Data time alignment method and system for information acquisition points of tobacco tube equipment

Abstract

The invention discloses a tobacco tube equipment information acquisition point data time alignment method and system, which are suitable for synchronous monitoring and data analysis of different equipment in an industrial production environment. By introducing an automatic time alignment technology and utilizing data translation and Pearson correlation analysis, the optimal time offset between different acquisition points is automatically determined, and accurate alignment of data acquired by different devices is realized. According to the method, the real-time performance and reliability of data analysis are improved, solid data support is provided for intelligent control and production optimization of tobacco equipment, and the high requirements of barrel type equipment for data real-time alignment and analysis are met.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing and analysis, and in particular to a method and system for aligning data time at information collection points of tobacco tube-type devices, and in particular to a method and system for aligning data time at information collection points of multiple devices based on time series data translation and correlation calculation. Background Art

[0002] With the rapid development of industrial automation and intelligent manufacturing, the demand for multi-device data collection and synchronous processing is growing in industrial production processes. Data collected by different devices is typically accompanied by unique timestamps. However, due to differences in device workflows and asynchrony in data collection systems, the data between devices exhibits temporal deviations, making it difficult to conduct effective analysis and comparison.

[0003] Traditional data alignment methods mostly rely on fixed time intervals preset by manual experience or through scrap methods. They have problems of high cost, low efficiency and poor accuracy, and cannot meet the high requirements for real-time data alignment and analysis in modern industrial production environments.

[0004] In the tobacco shred production process, most drum-type equipment (such as loosening and conditioning machines and thin-plate drying machines) is a large time-delay system, and the material usually needs to stay in the drum for 4 to 6 minutes. Therefore, the need for time alignment is more urgent for drum-type equipment, and the requirements for alignment accuracy are also higher.

[0005] Automatic time series alignment technology, through data shifting and correlation analysis, effectively resolves the issue of time asynchrony between different devices. By dynamically adjusting the data's time offset and combining it with correlation calculations, it can precisely identify the optimal alignment time for data from different collection points, thereby improving the accuracy and efficiency of data analysis. This approach not only ensures that device data is analyzed on the same timeline but also provides strong support for intelligent optimization of production processes.

[0006] With the growing demand for smart manufacturing, the application prospects of automated data alignment technology in the industrial field are very broad, especially in the fields of production monitoring and equipment optimization that require real-time data synchronization. Summary of the Invention

[0007] The purpose of the present invention is to provide a method and system for time alignment of data collected from information collection points on tobacco-container equipment based on time series data shift and correlation calculation. This method aims to address the problems of poor analysis accuracy and inefficiency caused by the asynchrony of data collection times across different devices in existing industrial production. By introducing automated time alignment technology, utilizing data shift and Pearson correlation analysis, the optimal time offset between different collection points is automatically determined, enabling precise alignment of data collected from different devices. This method improves the real-time performance and reliability of data analysis, providing solid data support for intelligent control and production optimization of tobacco-container equipment, and meeting the high demands of modern industry for real-time data alignment and analysis.

[0008] In order to adapt to the situation where there are multiple equipment collection points in the silk-making workshop, the overall concept of the technical solution adopted by the present invention is as follows:

[0009] 1. Point-by-point alignment, including:

[0010] For data from multiple collection points, perform the above data shift and Pearson correlation coefficient calculation for each pair of collection points to determine the optimal shift time. The specific steps are as follows:

[0011] (1) Selecting a benchmark collection point: First, select a collection point as the benchmark point for time alignment. Usually, a collection point with high data quality or a more critical location can be selected.

[0012] (2) Pairwise alignment: Align the data of other acquisition points with the data of the reference acquisition point in sequence. For each pair of acquisition points, use data translation and Pearson correlation coefficient calculation to determine the optimal translation time.

[0013] Furthermore, the data translation includes:

[0014] (a) Select data series: Select two time series data sets, A and B, to be aligned and set the initial translation step size (e.g., 1 second, 2 seconds, or 1 minute). The translation step size can be set based on the device's acquisition frequency and data characteristics.

[0015] (b) Data shift operation: Keeping data sequence A unchanged, data sequence B is shifted incrementally according to the set shift step size. After each shift, the time axis of data sequence B shifts relative to data sequence A, generating a new time point correspondence.

[0016] (c) Generate alignment relationships: With each translation, a new set of time point alignment relationships is generated between data sequence B and data sequence A. This allows the correlation between the two sets of data to be evaluated at different translation times.

[0017] (d) Dynamically adjust the shift range: The number and range of shifts are dynamically adjusted based on actual data and equipment operating characteristics. A maximum shift time range is typically set to avoid meaningless calculations caused by excessive shifts. The total shift time range can be determined based on factors such as the material transfer time between equipment during the production process. For equipment with shorter production time intervals, the maximum shift range may be as short as two minutes; for equipment with longer transfer times, the shift range can be appropriately extended.

[0018] Furthermore, the calculation of the Pearson correlation coefficient includes:

[0019] (a) Define the data sequence: Let the data sequence after translation be A = {A1, A2, …, An}, and the other data sequence be B = {B1, B2, …, Bn}

[0020] (b) Calculate the average: Calculate the average of two data series:

[0021]

[0022] (c) Calculate the covariance:

[0023] Covariance:

[0024]

[0025] variance:

[0026]

[0027] (d) Calculate the Pearson correlation coefficient r: After each translation operation, use the Pearson correlation coefficient to calculate the correlation between data series A and the translated data series B. The Pearson correlation coefficient is a statistic used to measure the degree of linear correlation between two series. Its calculation formula is:

[0028]

[0029] Where Cov(A,B) is the covariance of data sequences A and B.

[0030] Var(A) and Var(B) are the variances of data series A and B respectively.

[0031] (e) Correlation coefficient determination:

[0032] When r is close to 1, it means that the two data series have a strong positive correlation;

[0033] When r is close to -1, it means that the two data series have a strong negative correlation;

[0034] When r is close to 0, it means there is no obvious linear correlation between the two data series.

[0035] (f) Determine the optimal translation time:

[0036] After each data shift, calculate the corresponding Pearson correlation coefficient, r. Record the correlation coefficient at different shift times and plot the relationship between the correlation coefficient and shift time. By analyzing the curve, find the shift time at which the correlation coefficient reaches its maximum value. This is the optimal time alignment point for the data at the two acquisition points.

[0037] (g) Data alignment verification:

[0038] Apply the optimal shift time to the data series, adjust the time of the data, and recalculate the Pearson correlation coefficient to verify the alignment. If the correlation coefficient meets the preset threshold requirement (for example, r>0.8), the data alignment is considered successful.

[0039] (3) Data adjustment: Based on the determined optimal translation time, the corresponding acquisition point data is time-adjusted so that it is precisely aligned with the data of the reference acquisition point on the time axis.

[0040] 2. Global time alignment matrix, including:

[0041] During the multi-point alignment process, a global time alignment matrix is ​​established to record the optimal translation time between each acquisition point, thereby achieving global time alignment of all acquisition point data. The specific steps are as follows:

[0042] (1) Matrix construction: Assume that the total number of acquisition points is n, and establish a ×n×n time alignment matrix T, where the matrix elements T ij It represents the optimal translation time of acquisition point i relative to acquisition point j.

[0043] (2) Filling the matrix: Using the pairwise alignment method, calculate the optimal translation time between all acquisition points and fill the results into the corresponding positions in the matrix.

[0044] (3) Determine the global time reference: Taking the reference acquisition point as the reference, calculate the translation time of each acquisition point relative to the global time reference according to the translation time in the matrix T.

[0045] (4) Global data alignment: Time-adjust the data of all acquisition points to align them with the global time reference, so as to achieve precise synchronization of all data on the same time axis.

[0046] 3. Automation implementation, including:

[0047] To improve the efficiency and accuracy of data alignment, the automated implementation of the present invention specifically includes the following:

[0048] (1) Data alignment algorithm design: Integrate data translation, Pearson correlation coefficient calculation, and multi-point alignment steps into an automated algorithm to reduce manual intervention. The algorithm can dynamically adjust the translation step size, translation range, and correlation coefficient threshold based on the device operating status and data characteristics to improve alignment accuracy.

[0049] (2) Software platform development: Real-time acquisition of data from each device collection point ensures data integrity and timeliness. This includes functions such as data cleaning, outlier processing, data translation, and correlation calculation. This platform enables automated time alignment of multi-point data and generates a global time alignment matrix. The aligned data is stored and supports data query, analysis, and visualization.

[0050] (3) Real-time processing and monitoring: The system can process newly collected data in real time, perform time alignment and correlation analysis, and meet the real-time monitoring needs of the production process. The integrated anomaly detection function can identify abnormal values ​​and abnormal trends in the data, provide early warning information, and assist in production decision-making.

[0051] (4) User interaction and visualization: Provides an intuitive operation interface, allowing users to easily view data alignment status, correlation analysis results, etc. Supports graphical display of aligned data, such as trend charts, correlation charts, etc., to facilitate user understanding and analysis.

[0052] Specifically, the method for aligning the data time of tobacco tube device information collection points of the present invention includes the following steps:

[0053] S1 acquires and uniformly manages the raw data from each device collection point, ensuring data integrity and traceability, and providing a reliable data foundation for subsequent time alignment.

[0054] S2 calculates the initial time offset between devices based on the operating characteristics of each device and the material flow time. This includes: combining the process flow of the devices and the material transmission time between devices to provide a reasonable time offset estimate, providing a basis for data translation processing.

[0055] Furthermore, the method for determining the target single-machine maximum delay involved in the present invention includes:

[0056] In order to accurately determine the target maximum delay of a single tobacco tube device, the present invention first needs to obtain the historical operating data of the relevant equipment (such as inlet moisture, water addition, outlet moisture, etc.). In the silk-making workshop, multiple tobacco tube devices (such as loosening and rehumidification machines, leaf moistening feeders, and thin plate drying machines) are equipped with different collection points, and each collection point records the above parameters in real time through sensors. The data collection frequency of each collection point can be set to a time interval of 1 second, 2 seconds, or longer according to production requirements, and the data format is unified into a time series format;

[0057] Since there may be noise, missing data, or invalid values ​​during the acquisition process, the raw data needs to be cleaned and filled to ensure data integrity and accuracy. Invalid data can be processed by removing outliers, filling missing data with interpolation, or other standardized data processing methods to ensure that subsequent time alignment processing is based on high-quality data.

[0058] To solve this problem, the present invention adopts a data shift method to gradually shift the data sequence of the farthest acquisition point at the device entrance relative to the data sequence of the farthest acquisition point at the exit, and obtain the most relevant time point, which is the target single-machine maximum delay.

[0059] S3 gradually shifts the collected data from each device based on the target time offset, adjusting the position of the data on the time axis to ensure that the data from different collection points have preliminary consistency on the same time axis.

[0060] S4 Pearson correlation analysis evaluates the effectiveness of data alignment by calculating the Pearson correlation coefficient between the shifted data series. The data shift and correlation calculation process is repeated, gradually adjusting the time offset until the optimal time alignment point is found that maximizes the correlation coefficient.

[0061] S5 performs data adjustment and alignment. Based on the optimal time alignment point, it makes final adjustments to the collected data from each device, achieving precise alignment of data from different devices on the timeline. The aligned data can be used for subsequent analysis, monitoring, and intelligent control.

[0062] A computer-readable storage medium stores a computer program, which can be executed by a processor to implement the steps of a data time alignment method for information collection points of tobacco tube-type equipment described in the present invention.

[0063] Specifically, the data time alignment system of the tobacco tube device information collection point of the present invention includes:

[0064] (1) Data Management Center

[0065] The Data Management Center is responsible for the lifecycle management of the entire data collection and alignment process. It automatically identifies time deviations between devices and takes appropriate action to avoid analytical errors caused by data asynchrony. The Data Management Center ensures that data from each collection point is accurately recorded and stored in a chronological order, providing a reliable data foundation for subsequent time alignment.

[0066] (2) Data collection point management module

[0067] This module manages data collected from all devices, unifying timestamps and collection point numbers across all devices. Through strict time series recording, it ensures accurate data entry and ensures data integrity and traceability at each collection point. This provides a solid foundation for subsequent time alignment and analysis.

[0068] (3) Target Time Offset Management Module

[0069] Based on the workflow of different equipment and the material transfer time between devices, target time offsets are set between collection points as a preliminary reference value for data alignment. This module dynamically adjusts the time offset based on the equipment's operating characteristics and production status to ensure time synchronization within a reasonable range.

[0070] (4) Equipment data analysis module

[0071] Before time alignment, we conduct a preliminary analysis of the data collected by each device to identify outliers, noise, and other factors that may affect time alignment. Through data cleaning and preprocessing, we ensure the quality of the input data and remove outliers, providing a high-quality data foundation for high-precision time alignment.

[0072] (5) Data translation analysis module

[0073] The Data Shift module is responsible for time-shifting data from different acquisition points. By gradually adjusting the data's position on the timeline, it dynamically finds the optimal alignment. This module ensures that data from different devices is precisely aligned on the same timeline, improving the accuracy of data analysis.

[0074] (6) Pearson correlation analysis module

[0075] The Pearson correlation coefficient is used to calculate the correlation between data from different devices and evaluate the alignment of the shifted data. Through correlation analysis, the optimal time alignment point for data from different collection points is automatically found, ensuring data consistency and coherence across the time dimension.

[0076] (7) Data alignment result output module

[0077] After data translation and correlation analysis, the final aligned data is output. This aligned data facilitates subsequent analysis, production monitoring, and intelligent control. The system displays the alignment results to the user and securely stores them, ensuring accuracy and traceability.

[0078] (8) Acquisition system platform

[0079] The acquisition system platform provides the infrastructure for the whole data time alignment management, supports real-time data acquisition, processing and display. As the core platform of data alignment, it guarantees the automation and reliability of the whole data management process, and provides strong support for subsequent time alignment tasks and production optimization.

[0080] Advantages of the present application:

[0081] The present application correlates the acquisition point data of the equipment through time alignment and data translation, achieving synchronization and accuracy of analysis of different equipment data. This method improves the real-time and reliability of data analysis, providing technical support for data monitoring and optimization in industrial production process. The acquisition points of each equipment are correlated through time alignment and data translation, and through this process, synchronization and accuracy of analysis of different equipment data are achieved, providing technical support for data monitoring and optimization in industrial production process.

[0082] Through the above-mentioned multi-point data alignment method and automatic implementation scheme, the present application can effectively perform global time alignment on the acquisition point data of all tobacco cylinder equipment in the cut tobacco workshop. The system has high degree of automation, can process data in real time, reduce manual operation, and improve the efficiency and accuracy of data analysis. This method provides strong data support for production monitoring, equipment optimization and intelligent control, and has wide application prospects. BRIEF DESCRIPTION OF DRAWINGS

[0083] Figure 1 Flowchart of the method of the present application.

[0084] Figure 2 Block diagram of the system of the present application.

[0085] Figure 3 Inlet and outlet moisture before time alignment.

[0086] Figure 4 Inlet and outlet moisture before time alignment.

[0087] Figure 5 XGBOOST model outlet moisture prediction result graph before time alignment.

[0088] Figure 6 XGBOOST model outlet moisture prediction result graph before time alignment. DETAILED DESCRIPTION

[0089] The present application will be described in detail below with reference to the accompanying drawings.

[0090] Example 1

[0091] As Figure 1As shown, the present invention provides a data time alignment method for information collection points of tobacco tube equipment, which specifically includes the following steps:

[0092] Step S1: Obtain and uniformly manage the raw data collected by each device

[0093] In the tobacco-making workshop, multiple tobacco-making equipment (such as loosening and rehumidifying machines, leaf moistening feeders, and thin-plate drying machines) collect data at preset intervals. Parameters collected at these points include inlet moisture, added water volume, and outlet moisture. This data is recorded in real time by sensors and timestamped to form time series data. The collection frequency for each device can be set to 1, 2, or even longer intervals based on production needs.

[0094] Example:

[0095] Assume that the inlet moisture sensor of the loose moisture conditioning machine records data once per second, forming a data sequence A; the outlet moisture sensor of the thin plate drying machine records data once per second, forming a data sequence B.

[0096] Step S2: Calculate the initial time offset between devices

[0097] Calculate the initial time offset between devices based on the operating characteristics of each device and the material flow time.

[0098] Example:

[0099] According to the process flow, it takes about 300 seconds for the material to reach the thin plate drying machine from the loose conditioning machine. Therefore, the initial time offset is estimated to be 300 seconds.

[0100] Step S3: Perform a step-by-step translation operation on the collected data of each device

[0101] According to the target time offset, the collected data of each device is gradually shifted to ensure that the data has preliminary consistency on the same time axis.

[0102] Example:

[0103] Keeping data sequence A (inlet moisture) unchanged, perform a time shift on data sequence B (outlet moisture). The initial shift is set to 300 seconds, which means that the timestamp of data sequence B is shifted forward by 300 seconds.

[0104] Step S4: Pearson correlation analysis

[0105] The effect of data alignment was evaluated by calculating the Pearson correlation coefficient between the shifted data series.

[0106] Example:

[0107] Calculate the Pearson correlation coefficient r between data series A and the shifted data series B. If r does not meet expectations, adjust the shift amount (such as with a step size of 1 second) and recalculate the correlation coefficient until the maximum value of r is found.

[0108] Assuming that the correlation coefficient r reaches a maximum value of 0.93 when the translation time is 295 seconds, the optimal time offset is 295 seconds.

[0109] Step S5: Data adjustment and alignment

[0110] Based on the optimal time alignment point, the collected data of each device is finally adjusted to achieve accurate alignment of data from different devices on the time axis.

[0111] Example:

[0112] The timestamps of the data series B of the outlet moisture are shifted forward by 295 seconds so that they are precisely aligned with the data series A of the inlet moisture on the time axis.

[0113] Example 2

[0114] This embodiment analyzes the data aligned by embodiment 1, verifies the alignment effect, and applies it to production monitoring and intelligent control.

[0115] Data Validation:

[0116] Compare the inlet moisture and outlet moisture curves before and after alignment.

[0117] Figure 3 Showing the inlet and outlet moisture curves before alignment, the curves fail to move in sync.

[0118] Figure 4 The aligned inlet and outlet moisture curves are shown, and the curve change trends are basically consistent.

[0119] Model predictions:

[0120] The data before and after alignment were input into the XGBoost model to predict the export moisture.

[0121] Data before alignment:

[0122] The mean square error (MSE) of the prediction results is 0.027.

[0123] Figure 5 The comparison between the predicted value and the actual value is shown, and the error is large.

[0124] Aligned data:

[0125] The mean square error (MSE) of the prediction results is 0.019.

[0126] Figure 6 The comparison between the predicted value and the actual value is shown, and the error is significantly reduced.

[0127] Result analysis:

[0128] The time-aligned data reduced the forecast error by approximately 35%.

[0129] The prediction accuracy of the model is improved, and the effectiveness of the time alignment method is verified.

Claims

1. A method for aligning data time at information collection points of tobacco tube equipment, characterized in that: The following steps are involved: S1 obtains and uniformly manages the raw data collected by each device; S2 calculates the initial time offset between devices based on the operating characteristics of each device and the material flow time; S3 performs a step-by-step data shift operation on the collected data of each device according to the target time offset; S4 performed Pearson correlation analysis; S5 performs data adjustment and alignment, including: final adjustment of the collected data of each device based on the optimal time alignment point to achieve accurate alignment of data from different devices on the time axis; The aligned data is used for subsequent analysis, monitoring and intelligent control.

2. The data time alignment method according to claim 1, characterized in that: Step S2 further includes: Combining the equipment's process flow and the material transfer time between devices, a time offset estimate is provided, providing a basis for data translation processing.

3. The data time alignment method according to claim 1, characterized in that: Step S3 further includes: By adjusting the position of the data on the time axis, the consistency of data from different collection points on the same time axis can be achieved.

4. The data time alignment method according to claim 1, wherein: Step S4 further includes: The effect of data alignment was evaluated by calculating the Pearson correlation coefficient between the shifted data series; The data shifting and correlation calculation process is repeated, and the time offset is gradually adjusted until the optimal time alignment point is found that maximizes the correlation coefficient.

5. The data time alignment method according to claim 1, characterized in that: Step S3 further includes: (a) Select data series: Select two time series data A and B to be aligned and set the initial translation step size; (b) Data translation operation: Keeping data sequence A unchanged, data sequence B is translated step by step according to the set translation step size. After each translation, the time axis of data sequence B is shifted relative to data sequence A, generating a new time point correspondence relationship. (c) Generate alignment relationship: Through each translation, a new set of time point alignment relationships is generated between data sequence B and data sequence A; (d) Dynamically adjust the translation range: Dynamically adjust the number and range of translations based on actual data conditions and equipment operating characteristics.

6. The data time alignment method according to claim 5, characterized in that: Step S4 further includes: (a) Define the data sequence: Let the data sequence after translation be A = {A1, A2, …, An}, and the other data sequence be B = {B1, B2, …, Bn}; (b) Calculate the average: Calculate the average of two data series (c) Calculate the covariance Cov(A,B): (d) Calculating the Pearson correlation coefficient r: After each translation operation, the Pearson correlation coefficient is used to calculate the correlation between the data sequence A and the translated data sequence B; (e) Correlation coefficient determination: When r is close to 1, it means that the two data series have a strong positive correlation. When r is close to -1, it means that the two data series have a strong negative correlation. When r is close to 0, it means there is no obvious linear correlation between the two data series; (f) Determine the optimal translation time: After each data shift, calculate the corresponding Pearson correlation coefficient r; record the correlation coefficient at different shift times and draw a curve of the relationship between the correlation coefficient and the shift time; by analyzing the curve, find the shift time corresponding to the maximum correlation coefficient as the optimal time alignment point for the data of the two acquisition points; (g) Data alignment verification: The optimal translation time is applied to the data sequence. After time adjustment of the data, the Pearson correlation coefficient is calculated again to verify the alignment effect. If the correlation coefficient meets the preset threshold requirement, the data alignment is considered successful.

7. The data time alignment method according to claim 6, characterized in that: The calculation formula of the Pearson correlation coefficient r is: Among them, Cov(A,B) is the covariance of data sequences A and B; Var(A) and Var(B) are the variances of data sequences A and B respectively.

8. A computer-readable storage medium having a computer program stored thereon, wherein the computer program can be executed by a processor to implement the steps of the data time alignment method for information collection points of tobacco tube-type devices as described in any one of claims 1 to 7.

9. A data time alignment system for tobacco tube equipment information collection points, characterized in that: The system includes a computer and a computer-readable storage medium according to claim 8.

10. A data time alignment system for tobacco tube equipment information collection points, characterized in that: The system is used to implement a data time alignment method for information collection points of tobacco tube equipment as described in any one of claims 1 to 7, comprising: (1) Data Management Center It is used to automatically identify time deviations between devices and perform corresponding processing to avoid analysis errors caused by data asynchrony; The data management center ensures that the data of each collection point is accurately recorded and stored in time series, providing a reliable data foundation for subsequent time alignment; (2) Data collection point management module Responsible for managing the collection point data of all devices and unifying the timestamps and collection point numbers of each device; The data at each collection point is complete and traceable through time series recording; (3) Target Time Offset Management Module According to the workflow of different equipment and the material transmission time between equipment, set the target time offset between collection points as the preliminary reference value for data alignment; Ability to dynamically adjust time offset based on equipment operating characteristics and production conditions to ensure time synchronization within a reasonable range; (4) Equipment data analysis module Before time alignment, the data collected by each device is cleaned; (5) Data translation analysis module Perform time shift processing on data from different acquisition points; (6) Pearson correlation analysis module The Pearson correlation coefficient is used to calculate the correlation between data from different devices and evaluate the alignment effect of the data after translation; (7) Data alignment result output module After data translation and correlation analysis, the final aligned data results are output; (8) Acquisition system platform The acquisition system platform provides the entire infrastructure for data time alignment management, supporting real-time data acquisition, processing, and display.

Images