A Method and System for Quality Control in Pet Mat Production Based on Process Control

By constructing multi-source parameter features and evaluation dimensions, dynamically adjusting weights, and setting coupled penalty rules, the problems of lagging quality control and insufficient single-parameter monitoring in pet pad production are solved, achieving real-time and accurate quality assessment and early warning.

CN123117913APending Publication Date: 2026-11-17HANGZHOU DEBANG HOME TEXTILE CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611347534.0
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-09-02
Publication Date
2026-11-17

AI Technical Summary

Technical Problem

The existing quality control in the pet pad production process suffers from lag and the inability of single-parameter monitoring to identify coupled risks, resulting in the inability to correct process deviations in a timely manner and inaccurate evaluation results.

Method used

A process control-based approach is adopted to collect multi-source parameters, construct high-order, temporal, and interactive features, establish multiple evaluation dimensions, dynamically adjust weights, perform comprehensive membership judgment, set mandatory veto conditions and coupled penalty rules, and generate a comprehensive quality evaluation report.

Benefits of technology

It enables real-time, comprehensive, and accurate quality assessment during the pet pad production process, allowing for timely detection and correction of quality issues, and improving the sensitivity of anomaly detection and the ability to predict process parameter trends.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123117913A_ABST
    Figure CN123117913A_ABST
Patent Text Reader

Abstract

This application relates to the field of intelligent pet mat production technology, and discloses a method and system for quality control in pet mat production based on process control. The method includes: collecting multi-source parameters of the pet mat, including raw material parameters, process parameters, environmental parameters, and online detection parameters; preprocessing the multi-source parameters; constructing high-order features, temporal features, and interaction features based on the preprocessed multi-source parameters; establishing multiple evaluation dimensions and obtaining evaluation vectors for each dimension; dynamically adjusting the basic weights of each dimension, including environmental adjustment items and historical defect adjustment items, and simultaneously using high-order features to perform a secondary correction on the dynamically adjusted weights; obtaining a comprehensive membership vector based on the evaluation vector and the secondary corrected weights; and determining the initial quality level based on the comprehensive membership vector.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This application relates to the field of intelligent pet mat production technology, and in particular to a method and system for quality control in pet mat production based on process control. Background Technology

[0002] Pet mats are hygiene products used to collect pet excrement. They are typically made from materials such as a non-woven fabric surface, a composite core layer of superabsorbent polymer (SAP) and fluff pulp, and a PE waterproof bottom membrane, through processes including hot pressing, cutting, and edge sealing. The absorbency of the core layer directly determines the product's effectiveness, and the quality indicators of SAP are a key factor affecting the core layer's absorbency.

[0003] Currently, quality control in pet mat production mainly relies on end-of-line sampling and parameter threshold monitoring. End-of-line sampling involves quality inspectors taking samples from finished products after production and sending them to a laboratory for performance testing according to relevant standards. This method has a significant time lag; by the time quality problems are detected, a large number of products in the same batch have already been produced, making it impossible to correct process deviations in time and resulting in waste of raw materials and labor. Parameter threshold monitoring involves setting up several sensors on the production line to monitor process parameters such as speed, temperature, and pressure in real time, triggering an alarm when a parameter exceeds a preset threshold. While this method has a certain degree of real-time capability, each parameter is monitored independently, ignoring the coupling effect between parameters. This makes it impossible to provide a comprehensive and multi-dimensional evaluation of product quality and to give an intuitive quality grade determination. Therefore, how to achieve real-time, comprehensive, and accurate quality judgment during pet mat production is the fundamental problem that this invention aims to solve. Summary of the Invention

[0004] To achieve real-time, comprehensive, and accurate quality assessment during the pet mat production process, this application provides a process-control-based method and system for pet mat production quality management, employing the following technical solution: In a first aspect, the present invention provides a method for quality control in the production of pet mats based on process control, comprising: Collect multi-source parameters of the pet pad, including raw material parameters, process parameters, environmental parameters and online detection parameters, and perform parameter preprocessing on the multi-source parameters; High-order features, temporal features, and interaction features are constructed based on the multi-source parameters after data preprocessing. Multiple evaluation dimensions are established, and evaluation vectors for each dimension are obtained. The basic weights of each dimension are dynamically adjusted, including environmental adjustment items and historical defect adjustment items. At the same time, higher-order features are used to perform secondary correction on the dynamically adjusted weights. A comprehensive membership vector is obtained based on the evaluation vector and the secondary corrected weights. The initial quality level is determined based on the comprehensive membership vector. Check if any mandatory rejection condition is triggered. If any mandatory rejection condition is triggered, the system is directly judged as unqualified and an abnormal state vector is output. If no mandatory rejection condition is triggered, a coupling penalty correction is applied to the comprehensive membership vector of the interaction feature pair to obtain the final comprehensive membership vector and the final quality level. The stability index of the process is calculated based on the time series characteristics, and a comprehensive quality assessment report is generated and output by combining the final quality level and the abnormal state vector.

[0005] Optionally, the higher-order characteristics include core layer absorption potential index, molding quality index, proportioning deviation rate, comprehensive environmental factor, heat sealing sufficiency index, comprehensive raw material quality index, and deodorization performance index; The time series characteristics include short-term volatility, trend drift rate, extreme value deviation, and autocorrelation coefficient; The interactive features include absorption process coupling degree, environment-raw material sensitivity coefficient, heat sealing comprehensive effect coefficient, size-weight joint deviation, and speed-feeding following error.

[0006] Optionally, the raw material parameters include at least SAP residual monomers, SAP volatiles, SAP pH value, deodorization effect, SAP absorption rate, SAP absorption capacity, SAP water retention capacity, and SAP pressurized absorption capacity. The online detection parameters include at least the weight of a single product, the total length of the product, the total width of the product, the detection results of metal impurities, the detection results of surface defects, the quantitative value of the core layer, the thickness of the absorbent core layer, and the peel strength of the seal. The evaluation system includes the following dimensions: absorption performance, structural integrity, boundary quality, appearance, and environmental safety. The absorption performance was evaluated based on the SAP absorption rate, SAP absorption capacity, SAP water retention capacity, SAP pressurized absorption capacity, and core layer absorption potential index. The structural integrity dimension is evaluated based on the product's single-piece weight, core layer quantity, absorbent core layer thickness, proportioning deviation rate, and molding quality index. Evaluation was conducted on the boundary quality dimension based on the seal peel strength, heat seal adequacy index, and heat seal comprehensive effect coefficient. The evaluation is based on the results of metal impurity detection, surface defect detection, product length and width, and is conducted in terms of appearance. The evaluation was conducted in the environmental safety dimension based on SAP residual monomers, SAP volatiles, SAP pH value, deodorization effect, deodorization performance index, and environmental-raw material sensitivity coefficient.

[0007] Optionally, for any sub-indicator in each evaluation dimension, its membership value relative to each preset level is calculated, wherein the preset levels include excellent, good, qualified and unqualified. The evaluation vector for each dimension is obtained based on the membership value and corresponding weight of each sub-indicator.

[0008] Optionally, the process of dynamically adjusting the basic weights of each dimension includes: pass Dynamic adjustments are made; where d is the index of the dimension. The basic weights for dimension d. For environmental regulation items, Historical defect adjustment term; The environmental regulation item is calculated based on comprehensive environmental factors; The historical defect adjustment item is dynamically adjusted based on the defect frequency of similar products in various dimensions recently.

[0009] Optionally, the process of performing a second correction includes: When the core layer absorption potential index is 10% higher than the historical average, the weight of the absorption performance dimension increases by 2%, and the weight of the environmental safety dimension decreases by 1%. When the comprehensive environmental factor > 1.15, the weight of the environmental safety dimension increases by 3%; When the heat sealing adequacy index H5 < 0.4, the weight of the boundary quality dimension is increased by 5%; When the ratio deviation rate is greater than 3%, the weight of the structural integrity dimension increases by 4%, and the weight of the absorption performance dimension decreases by 2%.

[0010] Optionally, the process of determining the initial quality level based on the comprehensive membership vector includes: Determine the maximum membership degree: If the maximum membership degree is ≥0.55, it is directly determined as the level corresponding to the maximum membership degree; If 0.40 ≤ maximum membership degree < 0.55, it is judged as a critical state, and the level is taken as the weighted average of the levels corresponding to the first two membership degrees, and marked as needing review; If the maximum membership degree is less than 0.40, it is determined to be a conflict state, triggering a system alarm and requiring manual intervention for re-inspection.

[0011] Optionally, the mandatory veto condition includes: ①SAP residual monomer >800mg / kg; ② The test results for metal impurities were abnormal; ③ The surface defect detection result is abnormal; ④The SAP pH value is not within the preset range; ⑤ Seal peel strength <2.0N / 25mm.

[0012] Optionally, the penalty rules in the coupling penalty correction include absorption process coupling penalty, environment-raw material agglomeration penalty, heat sealing overheating penalty, size-weight coupling penalty, and speed-feeding following error penalty.

[0013] Secondly, the present invention provides a pet mat production quality control system based on process control, wherein the system employs the pet mat production quality control method based on process control as described in any of the preceding claims, comprising: The data acquisition and preprocessing module is used to collect multi-source parameters of the pet pad, including raw material parameters, process parameters, environmental parameters and online detection parameters, and to perform parameter preprocessing on the multi-source parameters; The feature engineering module is used to construct high-order features, temporal features, and interaction features; The evaluation module is used to establish multiple evaluation dimensions and obtain evaluation vectors for each dimension; dynamically adjust the basic weights of each dimension, including environmental adjustment items and historical defect adjustment items, and use higher-order features to perform secondary correction on the dynamically adjusted weights; obtain a comprehensive membership vector based on the evaluation vectors and the secondary corrected weights; and determine the initial quality level based on the comprehensive membership vector. The mandatory veto and penalty module is used to check whether a mandatory veto condition is triggered. If any mandatory veto condition is triggered, it is directly judged as unqualified and an abnormal state vector is output. If no mandatory veto condition is triggered, a coupling penalty correction is applied to the comprehensive membership vector of the interaction feature pair to obtain the final comprehensive membership vector and the final quality level. The stability assessment and output module is used to calculate the process stability index based on time-series characteristics, and generate and output a comprehensive quality assessment report by combining the final quality level and abnormal state vector.

[0014] In summary, this application includes at least one of the following beneficial technical effects: This invention performs real-time quality assessment based on online acquisition of multi-source process parameters, eliminating the need to wait for laboratory sampling results. Quality judgment can be completed simultaneously with product manufacturing, enabling immediate detection and correction of quality issues. By establishing multiple assessment dimensions, it covers all aspects of pet mat quality, resulting in more comprehensive and accurate evaluation results. Through constructing interactive features and setting coupling penalty rules, it effectively identifies coupling risks that single-parameter monitoring cannot detect (such as high-pressure particle breakage, high-humidity agglomeration, and heat-sealing overheating), significantly improving the sensitivity of anomaly detection. By extracting time-series features and calculating process stability indices, it can promptly detect drift trends and increased fluctuations in process parameters, providing early warnings of quality problems. Attached Figure Description

[0015] Figure 1 This is a flowchart of the steps in the pet mat production quality control method based on process control in this invention. Detailed Implementation

[0016] The embodiments of this application are described in detail below, and examples of the embodiments are shown in the accompanying drawings.

[0017] In the description of this specification, the references to "certain embodiments," "one embodiment," "some embodiments," "illustrative embodiment," "example," "specific example," or "some examples" refer to specific features, structures, materials, or characteristics described in connection with the described embodiment or example, which are 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.

[0018] Please see Figure 1 This application discloses a process control-based method for quality control in pet mat production, including: S1. Collect multi-source parameters of the pet pad, including raw material parameters, process parameters, environmental parameters and online detection parameters, and perform parameter preprocessing on the multi-source parameters; S2. Construct high-order features, temporal features, and interaction features based on the multi-source parameters after data preprocessing; S3. Establish multiple evaluation dimensions and obtain evaluation vectors for each dimension; dynamically adjust the basic weights of each dimension, including environmental adjustment items and historical defect adjustment items, and use higher-order features to perform secondary correction on the dynamically adjusted weights; obtain a comprehensive membership vector based on the evaluation vectors and the secondary corrected weights; determine the initial quality level based on the comprehensive membership vector. S4. Check if a mandatory rejection condition is triggered. If any mandatory rejection condition is triggered, the system will directly determine the system as unqualified and output an abnormal state vector. If no mandatory rejection condition is triggered, the system will apply a coupling penalty correction to the comprehensive membership vector of the interactive feature pair to obtain the final comprehensive membership vector and the final quality level. S5. Calculate the process stability index based on time-series characteristics, and generate and output a comprehensive quality assessment report by combining the final quality level and abnormal state vector.

[0019] This invention performs real-time quality assessment based on online acquisition of multi-source process parameters, eliminating the need to wait for laboratory sampling results. Quality judgment can be completed simultaneously with product manufacturing, enabling immediate detection and correction of quality issues. By establishing multiple assessment dimensions, it covers all aspects of pet mat quality, resulting in more comprehensive and accurate evaluation results. Through constructing interactive features and setting coupling penalty rules, it effectively identifies coupling risks that single-parameter monitoring cannot detect (such as high-pressure particle breakage, high-humidity agglomeration, and heat-sealing overheating), significantly improving the sensitivity of anomaly detection. By extracting time-series features and calculating process stability indices, it can promptly detect drift trends and increased fluctuations in process parameters, providing early warnings of quality problems.

[0020] In one embodiment, a process-controlled method for quality management in pet mat production includes: S1. Collect multi-source parameters for pet pads, including raw material parameters such as SAP residual monomer content, SAP volatile content, SAP pH value, SAP particle size distribution, SAP density, SAP absorption rate, SAP absorption capacity, SAP water retention, SAP pressurized absorption capacity, SAP deodorization effect, fluff pulp moisture content, and SAP to fluff pulp ratio; process parameters such as core layer forming speed, SAP feeding speed, fluff pulp feeding speed, core layer forming pressure, spray water volume, hot pressing temperature, hot pressing pressure, hot pressing time, composite roller gap, adhesive coating amount, nonwoven fabric unwinding tension, PE bottom film unwinding tension, cutting speed, cutting temperature, finished product drying temperature, finished product drying time, finished product compaction pressure, and sealing heat sealing temperature; environmental parameters such as workshop ambient temperature, workshop ambient humidity, raw material storage temperature, and raw material storage humidity; and online detection parameters such as single product weight, product length, product width, metal impurity detection results, surface defect detection results, core layer quantitative data, absorbent core layer thickness, and sealing peel strength.

[0021] Parameter preprocessing for multi-source parameters includes outlier handling, missing value imputation, and data standardization. Outlier handling employs physical threshold verification, directly classifying data exceeding physical limits as invalid. Missing value imputation uses linear imputation. Data standardization utilizes a median and interquartile range (IMR)-based standardization method, where the parameter value is subtracted from its median and then divided by the difference between its upper and lower quartiles. The standardized value is called the deviation; the closer to 0, the closer the parameter is to the normal level, and the larger the absolute value of the deviation, the more severe the deviation from the standard process.

[0022] S2. Construct higher-order features, including core layer absorption potential index H1, molding quality index H2, proportion deviation rate H3, comprehensive environmental factor H4, heat sealing sufficiency index H5, comprehensive raw material quality index H6, and deodorization performance index H7.

[0023] The core layer absorption potential index H1 is a weighted sum of SAP absorption, SAP water retention, and SAP pressurized absorption. The corresponding weight distribution is adaptively set according to the emphasis on absorption performance under rapid absorption, long-term retention, and pressure conditions. In this embodiment, the weights of SAP absorption, SAP water retention, and SAP are 0.4, 0.35, and 0.25, respectively. It should be noted that the unit of the core layer absorption potential index H1 is g / g, and it participates in the membership calculation as a sub-index.

[0024] The molding quality index H2 is determined based on the core layer molding pressure, core layer molding speed, and hot pressing temperature, and is expressed as: ,in, The coefficient of variation of the core layer forming speed. P06 is the coefficient of variation of the core layer forming pressure, and P06 is the hot pressing temperature (°C). The optimal temperature for hot pressing (°C) is set according to the process parameters; it should be noted here that the coefficient of variation refers to the ratio of the standard deviation to the mean within the window.

[0025] The ratio deviation rate H3 is the difference between the actual ratio of SAP and fluff pulp and the target ratio, and its ratio to the target ratio, reflecting the accuracy of the core material ratio.

[0026] The comprehensive environmental factor H4 is the product of the temperature factor and the humidity factor. The temperature factor is the ratio of the workshop ambient temperature to the reference temperature (25°C), and the humidity factor is the ratio of the workshop ambient humidity to the reference humidity (40%RH). The further the comprehensive environmental factor H4 deviates from 1.0, the greater the negative impact of environmental conditions on production.

[0027] The heat sealing adequacy index H5 is determined based on the hot-pressing pressure P07 (MPa), hot-pressing time P08 (s), and hot-pressing temperature P06 (°C), and is expressed as: Wherein, K (MPa·s / °C) is a process constant, which is material-dependent and obtained through experimental calibration.

[0028] The overall quality index H6 of raw materials is determined based on the SAP residual monomer content R01, SAP volatile matter content R02, and SAP pH value R03, and is expressed as: , , This indicates the content of SAP residual monomers and SAP volatiles after data standardization; , , The preset penalty coefficients are set to 0.25, 0.15, and 0.1 respectively in this embodiment. The higher the index, the better the overall quality of the raw materials.

[0029] The deodorization performance index H7 is determined based on the SAP deodorization effect R10 (%) and the workshop ambient humidity E02, and is expressed as: The deodorizing effect decreases as the ambient humidity increases. This index comprehensively reflects the deodorizing ability under actual use conditions.

[0030] It should be noted that the core layer absorption potential index H1, molding quality index H2, proportioning deviation rate H3, environmental comprehensive factor H4, heat sealing adequacy index H5, and deodorization performance index H7 in the higher-order characteristics are all calculated using the original measured values ​​of the parameters after outlier removal and missing value imputation; the residual monomer and volatile content in the raw material comprehensive quality index H6 are calculated using the deviation value after data standardization as a penalty term.

[0031] Construct time-series features, which are extracted using a sliding window method, with the window size being the most recent 30 consecutively produced products.

[0032] The short-term volatility T1 is the coefficient of variation of the first key parameter within the window, reflecting the degree of short-term volatility of the parameter. The first key parameter is SAP blanking, core layer molding pressure, and single-piece weight of the product.

[0033] The trend drift rate T2 is the trend slope of the second key parameter within the window, which is extracted by performing linear regression on the sequence of the second key parameter within the window. The second key parameter is the hot pressing temperature, the SAP feeding speed, and the gap between the composite rollers.

[0034] Extreme value deviation T3 reflects the extreme fluctuation range of the third key parameter within the window. It is the ratio of the difference between the maximum and minimum values ​​of the third key parameter within the window to the historical reference standard deviation of the corresponding third key parameter. The third key parameters are molding pressure, hot pressing temperature, and adhesive coating amount.

[0035] The autocorrelation coefficient T4 reflects the correlation between the preceding and following sequences of the fourth key parameter, and is expressed as: , This is the fourth key parameter. This is the average value of the fourth key parameter, which is the core layer forming speed and the SAP unloading speed.

[0036] We construct interactive features to capture coupling effects that cannot be detected by single-parameter analysis.

[0037] The coupling degree C1 of the absorption process is the product of the SAP absorption rate and the core layer molding pressure. The combination of high pressure and fast absorption will cause the SAP particles to break under pressure and reduce the water retention.

[0038] The environmental-raw material sensitivity coefficient C2 is the product of the workshop ambient humidity and the SAP volatile content. The combination of high humidity and high volatile content can cause the SAP surface to become viscous, increasing the risk of material discharge blockage.

[0039] The heat sealing comprehensive effect coefficient C3 is the product of the heat sealing temperature, heat sealing time, and heat sealing pressure, reflecting the total amount of heat sealing energy input. Too high (overheating) or too low (non-melting) are both abnormal.

[0040] The combined dimensional-weight deviation C4 is the product of the dimensional deviation coefficient and the weight deviation coefficient, where the weight deviation coefficient is... The dimensional deviation coefficient is D01 is the weight of a single product unit. The values ​​provided are reference values ​​for the weight of a single product; D02 is the overall length of the product; and D03 is the overall width of the product. This is a reference value for product area; this feature reflects the coordinated deviation between product size and weight. and Set according to product parameters.

[0041] The speed-to-material feeding error C5 is determined based on the SAP material feeding speed and the core layer forming speed, and is expressed as follows: , This represents the change in SAP feed rate (obtained by comparing it with the instantaneous value at the previous moment). This is a reference value for SAP material feeding speed. This represents the change in core layer forming speed (obtained by comparing it with the instantaneous value at the previous moment). This is a reference value for the core layer forming speed. This feature reflects whether the SAP unloading speed changes synchronously with the machine speed. It should be noted here that... and Set according to process data.

[0042] S3. Establish five evaluation dimensions: absorption performance, structural integrity, boundary quality, appearance, and environmental safety. Simultaneously, set four preset multi-level grades: "Excellent," "Good," "Qualified," and "Unqualified." The membership degree of each preset grade is determined based on the range of values ​​for sub-indicators under each evaluation dimension. Sub-indicators for the absorption performance dimension include SAP absorption rate, SAP absorption amount, SAP water retention, SAP pressurized absorption amount, and core layer absorption potential index. Sub-indicators for the structural integrity dimension include product unit weight, core layer quantity, absorbent core layer thickness, proportioning deviation rate, and molding quality index. Sub-indicators for the boundary quality dimension include seal peel strength, heat seal sufficiency index, and heat seal comprehensive effect coefficient. Sub-indicators for the appearance dimension include metal impurity detection results, surface defect detection results, product length, and product width. Sub-indicators for the environmental safety dimension include SAP residual monomers, SAP volatiles, SAP pH value, deodorization effect, deodorization performance index, and environmental-raw material sensitivity coefficient.

[0043] Taking SAP absorption volume as an example, its membership function for the superiority level is: The membership function for the "good" grade is: The membership function for the qualified level is: The membership function for the non-compliance grade is: For the remaining sub-indicators, the corresponding membership function parameters are set according to their physical characteristics and quality standards, which will not be detailed here.

[0044] We perform weighted aggregation of the membership degrees of sub-indicators within dimension d to obtain the evaluation vector of dimension d. ,in, Let be the weight of the i-th sub-indicator in dimension d, which is adaptively set based on the influence of the sub-indicators in the empirical data, satisfying the following conditions: , Let be the membership vector of the i-th sub-index in dimension d relative to the four levels. k is the index of the preset level.

[0045] Afterwards, through Obtain dynamically adjusted weights Where d is the index of the dimension. The basic weights for dimension d are set according to the core functions and product requirements. In this embodiment, the basic weights for the absorption performance dimension, structural integrity dimension, boundary quality dimension, appearance dimension, and environmental safety dimension are 30%, 20%, 18%, 12%, and 20%, respectively. For environmental regulation items, Historical defect adjustment term; environmental adjustment term is calculated based on comprehensive environmental factors; when comprehensive environmental factors > 1.15, the environmental adjustment term for the environmental safety dimension (d=5) is... The environmental adjustment items for the remaining dimensions are 0; the historical defect adjustment items are dynamically adjusted based on the defect frequency of similar products in each dimension recently, and the proportion of each dimension judged as "unqualified" in the past 1000 products is calculated. Obtain historical defect adjustment items , The average defect frequency across five dimensions.

[0046] The secondary correction process includes: when the core layer absorption potential index is 10% higher than the historical average index, the weight of the absorption performance dimension increases by 2%, and the weight of the environmental safety dimension decreases by 1%; when the comprehensive environmental factor is >1.15, the weight of the environmental safety dimension increases by 3%; when the heat sealing adequacy index H5 is <0.4, the weight of the boundary quality dimension increases by 5%; when the proportioning deviation rate is >3%, the weight of the structural integrity dimension increases by 4%, and the weight of the absorption performance dimension decreases by 2%. It should be noted that the historical average index is the average of all batches with an excellent quality rating within the past 7 days. If there are no excellent-rated batches within the past 7 days, data from an earlier time window is used as a substitute.

[0047] Finally, the comprehensive membership vector is obtained. The four components of V represent the comprehensive membership degree of the product belonging to "Excellent", "Good", "Qualified", and "Unqualified", respectively, and the sum of each component is 1. The process of determining the initial quality level based on the comprehensive membership degree vector includes: judging the maximum membership degree in each level: if the maximum membership degree is ≥ 0.55, it is directly determined as the level corresponding to the maximum membership degree; if 0.40 ≤ maximum membership degree < 0.55, it is determined as a critical state, and the level is the weighted average of the levels corresponding to the first two membership degrees, and marked as requiring review; if the maximum membership degree < 0.40, it is determined as a conflict state, triggering a system alarm and requiring manual intervention for re-inspection.

[0048] S4. Check if any mandatory rejection condition is triggered. If any mandatory rejection condition is triggered, it will be directly judged as unqualified and an abnormal status vector will be output. The mandatory rejection conditions include the following five items: ① SAP residual monomer > 800 mg / kg; indicating that the chemical safety standard is exceeded; ② The metal impurity detection result is abnormal, indicating the presence of metal foreign objects; ③ The surface defect detection result is abnormal, indicating serious appearance defects; ④ The SAP pH value is not within the preset range, indicating that the acidity or alkalinity exceeds the standard; ⑤ The seal peel strength < 2.0 N / 25 mm, indicating that the seal is seriously weak.

[0049] If the mandatory rejection condition is triggered, then the abnormal state vector , Each of the five veto conditions corresponds to a trigger flag: 1 for triggering and 0 for not triggering.

[0050] If the mandatory rejection condition is not triggered, the interaction feature is used to apply a coupling penalty correction to the comprehensive membership vector, resulting in the final comprehensive membership vector and the final quality level. The penalty rule in the coupling penalty correction includes absorption process coupling penalty: when C1 > 25 and SAP absorption rate < 25s, it is judged as a high-pressure particle risk, and the comprehensive membership vector is shifted 15% towards the non-conforming direction, i.e. ,in, The unit vector for the non-compliance level. .

[0051] The environmental-raw material agglomeration penalty is as follows: when C2 > 250 (%RH·%), it is judged as a high-humidity agglomeration risk, and the comprehensive membership vector is shifted 10% towards the unqualified direction.

[0052] The heat-sealing overheating penalty is as follows: when C3 > 350 (°C·s·MPa), it is judged as a risk of overheating and embrittlement, and the comprehensive membership vector is shifted 10% in the direction of non-compliance.

[0053] The size-weight coupling penalty is as follows: when C4 > 0.0025, it is judged as a systematic quantitative offset, and the comprehensive membership vector is shifted by 8% in the direction of non-compliance.

[0054] The speed-to-material feeding error penalty is as follows: when C5 > 0.1, it is judged as speed-to-material feeding mismatch, and the comprehensive membership vector shifts 5% in the unqualified direction.

[0055] It should be noted that when multiple penalty rules are triggered simultaneously, the penalties are executed sequentially and cumulatively. The offset calculation process refers to the expression in the absorption process coupling penalty. When the cumulative offset exceeds 30%, the final quality level is directly determined to be unqualified.

[0056] S5. Calculation of process stability index based on time series characteristics M is the number of parameters involved in the PSI calculation. , , , These represent the short-term volatility, trend drift rate, extreme value deviation, and autocorrelation coefficient of the j-th parameter, respectively. , , The preset normalized weights for each time series feature satisfy... Based on the normal production data of the production line over the past three months and the corresponding finished product sampling results, the PSI value range is (0,1], with the closer to 1 indicating a more stable production process. When PSI < 0.85, a yellow warning of "decreased production line stability" is issued; when PSI < 0.70, an orange warning of production line instability is issued, prompting a shutdown for inspection; when the PSI shows a downward trend for 10 consecutive products, all products in that time period are automatically marked as batches to be traced. The final comprehensive quality assessment report includes: product code and specification information, final quality grade and comprehensive membership vector, assessment vectors for each of the five assessment dimensions, process stability index PSI value and trend drift rate detection results, abnormal state vector, and quality traceability summary.

[0057] This invention also provides a process-control-based pet mat production quality control system, which employs a process-control-based pet mat production quality control method, including: The data acquisition and preprocessing module is used to collect multi-source parameters of the pet pad, including raw material parameters, process parameters, environmental parameters and online detection parameters, and to perform parameter preprocessing on the multi-source parameters; The feature engineering module is used to construct high-order features, temporal features, and interaction features; The evaluation module is used to establish multiple evaluation dimensions and obtain evaluation vectors for each dimension; dynamically adjust the basic weights of each dimension, including environmental adjustment items and historical defect adjustment items, and use higher-order features to perform secondary correction on the dynamically adjusted weights; obtain a comprehensive membership vector based on the evaluation vectors and the secondary corrected weights; and determine the initial quality level based on the comprehensive membership vector. The mandatory veto and penalty module is used to check whether a mandatory veto condition is triggered. If any mandatory veto condition is triggered, it is directly judged as unqualified and an abnormal state vector is output. If no mandatory veto condition is triggered, a coupling penalty correction is applied to the comprehensive membership vector of the interaction feature pair to obtain the final comprehensive membership vector and the final quality level. The stability assessment and output module is used to calculate the process stability index based on time-series characteristics, and generate and output a comprehensive quality assessment report by combining the final quality level and abnormal state vector.

[0058] 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 method for quality control in pet mat production based on process control, characterized in that, include: Collect multi-source parameters of the pet pad, including raw material parameters, process parameters, environmental parameters and online detection parameters, and perform parameter preprocessing on the multi-source parameters; High-order features, temporal features, and interaction features are constructed based on the multi-source parameters after data preprocessing. Multiple evaluation dimensions are established, and evaluation vectors for each dimension are obtained. The basic weights of each dimension are dynamically adjusted, including environmental adjustment items and historical defect adjustment items. At the same time, higher-order features are used to perform secondary correction on the dynamically adjusted weights. A comprehensive membership vector is obtained based on the evaluation vector and the secondary corrected weights. The initial quality level is determined based on the comprehensive membership vector. Check if any mandatory rejection condition is triggered. If any mandatory rejection condition is triggered, the system is directly judged as unqualified and an abnormal state vector is output. If no mandatory rejection condition is triggered, a coupling penalty correction is applied to the comprehensive membership vector of the interaction feature pair to obtain the final comprehensive membership vector and the final quality level. The stability index of the process is calculated based on the time series characteristics, and a comprehensive quality assessment report is generated and output by combining the final quality level and the abnormal state vector.

2. The pet mat production quality control method based on process control according to claim 1, characterized in that, The advanced features include the core layer absorption potential index, molding quality index, proportioning deviation rate, environmental comprehensive factor, heat sealing sufficiency index, raw material comprehensive quality index, and deodorization performance index. The time series characteristics include short-term volatility, trend drift rate, extreme value deviation, and autocorrelation coefficient; The interactive features include absorption process coupling degree, environment-raw material sensitivity coefficient, heat sealing comprehensive effect coefficient, size-weight joint deviation, and speed-feeding following error.

3. The pet mat production quality control method based on process control according to claim 2, characterized in that, The raw material parameters include at least SAP residual monomers, SAP volatiles, SAP pH value, deodorization effect, SAP absorption rate, SAP absorption capacity, SAP water retention capacity, and SAP pressurized absorption capacity. The online detection parameters include at least the weight of a single product, the total length of the product, the total width of the product, the detection results of metal impurities, the detection results of surface defects, the quantitative value of the core layer, the thickness of the absorbent core layer, and the peel strength of the seal. The evaluation system includes the following dimensions: absorption performance, structural integrity, boundary quality, appearance, and environmental safety. The absorption performance was evaluated based on the SAP absorption rate, SAP absorption capacity, SAP water retention capacity, SAP pressurized absorption capacity, and core layer absorption potential index. The structural integrity dimension is evaluated based on the product's single-piece weight, core layer quantity, absorbent core layer thickness, proportioning deviation rate, and molding quality index. Evaluation was conducted on the boundary quality dimension based on the seal peel strength, heat seal adequacy index, and heat seal comprehensive effect coefficient. The evaluation is based on the results of metal impurity detection, surface defect detection, product length and width, and is conducted in terms of appearance. The evaluation was conducted in the environmental safety dimension based on SAP residual monomers, SAP volatiles, SAP pH value, deodorization effect, deodorization performance index, and environmental-raw material sensitivity coefficient.

4. The pet mat production quality control method based on process control according to claim 3, characterized in that, For any sub-indicator in each evaluation dimension, calculate its membership value relative to each preset level, where the preset levels include excellent, good, qualified, and unqualified. The evaluation vector for each dimension is obtained based on the membership value and corresponding weight of each sub-indicator.

5. The pet mat production quality control method based on process control according to claim 4, characterized in that, The process of dynamically adjusting the basic weights of each dimension includes: pass Dynamic adjustments are made; where d is the index of the dimension. The basic weights for dimension d. For environmental regulation items, Historical defect adjustment term; The environmental regulation item is calculated based on comprehensive environmental factors; The historical defect adjustment item is dynamically adjusted based on the defect frequency of similar products in various dimensions recently.

6. The pet mat production quality control method based on process control according to claim 5, characterized in that, The process of making a second correction includes: When the core layer absorption potential index is 10% higher than the historical average, the weight of the absorption performance dimension increases by 2%, and the weight of the environmental safety dimension decreases by 1%. When the comprehensive environmental factor > 1.15, the weight of the environmental safety dimension increases by 3%; When the heat sealing adequacy index H5 < 0.4, the weight of the boundary quality dimension is increased by 5%; When the ratio deviation rate is greater than 3%, the weight of the structural integrity dimension increases by 4%, and the weight of the absorption performance dimension decreases by 2%.

7. The pet mat production quality control method based on process control according to claim 5, characterized in that, The process of determining the initial quality level based on the comprehensive membership vector includes: Determine the maximum membership degree: If the maximum membership degree is ≥0.55, it is directly determined as the level corresponding to the maximum membership degree; If 0.40 ≤ maximum membership degree < 0.55, it is judged as a critical state, and the level is taken as the weighted average of the levels corresponding to the first two membership degrees, and marked as needing review; If the maximum membership degree is less than 0.40, it is determined to be a conflict state, triggering a system alarm and requiring manual intervention for re-inspection.

8. The pet mat production quality control method based on process control according to claim 2, characterized in that, The mandatory veto conditions include: ①SAP residual monomer >800mg / kg; ② The test results for metal impurities were abnormal; ③ The surface defect detection result is abnormal; ④The SAP pH value is not within the preset range; ⑤ Seal peel strength <2.0N / 25mm.

9. The pet mat production quality control method based on process control according to claim 2, characterized in that, The penalty rules in the coupling penalty correction include absorption process coupling penalty, environment-raw material agglomeration penalty, heat sealing overheating penalty, size-weight coupling penalty, and speed-feeding following error penalty.

10. A pet mat production quality control system based on process control, characterized in that, The system employs the pet mat production quality control method based on process control as described in any one of claims 1-9, including: The data acquisition and preprocessing module is used to collect multi-source parameters of the pet pad, including raw material parameters, process parameters, environmental parameters and online detection parameters, and to perform parameter preprocessing on the multi-source parameters; The feature engineering module is used to construct high-order features, temporal features, and interaction features; The evaluation module is used to establish multiple evaluation dimensions and obtain evaluation vectors for each dimension; dynamically adjust the basic weights of each dimension, including environmental adjustment items and historical defect adjustment items, and use higher-order features to perform secondary correction on the dynamically adjusted weights; obtain a comprehensive membership vector based on the evaluation vectors and the secondary corrected weights; and determine the initial quality level based on the comprehensive membership vector. The mandatory veto and penalty module is used to check whether a mandatory veto condition is triggered. If any mandatory veto condition is triggered, it is directly judged as unqualified and an abnormal state vector is output. If no mandatory veto condition is triggered, a coupling penalty correction is applied to the comprehensive membership vector of the interaction feature pair to obtain the final comprehensive membership vector and the final quality level. The stability assessment and output module is used to calculate the process stability index based on time-series characteristics, and generate and output a comprehensive quality assessment report by combining the final quality level and abnormal state vector.