content and verifies the data integrity, and stores the parsed changed process parameter items and their parameter values ​​as a process parameter change list. The process parameter change list maintains a data table structure in memory.

[0007] As a further aspect of the present invention: in step S2, the specific process of querying the pre-constructed lace fabric process knowledge graph is as follows: The lace fabric technology knowledge graph uses a graph structure of nodes and edges for data storage. Node types include process parameter nodes and lace fabric quality attribute nodes. Edge types represent the influence relationship between process parameter nodes and lace fabric quality attribute nodes. Each edge contains an influence intensity coefficient and an influence direction identifier. Based on the name of the changed process parameter item in the process parameter change list, locate the corresponding process parameter node in the knowledge graph, traverse all outgoing edges from the process parameter node, obtain the set of directly connected lace fabric quality attribute nodes, calculate the influence quantification value of each lace fabric quality attribute node based on the influence intensity coefficient in the edge attribute and the change range of the process parameter value, mark the lace fabric quality attribute nodes whose influence quantification value exceeds the activation threshold as affected attributes, and form a list of lace fabric incoming material quality attributes.

[0008] As a further aspect of the present invention: in step S3, the specific process of generating the predictive detection scheme is as follows: Read each lace fabric quality attribute identifier from the incoming lace fabric quality attribute list, query the testing rule library, obtain the standard testing method number and benchmark testing parameters corresponding to each lace fabric quality attribute identifier, and adaptively adjust the benchmark testing parameters based on the influence quantification value calculated from the lace fabric process knowledge graph. The testing parameter adjustment includes increasing the sampling point density, improving the testing accuracy level, and extending the testing observation time. Bind the adjusted testing parameters to the lace fabric quality attribute identifier to form testing task items. All testing task items are combined into a predictive testing scheme, which includes a scheme number, generation timestamp, and a list of testing task items.

[0009] As a further aspect of the present invention: the specific process of adaptively adjusting the benchmark detection parameters is as follows: The adjustment level of the detection parameters is determined based on the magnitude of the influence quantification value. The influence quantification value is divided into multiple continuous intervals, and each interval corresponds to an adjustment level. The adjustment levels of the detection parameters include first-level adjustment, second-level adjustment, and third-level adjustment. Each adjustment level corresponds to a set proportion of the number of detection sampling points and a set level of detection accuracy.

[0010] As a further aspect of the present invention: in step S4, the specific process of determining the batch range of lace fabrics requiring testing is as follows: The warehouse management system's lace fabric inventory database is queried to retrieve inventory records for all in-stock lace fabric batches. These records include the batch number, arrival time, storage location, and fabric composition specifications. Based on the lace fabric quality attributes involved in the predictive testing scheme, and combined with the lace fabric composition specifications, the correlation score between each lace fabric batch and the testing scheme is calculated. A correlation score threshold is set, and lace fabric batches with correlation scores greater than or equal to the threshold are filtered out to form the range of lace fabric batches that need to be tested. The lace fabric batch range is then sorted from highest to lowest correlation score.

[0011] As a further aspect of the present invention: the specific process for calculating the correlation score is as follows: Construct a correlation weight matrix between the composition specifications and quality attributes of lace fabric. The rows of the matrix correspond to different types of lace fabric composition specifications, the columns of the matrix correspond to different lace fabric quality attributes, and the weight values ​​in the matrix represent the sensitivity of the lace fabric composition specifications to the quality attributes. For each batch of lace fabric, its lace fabric composition and specification type are extracted, and the corresponding row vector is located in the weight matrix. According to the list of lace fabric quality attributes involved in the predictive detection scheme, the corresponding weight values ​​are extracted from the row vectors. All extracted weight values ​​are weighted and summed. The weight is the quantified value of the influence of each quality attribute in the predictive detection scheme. The weighted sum is normalized to obtain the correlation score of the lace fabric batch.

[0012] As a further aspect of the present invention: in step S5, the specific process of initiating the predictive detection operation is as follows: Based on the batch range of the lace fabric, a testing task assignment order is generated. The testing task assignment order includes a list of lace fabric batch numbers and a copy of the predictive testing plan corresponding to each batch. The testing task assignment order is distributed to the automated testing equipment control center. The automated testing equipment control center parses the testing task assignment order, allocates testing equipment resources, and the testing equipment performs quality data acquisition according to the testing parameters in the predictive testing plan. The acquisition process includes acquiring surface images of the lace fabric, testing physical properties, and analyzing chemical composition. The acquired raw data is processed through format standardization and converted into a quality data file with a unified structure.

[0013] As a further aspect of the present invention, the specific process for generating the quality assessment record is as follows: Read the quality data file corresponding to each batch of lace fabric, extract the quality index measurement values, compare the quality index measurement values ​​with the allowable range in the lace fabric quality standard library, and generate a quality assessment conclusion for a single batch of lace fabric. The quality assessment conclusion includes the pass / fail result and abnormal index remarks. Summarize the quality assessment conclusions of all batches of lace fabric to form a total quality assessment record. The total quality assessment record includes the assessment date, the number of lace fabric batches involved, the overall pass rate statistics, and a detailed assessment data table.

[0014] The beneficial effects of this invention are: This invention accurately identifies affected incoming material quality attributes by automatically parsing the downstream process parameter change list and querying the lace fabric process knowledge graph. Based on the quantified impact value, it dynamically generates predictive testing schemes, effectively solving the queuing delay problem caused by the need to re-initiate testing tasks due to process adjustments in traditional quality assessment processes. By constructing a correlation weight matrix between lace fabric composition specifications and quality attributes, and calculating the correlation score between each batch and the testing scheme, intelligent screening of the testing scope is achieved, avoiding the waste of resources from testing all batches. The system adaptively optimizes the sampling point density and testing accuracy requirements according to the adjustment level of testing parameters, and performs standardized data collection through automated equipment, ensuring the consistency and reliability of testing results. Finally, by automatically integrating quality data and batch information to generate assessment records, a closed-loop management system from process changes to quality feedback is formed, significantly improving the timeliness and accuracy of incoming lace fabric quality assessment and providing timely and reliable data support for downstream process adjustments. Attached Figure Description

[0015] The invention will now be further described with reference to the accompanying drawings.

[0016] Figure 1 This is a flowchart illustrating the present invention. Detailed Implementation

[0017] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0018] Please see Figure 1 As shown, this invention is a data analysis-based method for assessing the quality of incoming lace fabric, comprising the following steps: S1. Receive a process parameter change list, which includes the changed process parameter items and their parameter values. S2. Based on the list of process parameter changes, query the pre-constructed lace fabric process knowledge graph to identify the incoming lace fabric quality attributes affected by the changed process parameters. S3. Based on the identified lace fabric quality attribute types, generate a predictive detection scheme, which includes a set of quality attributes to be detected and corresponding detection parameters. S4. Based on the predictive testing scheme, retrieve the inventory information of all batches of lace fabric in the current warehouse to determine the range of lace fabric batches that need to be tested. S5. Initiate predictive testing for the identified batch of lace fabric and collect quality data according to the testing parameters in the predictive testing plan. S6. Integrate the collected quality data with the batch information of the lace fabric to generate a quality assessment record.

[0019] In step S1, the specific process of receiving the process parameter change list is as follows: The downstream production management system has a built-in parameter configuration update monitoring module that monitors the modification status of the internal process parameter configuration files in real time. When an operator adjusts process parameters through the parameter configuration interface of the production management system and submits the changes, the configuration update monitoring module captures the write completion event of the configuration file and triggers the process parameter change notification file generation process. The system calls a preset file generation interface, which constructs the file content according to the structured text format specification to ensure the standardization and parsability of the data organization.

[0020] In the structured text format, the change timestamp field strictly adheres to the ISO8601 standard format, accurately recording the completion time of the parameter change operation and providing a time reference for subsequent traceability. The change operator identifier field is filled with a unique user identification code within the system. This code is bound to the operator's authorized account and can be linked to query the specific person responsible for the operation. The list of changed process parameter item names exists in the form of a string array. Each element in the array corresponds to a standard name of the changed process parameter item, and the name is consistent with the parameter item naming in the system's process parameter database. For each changed process parameter item in the list, the file allocates a separate data segment to record the parameter value before and after the change. The value type strictly matches the parameter item's definition type. For example, numeric parameters are stored as floating-point or integer types, and enumerated parameters are stored as preset enumerated value strings, ensuring the accuracy of the parameter values ​​and the consistency of data types.

[0021] After the process parameter change notification document is generated, the downstream production management system sends it to the corresponding quality assessment system of this invention via a file transfer service. The file transfer service is implemented based on an industrial-grade data transmission protocol, with common protocols including SFTP, FTP, or HTTP-based RESTful file transfer interfaces. A data verification mechanism is enabled during transmission by calculating the file's CRC32 checksum or MD5 hash value and transmitting the checksum along with the file for the receiving end to verify file integrity. The file receiving module of the quality assessment system continuously monitors the designated transmission port or file receiving directory. When a new file arrives, the receiving process is automatically triggered. First, data integrity is verified by comparing the file checksums at the transmitting and receiving ends. If the verification fails, a retransmission request is sent to the downstream production management system; if the verification passes, the file parsing process is initiated.

[0022] The file parsing process is executed by a structured text parser. The parser calls the corresponding parsing algorithm based on the file's format type, extracting the change timestamp, operator identifier, list of changed process parameter names, and the values ​​of each parameter before and after the change, field by field. During parsing, data validity checks are performed simultaneously, including field missing checks (ensuring no core fields are omitted), data type checks (verifying that the numerical types of each field match preset definitions), and parameter value validity checks (checking whether the parameter values ​​before and after the change are within the allowed range for that process parameter). If a validity issue is found, a parsing exception log is generated, and the process is paused for manual investigation and handling. If both parsing and validation pass, the extracted changed process parameters and their corresponding values ​​before and after the change, change timestamps, operator identifiers, etc., are stored as a process parameter change list.

[0023] The process parameter change list is maintained in memory as a two-dimensional data table structure. The data table is implemented using an in-memory database or a dynamic array. The table structure includes fixed columns: parameter item ID (a unique identifier associated with the system's process parameter database), parameter item name, parameter value before change, parameter value after change, change timestamp, and operator ID. The data table uses the parameter item ID as the primary key index, constructing a hash index structure to support subsequent modules in quickly querying and retrieving change information for specific parameter items. Simultaneously, the data table has a data expiration mechanism, retaining only the parameter change data required for the current evaluation process to avoid excessive memory resource consumption.

[0024] In step S2, the specific process of querying the pre-constructed lace fabric technology knowledge graph is as follows: The lace fabric technology knowledge graph is built on a graph database and uses a node-edge graph structure model for data storage. The graph database's storage engine manages the relationships between nodes and edges through adjacency lists or adjacency matrices, ensuring efficient query performance. Nodes in the knowledge graph are divided into two categories: process parameter nodes and lace fabric quality attribute nodes. Both types of nodes contain basic and extended attributes: Basic attributes of process parameter nodes include parameter ID, parameter name, parameter type (e.g., numeric, enumerated, Boolean), and parameter value range; extended attributes include a description of the parameter's technological function and common adjustment scenarios. Basic attributes of lace fabric quality attribute nodes include quality attribute ID, quality attribute name, attribute description (e.g., "fabric tensile strength," "pattern clarity," etc.), and standard threshold range (a value range conforming to industry or company quality standards); extended attributes include the quality attribute's testing method identifier and associated downstream finished product quality requirements.

[0025] The edges in the knowledge graph are used to represent the influence relationship between process parameter nodes and lace fabric quality attribute nodes. Each edge contains fixed attributes: influence intensity coefficient (within the range of 0-1, the larger the value, the higher the degree of influence of the process parameter on the corresponding quality attribute, the coefficient value is determined based on historical experimental data, industry standard documents and expert experience values), influence direction identifier (enumerated types, including "positive influence", "negative influence" and "bidirectional influence", which respectively indicate that the quality attribute value increases, decreases, and presents different influence directions according to different parameter value ranges), and the edge also contains the source identifier of the influence relationship (such as "experimental data derivation", "expert experience setting", "industry standard provision") and update timestamp, which are used to trace the validity and timeliness of the influence relationship.

[0026] Once the quality assessment system obtains the process parameter change list, it invokes the knowledge graph query module to initiate the related query process. The query module first extracts the names of the changed process parameter items from the change list, using them as query keywords. It then calls the node query interface of the graph database. The interface performs a precise query by matching the "parameter item name" attribute of the process parameter node. If multiple process parameter nodes with the same name but different parameter types or value ranges exist, the query module further compares the value ranges of the parameter items before and after the change in the change list with the node's "parameter value range" attribute to filter out the uniquely matching process parameter node, thus completing node location. If no matching node is found, a node mismatch log is generated, indicating that the knowledge graph data needs to be supplemented and improved.

[0027] After locating the process parameter node, the query module calls the edge traversal interface of the graph database to initiate the outgoing edge traversal process starting from the target process parameter node. The traversal algorithm adopts a breadth-first traversal strategy, prioritizing the acquisition of all outgoing edges directly connected to the process parameter node to avoid interference from irrelevant data due to excessive traversal depth. During the traversal, the interface returns the attribute information (influence strength coefficient, influence direction identifier) ​​of each outgoing edge in real time, as well as the terminal node (i.e., the lace fabric quality attribute node) connected to the outgoing edge. The query module temporarily stores this information in the memory cache, forming a related data set of "process parameter node - edge - quality attribute node".

[0028] Subsequently, the query module calculates the quantified impact value of each lace fabric quality attribute node based on the associated data set in the cache. The calculation process is based on the variation range of the process parameter values. The variation range is determined by the relative change relationship between the "parameter values ​​after the change" and the "parameter values ​​before the change" in the process parameter change list (such as the absolute difference or ratio of parameter values, and the specific calculation logic is preset according to the parameter type). Then, combined with the edge's influence intensity coefficient, the quantified impact value is generated through preset logical rules (such as the variation range and the influence intensity coefficient being positively correlated, and the two jointly determining the magnitude of the quantified impact value), ensuring that the quantified result can objectively reflect the actual degree of impact of process parameter changes on quality attributes.

[0029] The query module has a built-in activation threshold configuration unit. The activation threshold is set based on the enterprise's process quality requirements and historical quality risk data. Different types of quality attribute nodes can correspond to different activation thresholds (e.g., the activation threshold for critical quality attributes is lower, and the activation threshold for non-critical quality attributes is higher). After calculating the quantified impact value of each quality attribute node, the query module compares it with the activation threshold corresponding to that node. If the quantified impact value exceeds the activation threshold, the quality attribute node is marked as an "affected attribute," and core information such as the node's quality attribute ID, quality attribute name, and standard threshold range is extracted. If the value does not exceed the activation threshold, the quality attribute is determined to be unaffected by the current process parameter change and is not included in subsequent processes.

[0030] Finally, the query module sorts all quality attribute nodes marked as "affected attributes" by quality attribute ID and organizes them into a list of incoming lace fabric quality attributes. This list is stored in JSON array format, with each element containing fields such as quality attribute ID, quality attribute name, impact quantification value, and impact direction identifier. This facilitates direct access by the S3 predictive testing solution generation module. Simultaneously, the list is also synchronously stored in the system's temporary data directory and retained until the current batch's quality assessment process concludes, allowing for subsequent traceability and querying.

[0031] In step S3, the specific process of generating the predictive detection scheme is as follows: First, the process of reading the quality attribute identifier of the incoming lace fabric is initiated, calling the system's structured data reading interface to access the quality attribute list output by the S2 stage. This list is stored in a dedicated temporary data area of ​​the system, configured with a role-based access control policy, allowing only the current module to read the data through an encrypted authorized service link, preventing unauthorized tampering or access. The list is organized in a standardized structured format, with each entry containing core information such as the lace fabric quality attribute identifier, the impact quantification value, and the impact direction identifier. The reading interface uses streaming parsing technology to extract the quality attribute identifier line by line, avoiding memory resource overload caused by loading the full data, and improving reading efficiency through data fragmentation.

[0032] During the reading process, dual validity checks are performed simultaneously: First, the coding compliance of the quality attribute identifier is checked. This identifier must follow the fixed-length coding rules set by the enterprise's process standards. The code consists of a combination of letters and numbers to ensure uniqueness throughout the system. If an abnormal coding format is detected, the module immediately generates a verification exception log, clearly marking the location of the exception identifier and the type of violation. Second, the validity of the quantified value is checked. It must be confirmed that the quantified value is not empty and is within the reasonable value range preset by the system. If there is an empty value or an out-of-limit value, the module triggers an alarm mechanism to send an exception notification to the administrator terminal, suspending the reading process until the abnormal data is corrected and restarted.

[0033] After extracting the quality attribute identifiers, the module initiates a related search by calling the detection rule base query interface. The detection rule base is built on an enterprise-level relational database and adopts a storage architecture partitioned by quality attribute category. The core data table contains fields such as quality attribute identifier, standard detection method number, benchmark sampling point density, benchmark detection accuracy level, benchmark detection observation time, detection method applicable scope, and rule update timestamp. The quality attribute identifier serves as the primary key for indexing, ensuring efficient queries. The query interface employs an index-driven retrieval mechanism, using the quality attribute identifier as the search condition and quickly locating the target data through the database primary key index, avoiding delays caused by full table scans.

[0034] After the query results are returned, an integrity check is performed: If a quality attribute identifier does not match the corresponding testing rule, the interface automatically generates a rule missing report, which includes the quality attribute identifier of the missing rule and the associated process parameter information. This report is sent to the administrator via the system message push component, and the administrator can then query again after supplementing the rule data according to the enterprise's testing standards. If the results are complete, the interface extracts the standard testing method number and benchmark testing parameters into the module's dedicated memory cache. The cache uses a key-value pair mapping structure and establishes an associated index with "quality attribute identifier - testing rule data" to provide fast data support for subsequent parameter adjustments.

[0035] The module then adaptively adjusts the benchmark testing parameters based on the quantified impact values ​​in the process knowledge graph. Adjustments include increasing sampling point density, improving testing accuracy, and extending the testing observation time. After adjustment, the module generates a unique testing task ID for each quality attribute identifier. This ID uses a globally unique identifier generation mechanism to ensure traceability. The adjusted testing parameters are then bound to the quality attribute identifier and the standard testing method number to construct a testing task item containing the task ID, quality attribute identifier, standard testing method number, adjusted parameters, and task creation timestamp.

[0036] After all task items are generated, a consistency check is performed: This check verifies whether there are duplicate task items with the same quality attribute identifier, removing redundant entries by deduplication using ID; it also verifies the matching between adjusted parameters and the quantified values ​​affecting the results, ensuring the parameter adjustment range is logically sound; and it verifies the validity of the standard testing method number, comparing it with the database encoding to confirm usability. Task items that pass the check are arranged in ascending order by ID to form a task item list, which is then combined into a predictive testing scheme. The scheme includes core elements such as a scheme number (globally unique), a generation timestamp (ISO8601 standard format), and a list of testing task items. After generation, it is written to the scheme database and synchronized to the high-frequency access cache, simultaneously triggering a "scheme ready" event, which notifies the S4 stage via a message queue to initiate the batch range determination process.

[0037] The specific process of adaptively adjusting the benchmark detection parameters is as follows: The adjustment level of the detection parameters is determined based on the magnitude of the impact quantification value. The impact quantification value is divided into multiple continuous intervals. The interval division comprehensively refers to the historical detection data of lace fabric production, quality risk assessment reports, and the technical experience of industry experts. Multiple continuous intervals are set and can be dynamically adjusted by the administrator through the system backend. The quantification value falling into the first interval corresponds to the first level of adjustment, falling into the second interval corresponds to the second level of adjustment, and falling into the third interval corresponds to the third level of adjustment.

[0038] Each adjustment level corresponds to a preset parameter adjustment rule, which is stored in a standardized text format configuration file, specifying the exact adjustment method: Level 1 Adjustment: Applicable to scenarios with minimal impact on quantification values. The density of benchmark sampling points is increased according to a preset system ratio, with the ratio set to balance detection accuracy and cost. The benchmark detection accuracy level is increased by one grade. Accuracy levels are divided into basic, standard, precision, and ultra-high precision levels according to enterprise standards. Increasing one grade switches to the adjacent higher level. The benchmark detection observation time is extended according to a preset duration, with the duration based on capturing the basic changes in quality attributes.

[0039] Level 2 Adjustment: Applicable to scenarios with moderate impact on quantification values, with a higher adjustment range than Level 1. The density of benchmark sampling points is increased by a higher preset ratio to increase the amount of sampling data and enhance the reliability of the results; the benchmark detection accuracy level is increased by two levels to further reduce detection errors; the benchmark detection observation time is extended by a longer preset duration to ensure complete recording of the dynamic changes in quality attributes.

[0040] Level 3 Adjustment: Applicable to scenarios with significant impact on quantification values, with the highest adjustment range. The density of benchmark sampling points is increased according to the maximum preset ratio allowed by the system, and strictly controlled within the maximum sampling density range supported by the detection equipment hardware; the benchmark detection accuracy level is upgraded to the highest level (ultra-high precision level) to obtain the highest accuracy data; the benchmark detection observation time is extended to the maximum preset duration to fully cover the possible range of changes in quality attributes.

[0041] In step S4, the specific process for determining the range of lace fabric batches requiring testing is as follows: The process begins by integrating with the warehouse management system, establishing a communication link with the lace fabric inventory database through an industrial-grade data exchange protocol. This communication link employs an encrypted transmission mechanism to ensure data security and integrity during transmission. Simultaneously, authentication and permission verification through the warehouse management system are required to obtain read access to inventory data, allowing access only to inventory records related to lace fabrics to prevent unauthorized access to data from other product categories.

[0042] The inventory database adopts a relational database architecture. Core data tables are stored in partitions by year and quarter to optimize performance for large-volume data queries. The table structure includes fields such as batch number (a unique identifier automatically generated by the system, consisting of the production year, production line number, and serial number), arrival time (accurate to the minute, following the ISO8601 standard format), storage location (using a three-level coding rule of "warehouse area-shelf-location"), and fabric composition specifications (including key information such as fiber type percentage and yarn count). When executing a query, a pre-defined structured query statement is invoked, using "in-stock status" as the filter condition to extract inventory records for all in-stock lace fabric batches. The query statement uses an index optimization mechanism to link the batch number and storage status fields, avoiding query delays caused by full table scans.

[0043] After data extraction is completed, the inventory record preprocessing process is initiated: First, data integrity verification is performed to check each record for issues such as missing core fields (if the batch number or fabric composition specification is missing, it is marked as an abnormal record) and incorrect data format (such as the entry time not conforming to the standard format). Abnormal records are logged and temporarily stored in the abnormal data buffer until they are manually checked and supplemented before being reintroduced into the processing flow. Second, data standardization processing is performed to convert unstructured descriptions in the fabric composition specification (such as "high cotton content") into standardized numerical formats (such as "85% cotton") to ensure the comparability of composition specification data for different batches. The standardization rules are formulated based on the company's fabric composition coding specifications and can be dynamically updated through the system management backend.

[0044] After preprocessing, the correlation score calculation interface is invoked to associate the fabric composition specifications of each batch in the database with the quality attributes of the lace fabric involved in the predictive testing scheme, triggering the correlation score calculation process. Once the correlation scores for all batches are calculated, the system's preset correlation score threshold is read. This threshold is set by the quality control department based on historical testing data and quality risk assessment results, and can be dynamically adjusted according to fabric category or testing scenario. The correlation score of each batch is compared with the threshold, and batches with scores greater than or equal to the threshold are selected to form the initial set of testing batches.

[0045] Subsequently, the initial batch set of tests is sorted from highest to lowest based on the correlation score. A stable sorting strategy is employed to ensure that batches with the same score are arranged in reverse chronological order of their entry time, prioritizing batches with more recent entry times. After sorting, the final range of tests is output in a structured data format, including fields such as batch number, correlation score, entry time, and storage location, and is simultaneously stored in the test task scheduling database. A batch range confirmation log is also generated, recording key information such as screening conditions, sorting rules, and threshold values, providing a basis for subsequent process traceability.

[0046] The specific process for calculating the correlation score is as follows: A correlation weight matrix between the composition specifications and quality attributes of lace fabrics was constructed. This matrix was built based on lace fabric production process data, historical quality test results, and industry expert experience. The row dimension of the matrix corresponds to all defined lace fabric composition specification types, including fiber type combinations (such as cotton-polyester blends, cotton-spandex blends), yarn count ranges (such as 20S-30S, 30S-40S), weaving density grades, etc.; the column dimension corresponds to all lace fabric quality attributes, including tensile strength, pattern clarity, color fastness, elastic recovery rate, etc.

[0047] The weight values ​​in the matrix are determined through multi-dimensional data fusion: First, the fluctuation range of quality attributes under different component specifications in historical testing data is statistically analyzed. The greater the fluctuation range, the higher the sensitivity of the component specification to quality attributes, and the higher the corresponding weight value. Second, industry experts and enterprise process engineers are invited to calibrate the weight values, adjusting for weight deviations caused by extreme data based on actual production experience. Third, the weight values ​​are periodically iterated and updated based on newly added production and testing data, with the update cycle synchronized with the enterprise's process optimization cycle. The weight matrix is ​​stored in a matrix database, supporting efficient row vector retrieval and numerical updates. A mapping table is also established between component specification types and row indexes, and between quality attributes and column indexes, improving data location efficiency.

[0048] When calculating the correlation score for a particular batch, the fabric composition and specification type of that batch is first extracted from the pre-processed inventory data. The corresponding row vector is then located in the weight matrix using a composition specification-row index mapping table. Subsequently, the list of lace fabric quality attributes from the predictive detection scheme is read, and the column index of each attribute in the list is determined in the matrix using a quality attribute-column index mapping table. The weight value of the corresponding column index is then extracted from the located row vector to form the set of weight values ​​for that batch.

[0049] The influence quantification values ​​corresponding to each quality attribute are obtained from the predictive detection scheme. These values ​​are then used as weight coefficients and matched with corresponding elements in the aforementioned weight value set to initiate a weighted summation process. During the summation, the product of each weight value and its corresponding influence quantification value is accumulated to obtain the original summation result. Subsequently, the original summation result is normalized. The normalization range is set to 0-1, and a linear normalization strategy is used to map the original result to this range, ensuring that the correlation scores of different batches have a unified comparison benchmark.

[0050] After normalization, the obtained correlation scores are validated to check if they fall within the 0-1 range. If a score exceeds this range, it is considered a calculation anomaly, and the weight values ​​and influence quantification values ​​are retrieved for a second calculation. The validated correlation scores are then linked to their corresponding batch numbers, stored in the correlation score result database, and a calculation log is generated. This log records information such as the matrix version, weight value set, influence quantification value set, original results, and normalization parameters, ensuring the traceability of the calculation process.

[0051] In step S5, the specific process of initiating the predictive detection operation is as follows: First, a work order for testing is generated based on the determined batch range of lace fabric. The work order is constructed using a structured data format. Its core content includes a list of lace fabric batch numbers and a copy of the predictive testing plan for each batch. The batch number list is sorted in descending order of correlation score to ensure that high-priority batches enter the testing process first. The copy of the testing plan retains all fields of the original plan, including the plan number, adjusted testing parameters (sampling point density, testing accuracy level, observation time), and standard testing method number, avoiding deviations in testing operations due to missing plan information. When the work order is generated, a task identifier, generation timestamp, and task priority fields are added simultaneously. The task identifier uses a globally unique coding rule, consisting of the testing date and task sequence code, facilitating subsequent process tracking. The task priority is set according to the batch correlation score; the higher the score, the higher the priority, supporting the priority scheduling of high-priority tasks by testing equipment.

[0052] After a work order is generated, it is distributed to the control center of the automated testing equipment via an industrial-grade data transmission protocol. Before transmission, the work order data is compressed to reduce bandwidth consumption; a data checksum is also attached. Upon receiving the data, the control center verifies its integrity using the checksum. If verification fails, a retransmission request is sent to the sender until the data is fully received. Furthermore, a transmission log is recorded during the work order distribution process, including sending time, receiving time, and transmission status, providing a basis for subsequent troubleshooting of transmission anomalies.

[0053] After receiving a work order, the automated testing equipment control center initiates the parsing process. First, a syntax check is performed to verify that the work order's structured format conforms to preset specifications, such as missing fields or incorrect data types. If syntax errors are found, the work order is marked as invalid, and a parsing exception log is generated. Second, a completeness check is performed to confirm that the copy of the testing plan for each batch contains all necessary parameters. If the sampling point density is not specified, the work order is deemed incomplete, parsing is paused, and relevant personnel are notified to supplement the information. After successful parsing, the control center extracts the testing requirements (such as surface image acquisition, physical performance testing, and chemical composition analysis) from the work order and associates them with the testing parameters for each batch, establishing a "batch number - testing requirements - testing parameters" mapping table to provide data support for equipment resource allocation.

[0054] Following this, testing equipment resources are allocated. The control center has a built-in equipment management module that collects real-time data on the operating status (idle / busy / faulty), supported testing items, and current load rate of all automated testing equipment. Allocation follows two core principles: first, the equipment capability matching principle, which assigns equipment with corresponding functions based on testing needs. For example, surface image acquisition tasks are assigned to vision inspection equipment equipped with high-resolution industrial cameras, physical performance testing tasks are assigned to specialized equipment such as tensile testing machines and abrasion testing machines, and chemical composition analysis tasks are assigned to near-infrared spectroscopy detection equipment; second, the load balancing principle, which avoids single equipment handling too many tasks and causing testing delays. By calculating the ratio of the current number of tasks to be processed to the maximum processing capacity of the equipment, new tasks are assigned to idle equipment with lower load rates. After equipment allocation, the control center sends testing instructions to the corresponding equipment. The instructions include the target batch number, associated testing parameters, and testing sequence requirements, while simultaneously updating the equipment status to "busy" to prevent duplicate allocation.

[0055] After receiving the instruction, the testing equipment performs quality data acquisition according to the parameters in the predictive testing scheme. The acquisition process includes three core steps: First, image acquisition of the lace fabric surface, using an industrial camera with a resolution of no less than 20 megapixels, combined with a ring light source to eliminate shadow interference. The camera uniformly selects the acquisition area on the fabric surface according to the set sampling point density, taking 3-5 images per area to ensure image clarity. The image data is stored in a lossless format, and the acquisition time and camera parameters (focal length, exposure) are recorded simultaneously. Second, physical performance testing, calling the corresponding test program according to the standard testing method number. For example, tensile strength testing uses an electronic tensile testing machine, controlling the tensile speed and force measurement accuracy according to the set testing accuracy level, recording the force value change data in real time, and automatically calculating indicators such as tensile strength and elongation at break after the test. Third, chemical composition analysis, using a near-infrared spectrometer to perform non-destructive testing on the fabric. The spectral scanning range covers the characteristic wavelengths of the target components (such as fiber type, dye content). The scanning data is sampled at preset intervals to ensure that the data can accurately reflect the component proportion. The acquired raw data is processed through format standardization and converted into a quality data file with a unified structure.

[0056] In step S6, the specific process of generating the quality assessment record is as follows: Read the quality data file corresponding to each batch of lace fabric. Access the storage server through a file service interface. The interface uses a batch number-based retrieval mechanism; entering the batch number locates the corresponding quality data file, avoiding read delays caused by full file traversal. During the reading process, file validity is verified to check for file corruption (e.g., missing header, truncated data) and whether the file format conforms to standardization requirements. If the file is corrupted, a data recovery process is triggered, retrieving backup data from the testing device's local cache to regenerate the file. If the format is abnormal, a format error log is returned, notifying the data standardization stage to reprocess. After the file is read, a data cache is established by batch number to ensure efficient data retrieval in subsequent operations.

[0057] Subsequently, the quality indicator measurement values ​​are extracted from the quality data file. The extraction process is based on preset field mapping rules, associating the structured data in the file with standard quality indicator fields. For example, "tensile strength" and "elongation at break" fields are extracted from physical performance data; "pattern clarity score" and "number of surface defects" fields are extracted from surface image analysis results; and "fiber composition percentage" and "dye content" fields are extracted from chemical composition data. Data type validation is performed simultaneously during extraction to ensure that the data type of the measured values ​​(e.g., numeric, enumerated) is consistent with the requirements of the standard indicator fields. For example, "tensile strength" must be floating-point data; if character data appears, it is marked as an outlier and temporarily stored in an outlier list for manual review.

[0058] After extraction, the measured values ​​of quality indicators are compared with the allowable ranges in the lace fabric quality standard library. The quality standard library adopts a relational database architecture. The core data table contains fields such as quality attribute ID, quality attribute name, allowable upper limit value, allowable lower limit value, and judgment rules (e.g., "must simultaneously meet the upper limit ≥ measured value ≥ lower limit" and "if a single indicator fails, the whole batch fails"). The allowable ranges in the standard library are based on industry standards (such as GB / T series fabric standards) and enterprise internal control standards, and can be dynamically adjusted according to fabric category and application scenario. During comparison, the measured value is associated with the standard allowable range by "quality attribute ID", and each indicator is judged to see if the measured value is within the allowable range: if all indicators meet the requirements, the batch is initially judged to be qualified; if at least one indicator exceeds the allowable range, it is initially judged to be unqualified, and the name, measured value, and allowable range of the indicator that exceeds the range are recorded.

[0059] The quality assessment conclusion for each batch of lace fabric is generated based on the comparison results. The assessment conclusion includes a pass / fail judgment ("pass" or "fail") and remarks on abnormal indicators: a pass / fail conclusion only needs to clearly state the judgment result; a fail / fail conclusion requires detailed remarks on abnormal indicators, including the name of the abnormal indicator, the deviation of the measured value from the allowable range (e.g., "tensile strength measured value 2.8MPa, below the allowable lower limit of 3.0MPa"), and process-related hints that may lead to the abnormality (e.g., "suggested to be related to yarn count adjustment"). When generating the assessment conclusion, a conclusion identifier, a generation timestamp, and an assessor identifier (if manual review is involved) are added. The conclusion identifier is bound to the corresponding batch number to ensure uniqueness and traceability.

[0060] After the quality assessment conclusions for all individual batches are generated, they are summarized to form a total quality assessment record. The core content of the total record includes the assessment date, the number of lace fabric batches involved, the overall pass rate statistics, and a detailed assessment data table: the assessment date follows the ISO8601 standard format and is accurate to the hour; the batch quantity statistics are the total number of batches tested in this instance, distinguishing between "assessed and completed" and "pending review" statuses; the overall pass rate is calculated as "number of qualified batches / total number of assessed batches", with the result rounded to two decimal places; the detailed assessment data table adopts a tabular structure, including the batch number, pass / fail result, anomaly indicator remarks, and assessment time field for each batch, arranged in ascending order by batch number.

[0061] The foregoing has provided a detailed description of one embodiment of the present invention, but this description is merely a preferred embodiment and should not be construed as limiting the scope of the invention. All equivalent variations and modifications made within the scope of the claims of this invention should still fall within the patent coverage of this invention.

Claims

1. A data analysis-based method for assessing the quality of incoming lace fabric, characterized in that, Includes the following steps: S1. Receive a process parameter change list, which includes the changed process parameter items and their parameter values. S2. Based on the list of process parameter changes, query the pre-constructed lace fabric process knowledge graph to identify the incoming lace fabric quality attributes affected by the changed process parameters. S3. Based on the identified lace fabric quality attribute types, generate a predictive detection scheme, which includes a set of quality attributes to be detected and corresponding detection parameters. S4. Based on the predictive testing scheme, retrieve the inventory information of all batches of lace fabric in the current warehouse to determine the range of lace fabric batches that need to be tested. S5. Initiate predictive testing for the identified batch of lace fabric and collect quality data according to the testing parameters in the predictive testing plan. S6. Integrate the collected quality data with the batch information of the lace fabric to generate a quality assessment record.

2. The data analysis-based method for evaluating the quality of incoming lace fabric as described in claim 1, characterized in that, In step S1, the specific process of receiving the process parameter change list is as follows: After the downstream production management system completes the process parameter configuration update, it automatically generates a process parameter change notification file. The process parameter change notification file adopts a structured text format, which includes a change timestamp, the operator's identifier, and a list of changed process parameter item names. For each changed process parameter item, the parameter value before the change and the parameter value after the change are recorded. The system receives the process parameter change notification file through the file transfer service, parses the file content and verifies the data integrity, and stores the parsed changed process parameter items and their parameter values ​​as a process parameter change list. The process parameter change list maintains a data table structure in memory.

3. The method for evaluating the quality of incoming lace fabric based on data analysis according to claim 1, characterized in that, In step S2, the specific process of querying the pre-constructed lace fabric technology knowledge graph is as follows: The lace fabric technology knowledge graph uses a graph structure of nodes and edges for data storage. Node types include process parameter nodes and lace fabric quality attribute nodes. Edge types represent the influence relationship between process parameter nodes and lace fabric quality attribute nodes. Each edge contains an influence intensity coefficient and an influence direction identifier. Based on the name of the changed process parameter item in the process parameter change list, locate the corresponding process parameter node in the knowledge graph, traverse all outgoing edges from the process parameter node, obtain the set of directly connected lace fabric quality attribute nodes, calculate the influence quantification value of each lace fabric quality attribute node based on the influence intensity coefficient in the edge attribute and the change range of the process parameter value, mark the lace fabric quality attribute nodes whose influence quantification value exceeds the activation threshold as affected attributes, and form a list of lace fabric incoming material quality attributes.

4. The method for evaluating the quality of incoming lace fabric based on data analysis according to claim 1, characterized in that, In step S3, the specific process of generating the predictive detection scheme is as follows: Read each lace fabric quality attribute identifier from the incoming lace fabric quality attribute list, query the testing rule library, obtain the standard testing method number and benchmark testing parameters corresponding to each lace fabric quality attribute identifier, and adaptively adjust the benchmark testing parameters based on the influence quantification value calculated from the lace fabric process knowledge graph. The testing parameter adjustment includes increasing the sampling point density, improving the testing accuracy level, and extending the testing observation time. Bind the adjusted testing parameters to the lace fabric quality attribute identifier to form testing task items. All testing task items are combined into a predictive testing scheme, which includes a scheme number, generation timestamp, and a list of testing task items.

5. The method for evaluating the incoming quality of lace fabric based on data analysis according to claim 4, characterized in that, The specific process for adaptively adjusting the benchmark detection parameters is as follows: The adjustment level of the detection parameters is determined based on the magnitude of the influence quantification value. The influence quantification value is divided into multiple continuous intervals, and each interval corresponds to an adjustment level. The adjustment levels of the detection parameters include first-level adjustment, second-level adjustment, and third-level adjustment. Each adjustment level corresponds to a set proportion of the number of detection sampling points and a set level of detection accuracy.

6. The method for evaluating the quality of incoming lace fabric based on data analysis according to claim 1, characterized in that, In step S4, the specific process for determining the range of lace fabric batches requiring testing is as follows: The warehouse management system's lace fabric inventory database is queried to retrieve inventory records for all in-stock lace fabric batches. These records include the batch number, arrival time, storage location, and fabric composition specifications. Based on the lace fabric quality attributes involved in the predictive testing scheme, and combined with the lace fabric composition specifications, the correlation score between each lace fabric batch and the testing scheme is calculated. A correlation score threshold is set, and lace fabric batches with correlation scores greater than or equal to the threshold are filtered out to form the range of lace fabric batches that need to be tested. The lace fabric batch range is then sorted from highest to lowest correlation score.

7. The method for evaluating the quality of incoming lace fabric based on data analysis according to claim 6, characterized in that, The specific process for calculating the correlation score is as follows: Construct a correlation weight matrix between the composition specifications and quality attributes of lace fabric. The rows of the matrix correspond to different types of lace fabric composition specifications, the columns of the matrix correspond to different lace fabric quality attributes, and the weight values ​​in the matrix represent the sensitivity of the lace fabric composition specifications to the quality attributes. For each batch of lace fabric, its lace fabric composition and specification type are extracted, and the corresponding row vector is located in the weight matrix. According to the list of lace fabric quality attributes involved in the predictive detection scheme, the corresponding weight values ​​are extracted from the row vectors. All extracted weight values ​​are weighted and summed. The weight is the quantified value of the influence of each quality attribute in the predictive detection scheme. The weighted sum is normalized to obtain the correlation score of the lace fabric batch.

8. The method for evaluating the quality of incoming lace fabric based on data analysis according to claim 1, characterized in that, In step S5, the specific process of initiating the predictive detection operation is as follows: Based on the batch range of the lace fabric, a testing task assignment order is generated. The testing task assignment order includes a list of lace fabric batch numbers and a copy of the predictive testing plan corresponding to each batch. The testing task assignment order is distributed to the automated testing equipment control center. The automated testing equipment control center parses the testing task assignment order, allocates testing equipment resources, and the testing equipment performs quality data acquisition according to the testing parameters in the predictive testing plan. The acquisition process includes acquiring surface images of the lace fabric, testing physical properties, and analyzing chemical composition. The acquired raw data is processed through format standardization and converted into a quality data file with a unified structure.

9. The method for evaluating the quality of incoming lace fabric based on data analysis according to claim 1, characterized in that, In step S6, the specific process of generating the quality assessment record is as follows: Read the quality data file corresponding to each batch of lace fabric, extract the quality index measurement values, compare the quality index measurement values ​​with the allowable range in the lace fabric quality standard library, and generate a quality assessment conclusion for a single batch of lace fabric. The quality assessment conclusion includes the pass / fail result and abnormal index remarks. Summarize the quality assessment conclusions of all batches of lace fabric to form a total quality assessment record. The total quality assessment record includes the assessment date, the number of lace fabric batches involved, the overall pass rate statistics, and a detailed assessment data table.

Images