Textile pH detection data management method based on cloud collaboration

Abstract

The invention discloses a textile pH detection data management method based on cloud collaboration, and relates to the technical field of Internet of Things management, and the method comprises the steps: receiving a synchronous clock signal issued by a cloud NTP through an Internet of Things sensing node deployed in a textile production line, carrying out the collaborative collection and multi-source heterogeneous information fusion of pH original parameters, and generating a multi-modal original data packet; differentiated smooth filtering and feature extraction are carried out on the multi-modal original data packet, low-confidence data fragments are removed, and a structured trusted pH data packet is generated and written into a cloud time sequence database; and converting the optimization and improvement suggestions into real-time execution data oriented to multiple subjects, and writing the real-time execution data back to a cloud time sequence database. According to the method, through time sequence causal analysis and optimization improvement suggestion generation, accurate recognition of pH abnormal root cause nodes and process parameter adjustment, quality control and decision support in the production process are optimized, and the production efficiency and the product quality are improved.

Description

Technical Field

[0001] This invention relates to the field of Internet of Things (IoT) management technology, and in particular to a cloud-based collaborative method for managing textile pH detection data. Background Technology

[0002] With the rapid development of the Internet of Things (IoT), IoT sensor nodes are applied to various stages of the textile production process, especially in the field of quality monitoring. By collecting various parameters on the textile production line in real time, IoT sensor nodes provide rich data sources for monitoring the production process. They also provide computing power for the storage, processing, and analysis of data on cloud platforms, ensuring collaborative processing across devices and platforms in terms of time synchronization and data fusion.

[0003] However, despite the automation of data acquisition, storage and analysis, there are still some shortcomings in the efficient integration of multi-source heterogeneous data, the accurate synchronization of time-series data and the collaborative decision-making of multiple stakeholders. When processing multi-source heterogeneous data, there are problems such as data time alignment, inconsistent data quality and lack of real-time optimization and collaborative feedback for multiple stakeholders. These are the main challenges in the complex quality control and optimization decision-making of multi-party collaborative textile production processes. Summary of the Invention

[0004] In view of the aforementioned existing problems, the present invention is proposed.

[0005] Therefore, this invention provides a cloud-based collaborative method for managing textile pH detection data to address the issues of fusion of multi-source heterogeneous data and multi-party collaborative optimization decision-making.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution:

[0007] This invention provides a cloud-based collaborative method for managing pH detection data in textiles. The method includes: using IoT sensor nodes deployed on the textile production line to receive a synchronous clock signal from a cloud-based NTP (Network Transmission Platform) to collaboratively collect raw pH parameters and fuse multi-source heterogeneous information, generating multimodal raw data packets; performing differentiated smoothing filtering and feature extraction on the multimodal raw data packets to remove low-confidence data segments, generating structured and reliable pH data packets, and writing them into a cloud-based time-series database; performing collaborative clustering and association analysis based on the structured and reliable pH data packets in the cloud-based time-series database to obtain a dynamic production line status diagram shared by multiple stakeholders; conducting time-series causal analysis based on the dynamic production line status diagram to identify root cause nodes of pH anomalies, perform adjustment and deduction, and generate optimization and improvement suggestions; converting the optimization and improvement suggestions into real-time execution data for multiple stakeholders and writing it back to the cloud-based time-series database.

[0008] As a preferred embodiment of the cloud-based collaborative textile pH detection data management method of the present invention, the Internet of Things (IoT) sensing node refers to an edge sensing device at the quality control point of the textile production line that senses and collects raw pH parameters, process parameters, and equipment identification data in real time, and uploads them to the cloud to participate in NTP time synchronization and collaborative processing.

[0009] As a preferred embodiment of the cloud-based collaborative textile pH detection data management method of the present invention, the original pH parameter refers to the original pH detection data directly collected by IoT sensor nodes on the textile production line without filtering, correction and feature extraction processing.

[0010] As a preferred embodiment of the cloud-based collaborative textile pH detection data management method of the present invention, the specific steps for generating the multimodal raw data package are as follows:

[0011] By combining the synchronous clock signal sent by the cloud NTP, the raw pH parameters collected by the IoT sensor nodes of the textile production line are time-synchronized and calibrated to obtain the raw pH data sequence.

[0012] A multimodal time-series alignment algorithm was used to align and integrate the original pH data sequences to obtain a multi-source sensing data set;

[0013] By integrating data from different IoT sensor nodes, the multi-source sensing data set is structured and encapsulated to generate multimodal raw data packets.

[0014] As a preferred embodiment of the cloud-based collaborative textile pH detection data management method of the present invention, the cloud time-series database refers to the time-stamped data of pH detection data, environmental parameters, and process parameters stored and managed in the cloud in chronological order.

[0015] As a preferred embodiment of the cloud-based collaborative method for managing textile pH detection data according to the present invention, the specific steps for performing differential smoothing filtering and feature extraction on the multimodal raw data packets, removing low-confidence data segments, generating structured and reliable pH data packets, and writing them into the cloud time-series database are as follows.

[0016] A differential smoothing filter method is used to denoise and smooth the feature data of the original multimodal data packets to obtain a multimodal data sequence;

[0017] Statistical and temporal feature extraction is performed on multimodal data sequences, and trend analysis is conducted by combining historical data in a cloud-based time-series database to obtain a multimodal feature data set.

[0018] By using confidence filtering and structured recombination, low-confidence fragments are removed and fields are recombined in the multimodal feature data set to generate structured and reliable pH data packages, which are then written to a cloud-based time-series database.

[0019] As a preferred embodiment of the cloud-based collaborative textile pH detection data management method of the present invention, the method involves: performing collaborative clustering and association analysis on structured and reliable pH data packets in a cloud-based time-series database to obtain a dynamic production line status diagram shared by multiple stakeholders. The specific steps are as follows:

[0020] A collaborative clustering method is used to cluster and integrate structured and reliable pH data packets according to feature dimensions to obtain the cluster groups of pH data;

[0021] Based on the clustering and grouping of pH data, correlation analysis and data mining are performed with process parameters and equipment identification data to obtain a set of state node data.

[0022] Based on structured and reliable pH data packets in a cloud-based time-series database, a collaborative clustering method is used to integrate and analyze the state node data set, resulting in a dynamic production line state diagram shared by multiple stakeholders.

[0023] As a preferred embodiment of the cloud-based collaborative textile pH detection data management method described in this invention, the steps of conducting time-series causal analysis based on a dynamic production line status diagram, identifying the root cause nodes of pH anomalies, adjusting and extrapolating them, and generating optimization and improvement suggestions are as follows.

[0024] Based on the dynamic production line state diagram, time-series causal analysis was used to perform correlation analysis on the pH change sequence of state nodes to obtain an abnormal impact dataset.

[0025] The contribution of the abnormal impact dataset is quantitatively evaluated, and the influence of the pH change sequence of each state node on the pH anomaly is assessed to obtain the root cause node set of contribution scores.

[0026] By using time-series causal analysis, parameter adjustment and deduction are performed on the process parameters and equipment identifiers of the root cause node set to obtain optimization and improvement suggestions.

[0027] As a preferred embodiment of the cloud-based collaborative textile pH detection data management method of the present invention, the real-time execution data refers to the process execution data obtained in the cloud collaborative scheduling based on optimization and improvement suggestions, which records the actual execution adjustment instructions, quality inspection marks and task response status of the dynamic production line at the current moment.

[0028] As a preferred embodiment of the cloud-based collaborative textile pH detection data management method of the present invention, the specific steps for converting optimization and improvement suggestions into real-time execution data for multiple stakeholders and writing it back to the cloud time-series database are as follows.

[0029] The optimization and improvement suggestions are broken down and formatted by multiple stakeholders to obtain collaborative adjustment instruction data;

[0030] A time-series archiving method is used to associate, index, merge, and reassemble the collaborative adjustment instruction data with the structured trusted pH data packets in the cloud time-series database to obtain an updated pH detection management time-series dataset, which is then written back to the cloud time-series database.

[0031] The beneficial effects of this invention are as follows: by adopting a multimodal time-series alignment algorithm, the data from different IoT sensor nodes are accurately synchronized in time and multi-source data are fused, which optimizes the consistency and integrity of the data and provides a reliable basis for analysis; through time-series causal analysis and optimization improvement suggestions, the root cause nodes of pH abnormalities are accurately identified and process parameters are adjusted, which optimizes the quality control and decision support in the production process and improves production efficiency and product quality. Attached Figure Description

[0032] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0033] Figure 1 This is a flowchart of a cloud-based collaborative method for managing textile pH testing data.

[0034] Figure 2 A flowchart for generating multimodal raw data packets.

[0035] Figure 3 A flowchart for generating structured, trusted pH data packets.

[0036] Figure 4 A flowchart for optimizing and extrapolating the dynamic production line status diagram. Detailed Implementation

[0037] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0038] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0039] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0040] Reference Figures 1-4 This is one embodiment of the present invention, which provides a cloud-based collaborative method for managing textile pH detection data, including the following steps:

[0041] S1: By deploying IoT sensor nodes on the textile production line, the system receives the synchronous clock signal sent by the cloud NTP to collaboratively collect the original pH parameters and fuse multi-source heterogeneous information to generate multimodal raw data packets.

[0042] S1.1: IoT sensor nodes refer to edge sensing devices at quality control points on textile production lines that collect raw pH parameters, process parameters, and equipment identification data in real time and upload them to the cloud to participate in NTP time synchronization and collaborative processing.

[0043] S1.2: The raw pH parameter refers to the raw pH value detection data directly collected by IoT sensor nodes on the textile production line without filtering, correction and feature extraction processing.

[0044] S1.3: Combine the synchronous clock signal sent by the cloud NTP to perform time synchronization calibration on the raw pH parameters collected by the IoT sensor nodes of the textile production line to obtain the raw pH data sequence.

[0045] Furthermore, the IoT sensor node receives the synchronization clock signal sent by the cloud NTP and corrects the timestamp of the collected pH raw parameters according to the synchronization clock signal, so that the pH raw parameters collected by different IoT sensor nodes have a unified time reference. The pH raw parameters after timestamp correction are aligned and arranged in chronological order to obtain the pH raw data sequence.

[0046] It should be noted that cloud-based NTP refers to the cloud sending a synchronized clock signal through the Network Time Protocol (NTP) to ensure that the data collected by IoT sensor nodes on the textile production line has a consistent time stamp, thus guaranteeing the accuracy of the data's timing.

[0047] Synchronous clock signals refer to time signals transmitted via NTP from the cloud to ensure the time accuracy and consistency of IoT sensor nodes when collecting data, and to reduce data alignment problems caused by time differences.

[0048] S1.4: A multimodal time-series alignment algorithm is used to align and integrate the original pH data sequence to obtain a multi-source sensing data set.

[0049] Furthermore, based on the synchronous clock signal issued by the cloud-based NTP, the raw pH parameters collected by each IoT sensor node are timestamped to form a time-aligned raw pH data sequence. The raw pH data sequence is then matched with the process parameters and equipment identification data synchronously collected by the IoT sensor nodes within the same time window. A multimodal time-series alignment algorithm is used to interpolate and align the data from different sources under a unified time coordinate, obtaining a one-to-one temporal correspondence between the raw pH parameters, process parameters, and equipment identification data. The corresponding multidimensional data are then combined in chronological order to obtain a multi-source sensing data set.

[0050] Specifically, multimodal time-series alignment algorithms refer to methods that align and integrate time-series data from different sources to ensure consistency of data from different IoT sensor nodes in the time dimension.

[0051] The time dimension refers to recording and analyzing the changes of parameters (pH value, temperature and humidity, process parameters and equipment status) in the textile production process at different times, and tracking the production status at each point in time.

[0052] It should be noted that pH value is a quality control parameter in the textile production process. For example, in the dyeing process, a pH value that is too high or too low will lead to incomplete binding of dye and fiber, affecting the quality and uniformity of dyeing.

[0053] Interpolation and alignment refer to the process of estimating missing data points among collected data and the process of synchronizing heterogeneous data from different IoT sensor nodes according to time, and comparing and analyzing them under the same time reference.

[0054] Multidimensional data refers to raw pH parameters, process parameters, and equipment identification data.

[0055] Raw pH parameters refer to the raw pH parameters directly collected on the textile production line at the IoT sensor nodes, without filtering, correction, or feature extraction processing.

[0056] Process parameters refer to parameters used to control and monitor production quality during textile production, such as temperature, humidity, pressure, and flow rate.

[0057] Equipment identification data refers to the identification information of equipment on the production line, including equipment number, model, manufacturer, and installation location.

[0058] S1.5: By integrating data from different IoT sensor nodes, the multi-source sensing data set is structured and encapsulated to generate multimodal raw data packets.

[0059] Furthermore, the multi-source sensing data set obtained after alignment and integration by the multi-modal time-series alignment algorithm is aligned with the original pH parameters, process parameters and equipment identification data from each IoT sensing node according to the timestamp order. Fields are divided according to data type and source node, and the aligned multi-source sensing data set is structured and encapsulated using a unified data format to generate multi-modal raw data packets.

[0060] Specifically, multimodal raw data packets refer to the data set after time-series alignment and integration of multi-source sensing data.

[0061] A unified data format refers to the consistency and comparability of data from different sources during storage, transmission, and analysis.

[0062] Structured encapsulation processing refers to organizing and arranging multi-source heterogeneous data (such as raw pH parameters, process parameters, and equipment identification data from different IoT sensor nodes) according to a unified format and standard. Structured encapsulation processing includes data integration, standardization processing, field division, and data encapsulation.

[0063] Data integration involves combining data from different sensors (pH value, temperature, humidity, equipment status) according to time and sensor type.

[0064] Standardization processing involves uniformly converting the integrated multi-source heterogeneous data according to a unified format (field names and data types) and timestamp standards.

[0065] Field partitioning involves dividing the integrated multi-source heterogeneous data according to fields and encapsulating it in standardized formats (such as tables, JSON, and XML).

[0066] Data encapsulation: Encapsulating multi-source heterogeneous data that has been integrated and standardized into a single data packet.

[0067] It should be noted that the packaged contents include the original pH parameter sequence after time synchronization calibration, the process parameter sequence, and the equipment identification data.

[0068] S2: Perform differential smoothing filtering and feature extraction on the multimodal raw data packets, remove low-confidence data segments, generate structured and reliable pH data packets, and write them into the cloud time series database.

[0069] S2.1: Cloud-based time-series database refers to the time-stamped data of pH detection data, environmental parameters, and process parameters stored and managed in the cloud in chronological order.

[0070] S2.2: Differentiated smoothing filtering method is used to denoise and smooth the feature data of the original multimodal data packets to obtain multimodal data sequences.

[0071] Furthermore, a differentiated smoothing filtering method is used to denoise each dimension of the feature data in the data packet, remove noise signals from the feature data, and smooth the noise signals after removing the feature data to reduce irregular fluctuations in the noise signals and obtain a multimodal data sequence.

[0072] Specifically, the differential smoothing filtering method selectively smooths and removes noise while preserving effective signal features by taking advantage of the differences in the variation of feature data across different feature dimensions.

[0073] Differential smoothing filtering methods select the filtering window based on the data fluctuation characteristics of the feature dimension; for example, filtering is used to retain the abrupt change edge for the original pH parameter, moving average filtering is used for the temperature process parameter, and no smoothing is performed on the equipment identification data.

[0074] Feature data refers to information extracted from raw data, such as pH change trends, process conditions, and equipment operating status, which is extracted from raw pH parameters, process parameters, and equipment identification data through processing methods (smoothing filtering, statistical analysis).

[0075] The noise reduction process uses a differential smoothing filter method to process the original multimodal data packets and remove noise from the feature data.

[0076] Smoothing refers to reducing fluctuations and noise in feature data. Based on the differences in the variation of feature data in different dimensions, it prioritizes smoothing to remove noise signals with large fluctuations and irregularities, while retaining effective signal features.

[0077] S2.3: Perform statistical and temporal feature extraction on the multimodal data sequence, and combine it with historical data in the cloud time series database for trend analysis to obtain a multimodal feature data set.

[0078] Furthermore, by extracting statistical and temporal features from the multimodal data sequence, and utilizing historical pH detection data (previously collected raw pH parameters during textile production) and time-stamped data of environmental and process parameters stored in the cloud-based time-series database, trend analysis is performed by combining historical data and the cloud-based time-series database to identify the evolution direction of multimodal feature data, and mutation point analysis is performed to identify the state transition locations. The statistical features, temporal features, trend analysis results, and mutation point analysis results are then fused to form a multimodal feature data set containing multidimensional semantic information.

[0079] Specifically, trend analysis refers to analyzing changes in time series data (pH value, process parameters, equipment identification) to identify long-term trends, periodic fluctuations, and abrupt changes in the data, thereby obtaining the evolution patterns of various indicators in the production process.

[0080] Time series characteristics refer to the time dependencies and changes in time series data extracted through analysis, such as the volatility, periodicity, and abrupt change characteristics of time series data.

[0081] Mutation point analysis refers to the examination of time series data to identify points of change and abnormal fluctuations in the data. Changes represent sudden events and rapid changes in process conditions during the production process.

[0082] S2.4: Through confidence screening and structured recombination, low-confidence fragments are removed and fields are recombined in the multimodal feature data set to generate structured and reliable pH data packages and write them to the cloud time series database.

[0083] Furthermore, based on the degree of agreement between the data segments in the multimodal feature dataset and the mutation point analysis results in the cloud time series database for statistical and temporal feature extraction, a confidence score is calculated for each data segment. Data segments with confidence scores below the acceptable level are removed. The data segments are then reorganized according to the field structure of the original pH parameters, process parameters, and equipment identification data to form a structured and reliable pH data package with complete fields, time alignment, and semantic consistency. This structured and reliable pH data package is then written into the cloud time series database.

[0084] Confidence score, expressed as:

[0085] ;

[0086] in, Indicates the first Confidence score for each data segment Indicates the first Each characteristic dimension (pH value, temperature, humidity) Indicates the first The data segment in the _ ... The values ​​in each feature dimension Indicates the number of feature dimensions. The index represents the feature dimension, sequentially labeling each feature dimension involved in the calculation, such as raw pH parameters, process parameters, and equipment identification data. This represents a data segment from the sensor, with features including pH raw parameter features, process parameter features, and equipment identification data features.

[0087] It should be noted that confidence screening refers to removing low-confidence data segments from a multimodal feature dataset based on the reliability and accuracy of credible pH data.

[0088] The structured, reliable pH data package includes timestamps, raw pH parameters, process parameters, and equipment identification information.

[0089] The degree of agreement refers to the consistency of multimodal data across different dimensions. The range is obtained through statistical analysis (from 0 to 1), where 0 indicates a complete mismatch and 1 indicates a complete agreement.

[0090] Confidence score refers to the evaluation of each segment and feature based on the quality and reliability of pH data, measuring the accuracy and reliability of pH data. The value ranges from 0 to 1, where 0 indicates that the data is completely unreliable and 1 indicates that the data is completely reliable.

[0091] Acceptable level refers to the range set after evaluating multimodal data to select data segments that meet quality standards; it is obtained based on confidence scores, and the acceptable level is set between 0.6 and 1.

[0092] S3: Based on the structured and reliable pH data packets in the cloud time series database, collaborative clustering and association analysis are performed to obtain a dynamic production line status diagram shared by multiple stakeholders.

[0093] S3.1: A collaborative clustering method is used to cluster and integrate the structured and reliable pH data packets according to their feature dimensions to obtain the cluster groups of the pH data.

[0094] Furthermore, the feature vectors of pH values, process parameters, and equipment identification data contained in the structured trusted pH data package, after being aligned by time, are used with a collaborative clustering method to simultaneously optimize cluster division, integration, and optimization across multiple feature dimensions. This identifies the pH data set, and through the collaborative clustering method, the data is clustered according to the feature dimensions of pH change trends, process conditions, and equipment operating status. Similarities are observed within the same cluster group, while different cluster groups maintain distinguishability in the feature space, thus obtaining the cluster groups of the pH data.

[0095] It should be noted that collaborative clustering refers to identifying and integrating data groups with similar characteristics by simultaneously clustering data across multiple feature dimensions (pH data, process parameters, equipment identifiers).

[0096] In a better way, by co-optimizing the data clustering of different dimensions, data groups within the same cluster group show consistency in multiple features, while different cluster groups maintain obvious differences in the feature space. Co-clustering helps to better understand the behavior of different states in the textile production line and supports the construction and analysis of dynamic production line state diagrams.

[0097] Feature dimension refers to the dimension corresponding to the different attributes and features of multimodal data when analyzing and processing multimodal data. For example, in pH detection data, feature dimensions may include pH value, process parameters, equipment status, and timestamp. Each dimension represents the characteristics and attributes of the data, and the dimensions together form the complete feature space of the data.

[0098] pH change trend refers to the pattern of pH value change over time during textile production, which is manifested as the increase, decrease and fluctuation of pH value.

[0099] Process conditions refer to the process parameters involved in textile production, such as temperature, humidity, pressure, and flow rate.

[0100] Equipment operating status refers to the working status of equipment during the production process, including operating status, shutdown status, load, and efficiency parameters.

[0101] Distinctiveness refers to the ability to analyze different cluster groups, where data within the same cluster group are similar in terms of pH variation trends, process conditions, and equipment status characteristics, while different cluster groups exhibit differences in their feature spaces.

[0102] S3.2: Based on the clustering and grouping of pH data, and the correlation analysis and data mining processing of process parameters and equipment identification data, a set of state node data is obtained.

[0103] Furthermore, each cluster in the pH data clustering group is time-aligned with the process parameter and equipment identification data. Correlation analysis and data mining are used to mine and integrate the feature dimensions between the pH data, process parameter combination and equipment identification data, and integrate the pH data, process parameter combination and equipment identification to form a data set of status nodes.

[0104] Specifically, correlation analysis and data mining processing refer to in-depth analysis of multiple datasets (pH data, process parameters, equipment identifiers) to uncover the correlations and relationships between different data, and to analyze the relationship between pH value changes and process parameters and equipment status under different process conditions, so as to obtain the factors affecting production quality and performance.

[0105] S3.3: Based on the structured and reliable pH data packets in the cloud time series database, a collaborative clustering method is used to integrate and analyze the state node data set to obtain a dynamic production line state diagram that is shared by multiple stakeholders.

[0106] Furthermore, the pH data of each state node in the state node dataset, along with the corresponding process parameters and equipment identification data, are used as multi-dimensional features input to a collaborative clustering method. The state nodes are integrated through the collaborative clustering method, and combined with the time stamp data in the cloud time series database for correlation analysis, the temporal correlation between state nodes is obtained. A graph structure is constructed based on the pH evolution path of the state nodes, with the state nodes as vertices and the correlations between nodes based on process continuity and temporal causality as edges, forming a dynamic production line state graph.

[0107] Collaborative clustering and association analysis, expressed as:

[0108] ;

[0109] in, This represents the optimized results of grouping the dataset obtained after cluster analysis based on pH value trends, process conditions, and equipment status. Indicates the first Data from a data segment, Indicates the first The feature values ​​of each data segment are multidimensional data including pH value, process parameters, and equipment identification. This represents the total number of data points.

[0110] It should be noted that time-series correlation refers to the correlation between various data obtained by analyzing pH data, process parameters, and equipment identification data at different time points.

[0111] pH evolution path refers to the trajectory of pH value changes over time during textile production.

[0112] S4: Conduct time-series causal analysis based on the dynamic production line state diagram, identify the root cause nodes of pH abnormalities, perform adjustment and deduction, and generate optimization and improvement suggestions.

[0113] S4.1: Based on the dynamic production line state diagram, time-series causal analysis is used to perform correlation analysis on the pH change sequence of state nodes to obtain an abnormal impact dataset.

[0114] Furthermore, by extracting the pH change sequence of each state node continuously recorded in the time dimension from the dynamic production line state diagram, the pH change sequence is used as the input of time-series causal analysis. Time-series causal analysis is used to analyze the causal dependence of pH changes between any two state nodes, identify state node pairs with response relationships in the time dimension, and analyze the correlation strength of state nodes with time-series causal correlation with abnormal performance during the period of pH anomaly occurrence. The state nodes with causal dependence, pH change characteristics, process parameters and equipment identification data are integrated to obtain an anomaly impact dataset.

[0115] It should be noted that time-series causal analysis refers to the analysis of the causal relationships between various variables in time series data, and the time-series analysis of pH change sequences and data (such as process parameters and equipment status) between different state nodes to identify the causal relationships that occur sequentially in the time dimension.

[0116] Time-series causal analysis identifies the influence and dependency relationships between variables at different time points by analyzing the causal relationships among variables in time series data. In textile production, time-series causal analysis can be used to analyze the patterns of change in various parameters (pH value, process parameters, equipment status) over time, identifying the root causes of pH anomalies.

[0117] Specifically, time-series causal analysis correlates pH value changes with time-series data of process parameters and equipment status to determine the order in which they occur. Through causal deduction, it reveals that changes in parameters (temperature, pressure) lead to pH value anomalies, providing real-time optimization suggestions for adjusting production processes, optimizing equipment settings, and improving quality control.

[0118] Causality analysis refers to identifying the temporal interactions and dependencies between different state nodes through time-series causal analysis. Changes in one state node can trigger and influence changes in another. For example, changes in pH value can be affected by process parameters and equipment identification data, leading to changes in the node.

[0119] Association analysis refers to comparing the time series of pH value changes with process parameters and equipment identification data to identify state nodes (temperature and humidity changes) that are related to pH anomalies in the time dimension; through time-series causal analysis, the causal relationship between different state nodes is explored, and the influence intensity of each node on pH anomalies is evaluated.

[0120] S4.2: Quantitatively assess the contribution of the abnormal impact dataset and evaluate the degree of influence of the pH change sequence of each state node on the pH anomaly, and obtain the root cause node set of contribution scores.

[0121] Furthermore, the contribution of the pH change sequence of each state node in the abnormal impact dataset to the normal pH change characteristics of the corresponding nodes in the dynamic production line state diagram is quantitatively evaluated. Time-series causal analysis is used to assess the pH change characteristics of each state node on the abnormal propagation path state nodes in the time dimension. Combined with the influence of process parameter changes and equipment identification, the pH change sequence of each state node is correlated to obtain a contribution score for each state node. The state nodes with the highest contribution scores are selected after sorting them from high to low to obtain the root cause node set of contribution scores.

[0122] Specifically, quantitative contribution assessment refers to quantifying the role of each node in initiating and influencing pH changes by analyzing the degree of influence of each state node on pH anomalies.

[0123] The degree of influence is a ratio from 0 to 1, where 0 indicates that the node contributes nothing to the pH anomaly, and 1 indicates that the node contributes the most to the pH anomaly; the median value represents the partial contribution of the node to the anomaly.

[0124] S4.3: Through time-series causal analysis, parameter adjustment and deduction are performed on the process parameters and equipment identifiers of the root cause node set to obtain optimization and improvement suggestions.

[0125] Furthermore, taking the process parameters and equipment identifiers corresponding to each root node in the root node set as input, and based on the deviation direction between the pH change sequence and pH change characteristics of the root nodes in the dynamic production line status diagram, time-series causal analysis is used to adjust the direction and magnitude of process parameters that allow the pH change sequence to return to the normal range. At the same time, the feasibility of the adjustment scheme is verified by combining the equipment identifiers. The verified process parameter adjustment schemes are then bound to the corresponding equipment identifiers to generate optimization and improvement suggestions.

[0126] It should be noted that the optimization and improvement suggestions specifically refer to identifying the root cause nodes affecting pH abnormalities through time-series causal analysis of the dynamic production line status diagram, and proposing feasible process parameter adjustment schemes and optimization measures in combination with process parameters and equipment identification data.

[0127] For example, in the textile production process, time-series causal analysis revealed a causal relationship between abnormal pH fluctuations and excessively high heating temperatures in the dyeing machine. When the analysis found that the pH fluctuated drastically when the heating temperature exceeded the set value, the dye and fiber did not bind completely, affecting the dyeing quality. Based on the analysis, optimization and improvement suggestions included lowering the heating temperature to 85°C and shortening the heating time to 30 minutes; and adding a cooling stage at the end of the dyeing process to ensure that the pH value remained stable.

[0128] For example, in the textile dyeing process, the dyeing machine is set to a temperature of 80℃ to control the binding of dye and fiber. According to time-series causal analysis, when the heating temperature of the dyeing machine exceeds the set value of 80℃, the pH value is monitored to begin to fluctuate drastically (from 7.0 to between 6.5 and 8.0). When the temperature exceeds 90℃, the pH value fluctuates significantly, which is caused by changes in dye solubility and the intensity of chemical reactions. The pH fluctuation is caused by the unstable binding of dye and fiber due to excessively high temperature, which affects the dyeing effect and leads to uneven color and inconsistent dyeing depth.

[0129] The generation of optimization and improvement suggestions involves identifying process parameters and equipment identification data related to pH changes through time-series causal analysis, discovering a causal relationship between temperature and pH, and deducing the impact of temperature adjustments based on dynamic production line status diagrams, combined with equipment operation and process parameters. By lowering equipment temperature settings and adjusting process parameters, such as reducing heating time and stabilizing pH, the optimization suggestions are transformed into real-time execution data. This includes adjustment instructions (lowering temperature), quality inspection markers (checking pH changes), and task response status (equipment adjustment execution status), which are fed back to the cloud-based time-series database to form a complete closed-loop control.

[0130] The process of generating optimization and improvement suggestions involves using time-series causal analysis to deduce the root causes of pH anomalies. By combining the temporal changes in process parameters and equipment identification data, factors affecting pH fluctuations are identified. For example, during the dyeing process, time-series causal analysis revealed that excessively high temperatures were the cause of abnormal pH fluctuations. This analysis indicated that heating temperatures exceeding the set value led to incomplete binding of the dye to the fiber, affecting dyeing quality. Optimization and improvement suggestions included adjusting the heating temperature to 85°C, shortening the heating time to 30 minutes, and adding a cooling stage to ensure pH stability.

[0131] Process parameter adjustment schemes refer to optimization and adjustment measures for process parameters (such as temperature, humidity, and flow rate) in the production process, based on the root cause nodes and contribution scores identified through analysis of dynamic production line status diagrams and time-series causal analysis.

[0132] Ideally, the production process should be optimized to reduce abnormal pH fluctuations, ensure product quality, and improve production efficiency and stability by adjusting process parameters.

[0133] Feasibility verification refers to checking and verifying the collected process parameters and equipment identification data to ensure their reliability and effectiveness in practical applications. This includes confirming the quality, completeness, accuracy, and consistency of the data, and ensuring that the process parameters and equipment identification data are synchronized in time, standardized in format, and meet the requirements of the process parameter adjustment plan.

[0134] S5: Convert optimization and improvement suggestions into real-time execution data for multiple stakeholders and write it back to the cloud time-series database.

[0135] S5.1: Real-time execution data refers to the process execution data obtained in the cloud-based collaborative scheduling based on optimization and improvement suggestions, which records the actual execution adjustment instructions, quality inspection marks, and task response status of the dynamic production line at the current moment.

[0136] S5.2: Perform instruction decomposition and format conversion on optimization and improvement suggestions from multiple subject dimensions to obtain collaborative adjustment instruction data.

[0137] Furthermore, based on the process parameter adjustment content, target equipment identification, and execution conditions of the optimization and improvement suggestions, and according to the division of responsibilities among the three parties—quality inspectors, process engineers, and equipment maintenance personnel—the optimization and improvement suggestions are decomposed into sub-instructions for quality inspection marking tasks, process parameter modification tasks, and equipment control tasks, respectively. According to the interface specifications of the collaborative platform used by the three parties, each sub-instruction is formatted and converted, and the data receiving protocol of the corresponding execution terminal of the converted sub-instruction is used to generate collaborative adjustment instruction data.

[0138] Specifically, the multiple stakeholders refer to quality inspectors, process engineers, and equipment maintenance personnel.

[0139] The data receiving protocol refers to the collaborative platform interface specifications used by different execution entities (quality inspectors, process engineers, and equipment maintenance personnel) to correctly receive and process instruction data from the cloud.

[0140] S5.3: Using a time-series archiving method, the collaborative adjustment instruction data and the structured trusted pH data packets in the cloud time-series database are associated, indexed, merged, and recombined to obtain an updated pH detection management time-series dataset, which is then written back to the cloud time-series database.

[0141] Furthermore, based on the time stamps contained in the collaborative adjustment instruction data and the time stamp data in the cloud time series database, time alignment is performed. A one-to-one association is established between the actual adjustment instructions, quality inspection marks, and task response status in the collaborative adjustment instruction data and the structured trusted pH data packets at the corresponding time points, forming time series record data with execution feedback tags. All time series record data with execution feedback tags are merged and restructured with the original structured trusted pH data packets in chronological order at the field level to generate an updated pH detection management time series dataset containing pH detection results, process parameters, equipment identification, and execution feedback information. The updated pH detection management time series dataset is then written into the cloud time series database.

[0142] It should be noted that the one-to-one association refers to matching the adjustment instructions, quality inspection marks, and task response status in the real-time execution data with the structured trusted pH data packets at the corresponding time points in the cloud time-series database.

[0143] Time-series data refers to real-time execution data and state changes of specific variables and parameters at different points in time, arranged in chronological order.

[0144] In summary, this invention optimizes data consistency and integrity by employing a multimodal time-series alignment algorithm to achieve precise time synchronization and multi-source data fusion from different IoT sensor nodes, providing a reliable foundation for analysis. Furthermore, through time-series causal analysis and the generation of optimization and improvement suggestions, it accurately identifies root causes of pH anomalies and adjusts process parameters, thereby optimizing quality control and decision support in the production process and improving production efficiency and product quality.

[0145] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A cloud-based collaborative method for managing textile pH detection data, characterized in that: include, By deploying IoT sensor nodes on the textile production line, the system receives synchronous clock signals from the cloud NTP to collaboratively collect pH raw parameters and fuse multi-source heterogeneous information, generating multimodal raw data packets. Differential smoothing filtering and feature extraction are performed on the original multimodal data packets to remove low-confidence data fragments, generate structured and reliable pH data packets, and write them into the cloud time series database; Collaborative clustering and association analysis are performed based on structured and reliable pH data packets in a cloud-based time-series database to obtain a dynamic production line status diagram shared by multiple stakeholders. Based on the dynamic production line state diagram, a time-series causal analysis was conducted to identify the root cause nodes of pH abnormalities, and adjustment and deduction were performed to generate optimization and improvement suggestions. The optimization and improvement suggestions are converted into real-time execution data for multiple stakeholders and written back to the cloud time-series database.

2. The cloud-based collaborative method for managing textile pH detection data as described in claim 1, characterized in that: The IoT sensing node refers to an edge sensing device located at the quality control point of the textile production line, which senses and collects raw pH parameters, process parameters, and equipment identification data in real time, and uploads them to the cloud to participate in NTP time synchronization and collaborative processing.

3. The cloud-based collaborative method for managing textile pH detection data as described in claim 2, characterized in that: The raw pH parameter refers to the raw pH value detection data directly collected by IoT sensor nodes on the textile production line without filtering, correction, or feature extraction processing.

4. The cloud-based collaborative method for managing textile pH detection data as described in claim 3, characterized in that: The specific steps for generating the multimodal raw data packet are as follows. By combining the synchronous clock signal sent by the cloud NTP, the raw pH parameters collected by the IoT sensor nodes of the textile production line are time-synchronized and calibrated to obtain the raw pH data sequence. A multimodal time-series alignment algorithm was used to align and integrate the original pH data sequences to obtain a multi-source sensing data set; By integrating data from different IoT sensor nodes, the multi-source sensing data set is structured and encapsulated to generate multimodal raw data packets.

5. The cloud-based collaborative method for managing textile pH detection data as described in claim 4, characterized in that: The cloud-based time-series database refers to the time-stamped data of pH detection data, environmental parameters, and process parameters stored and managed in the cloud in chronological order.

6. The cloud-based collaborative method for managing textile pH detection data as described in claim 5, characterized in that: The specific steps for performing differential smoothing filtering and feature extraction on the multimodal raw data packets, removing low-confidence data segments, generating structured and reliable pH data packets, and writing them into the cloud time-series database are as follows. A differential smoothing filter method is used to denoise and smooth the feature data of the original multimodal data packets to obtain a multimodal data sequence; Statistical and temporal feature extraction is performed on multimodal data sequences, and trend analysis is conducted by combining historical data in a cloud-based time-series database to obtain a multimodal feature data set. By using confidence filtering and structured recombination, low-confidence fragments are removed and fields are recombined in the multimodal feature data set to generate structured and reliable pH data packages, which are then written to a cloud-based time-series database.

7. The cloud-based collaborative method for managing textile pH detection data as described in claim 6, characterized in that: The method involves collaborative clustering and association analysis based on structured, reliable pH data packets from a cloud-based time-series database to obtain a dynamic production line status diagram shared by multiple stakeholders. The specific steps are as follows: A collaborative clustering method is used to cluster and integrate structured and reliable pH data packets according to feature dimensions to obtain the cluster groups of pH data; Based on the clustering and grouping of pH data, correlation analysis and data mining are performed with process parameters and equipment identification data to obtain a set of state node data. Based on structured and reliable pH data packets in a cloud-based time-series database, a collaborative clustering method is used to integrate and analyze the state node data set, resulting in a dynamic production line state diagram shared by multiple stakeholders.

8. The cloud-based collaborative method for managing textile pH detection data as described in claim 7, characterized in that: The process involves conducting time-series causal analysis based on a dynamic production line state diagram to identify the root cause of pH anomalies, perform adjustment and deduction, and generate optimization and improvement suggestions. The specific steps are as follows: Based on the dynamic production line state diagram, time-series causal analysis was used to perform correlation analysis on the pH change sequence of state nodes to obtain an abnormal impact dataset. The contribution of the abnormal impact dataset is quantitatively evaluated, and the influence of the pH change sequence of each state node on the pH anomaly is assessed to obtain the root cause node set of contribution scores. By using time-series causal analysis, parameter adjustment and deduction are performed on the process parameters and equipment identifiers of the root cause node set to obtain optimization and improvement suggestions.

9. The cloud-based collaborative method for managing textile pH detection data as described in claim 8, characterized in that: The real-time execution data refers to the process execution data obtained in the cloud-based collaborative scheduling based on optimization and improvement suggestions, which records the actual execution adjustment instructions, quality inspection marks, and task response status of the dynamic production line at the current moment.

10. The cloud-based collaborative method for managing textile pH detection data as described in claim 9, characterized in that: The specific steps for converting optimization and improvement suggestions into real-time execution data for multiple stakeholders and writing it back to the cloud-based time-series database are as follows: The optimization and improvement suggestions are broken down and formatted by multiple stakeholders to obtain collaborative adjustment instruction data; A time-series archiving method is used to associate, index, merge, and reassemble the collaborative adjustment instruction data with the structured trusted pH data packets in the cloud time-series database to obtain an updated pH detection management time-series dataset, which is then written back to the cloud time-series database.

Images