A PVC pipe automatic production line production data analysis processing method and system

Abstract

The application discloses a PVC pipe automatic production line production data analysis processing method and system, relates to the technical field of production data analysis, and aims to solve the technical problem that in the existing PVC pipe automatic production line, the source of the current fluctuation of the main motor of the extruder is difficult to accurately distinguish, resulting in misjudgment and the decrease of production efficiency. The method comprises the following steps: collecting real-time running current data of the main motor of the extruder, and collecting power grid voltage data and power grid frequency data corresponding to the branch where the main motor of the extruder is located in the shared power grid of the workshop; performing preset multi-scale decomposition processing on the residual signal, and extracting at least one microscopic characteristic parameter reflecting the change trend of the physical characteristics of raw materials from the decomposed signal components; and determining the fluctuation source category of the real-time running current data according to the matching result of the microscopic characteristic parameter and the preset raw material characteristic fingerprint, and the fluctuation record of the power grid voltage data or the power grid frequency data.

Description

Technical Field

[0001] This application relates to the field of production data analysis technology, and in particular to a method and system for analyzing and processing production data from an automated PVC pipe production line. Background Technology

[0002] In automated PVC pipe production lines, analyzing and processing the data generated during production is essential to ensure product quality and production efficiency. Especially when using the high-frequency current information from the extruder's main motor as an indirect measurement method to infer minute changes in the physical properties of raw materials in real time, accurately identifying the signal source becomes crucial. Summary of the Invention

[0003] This application provides a method and system for analyzing and processing production data of an automated PVC pipe production line, which aims to solve the technical problem in existing automated PVC pipe production lines where it is difficult to accurately distinguish the source of current fluctuations in the main motor of the extruder, leading to misjudgments and reduced production efficiency.

[0004] Firstly, to address the aforementioned technical problems, this invention provides a method for analyzing and processing production data from an automated PVC pipe production line. This method includes: collecting real-time operating current data of the extruder's main motor, and collecting grid voltage and frequency data corresponding to the branch of the shared power grid where the extruder's main motor is located; responding to instantaneous fluctuations in the grid voltage or frequency data, compensating the real-time operating current data based on a pre-established mapping relationship between grid disturbances and motor current response to obtain first current data, which characterizes the extruder's main motor current after excluding power quality disturbances corresponding to the grid voltage and frequency data; and based on the extruder's main motor's operating current data... The process parameter settings under the current production task are used to determine the reference operating current mode. Based on the difference between the first current data and the reference operating current mode, a residual signal is generated. The residual signal is used to characterize the abnormal current fluctuation component after excluding the normal operating current component of the extruder itself. The residual signal is subjected to preset multi-scale decomposition processing, and at least one microscopic feature parameter reflecting the trend of change in the physical properties of the raw material is extracted from the decomposed signal components. Based on the matching result of the microscopic feature parameter and the preset raw material characteristic fingerprint, as well as the fluctuation record of the grid voltage data or grid frequency data, the fluctuation source category of the real-time operating current data is determined, and the warning information corresponding to the fluctuation source category is output.

[0005] Secondly, this application provides a production data analysis and processing system for an automated PVC pipe production line. The system includes: a data acquisition unit for acquiring real-time operating current data of the extruder's main motor, and for acquiring grid voltage and frequency data corresponding to the branch where the extruder's main motor is located in the workshop's shared power grid; an acquisition unit for, in response to instantaneous fluctuations in grid voltage or frequency data, compensating the real-time operating current data according to a pre-established mapping relationship between grid disturbances and motor current response to obtain first current data, which characterizes the extruder's main motor current after excluding power quality disturbances corresponding to the grid voltage and frequency data; and a generation unit for, based on the extruder's... The process parameter settings under the current production task determine the reference operating current mode, and generate a residual signal based on the difference between the first current data and the reference operating current mode. The residual signal is used to characterize the abnormal current fluctuation component after excluding the normal operating current component of the extruder itself. The processing unit is used to perform preset multi-scale decomposition processing on the residual signal and extract at least one microscopic feature parameter reflecting the trend of change of the physical properties of the raw material from the decomposed signal components. The output unit is used to determine the fluctuation source category of the real-time operating current data based on the matching result of the microscopic feature parameter and the preset raw material characteristic fingerprint, as well as the fluctuation record of the grid voltage data or grid frequency data, and output the warning information corresponding to the fluctuation source category.

[0006] This application has at least the following beneficial effects: The PVC pipe automatic production line production data analysis and processing method disclosed in this application collects real-time operating current data of the extruder's main motor and the voltage and frequency data of the shared power grid in the workshop. Responding to instantaneous fluctuations in the power grid, the real-time operating current data is compensated based on a pre-established mapping relationship between power grid disturbances and motor current response, obtaining first current data after excluding power quality disturbances. The method further determines a reference operating current mode based on the process parameter settings under the current production task of the extruder, and generates a residual signal based on the difference between the first current data and the reference operating current mode. This residual signal characterizes the abnormal current fluctuation component after excluding the normal operating current component of the extruder itself. Subsequently, the residual signal undergoes a pre-set multi-scale decomposition process, and microscopic feature parameters reflecting the changing trends of the raw material's physical properties are extracted. Finally, based on the matching results of the microscopic feature parameters and the pre-set raw material characteristic fingerprint, as well as the power grid fluctuation records, the fluctuation source category of the real-time operating current data is determined, and corresponding early warning information is output.

[0007] Through the above technical solution, this application effectively solves the problem in the prior art of distinguishing between current fluctuations caused by external power grid disturbances and actual current fluctuations caused by changes in the physical properties of raw materials. Specifically, by accurately compensating for power grid disturbances, the influence of power grid factors on motor current can be eliminated, allowing subsequent analysis to focus on changes in the process state inside the extruder. Simultaneously, by establishing a benchmark operating current pattern and generating residual signals, the current components during normal extruder operation can be effectively filtered out, highlighting abnormal fluctuations. Furthermore, multi-scale decomposition and micro-feature parameter extraction technologies make the identification of minute changes in the physical properties of raw materials more sensitive and accurate. Combining the raw material characteristic fingerprint database and power grid fluctuation records, this application can accurately identify whether the source of current fluctuations is internal batch differences in raw materials or external power grid disturbances, thereby avoiding misjudgments and improving the reliability and practicality of production data analysis. Accordingly, unnecessary downtime for inspections can be effectively reduced, production costs lowered, and genuine quality defects can be promptly identified, avoiding batch defects and significantly improving the production efficiency and product quality control level of automated PVC pipe production lines. Attached Figure Description

[0008] Figure 1 This is a flowchart illustrating a method for analyzing and processing production data from an automated PVC pipe production line, as provided in this application. Detailed Implementation

[0009] The technical solutions of this application will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are merely some embodiments of this application, and not all embodiments. The components of this application described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but merely to illustrate selected embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0010] It should be noted that similar reference numerals and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. Furthermore, in the description of this application, terms such as "first," "second," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0011] In automated PVC pipe production lines, analyzing and processing the data generated during production is essential to ensure product quality and production efficiency. Especially when using the high-frequency current information from the extruder's main motor as an indirect measurement method to infer minute changes in the physical properties of raw materials in real time, accurately identifying the signal source becomes crucial. In actual production, current fluctuations in the extruder's main motor can be caused by various factors. On one hand, minute differences within raw material batches, such as fluctuations in molecular weight distribution or plasticizer absorption performance, can affect the melting and plasticizing behavior of the material during extrusion, leading to changes in the torque required for screw shearing and material conveying, ultimately reflected in the current fluctuations of the drive motor. This "batch variation" is information that raw material suppliers cannot provide in quality inspection reports and is beyond the current data acquisition system's scope, yet it significantly impacts product quality. On the other hand, the start-up and shutdown of other high-power equipment in the shared power grid, such as the start-up of a large crusher, can cause instantaneous fluctuations in grid voltage or frequency, i.e., power quality disturbances. These external grid disturbances can also be captured by the current sensor of the extruder's main motor, manifesting as drastic jumps in current readings. Existing data analysis systems often struggle to distinguish between current fluctuations caused by external power grid disturbances and genuine current fluctuations caused by changes in the physical properties of raw materials. When a system cannot effectively differentiate between these two highly similar current signal fluctuations, misjudgments are prone to occur. For example, current fluctuations caused by external power grid disturbances may be misjudged as sudden changes in raw material properties, leading to unnecessary downtime for inspection, resulting in decreased production efficiency and economic losses. Alternatively, genuine changes in process conditions caused by batch-to-batch variations in raw materials may be misjudged as normal fluctuations, missing genuine quality defect warnings and resulting in batches of substandard products. This overlap in signal characteristics severely reduces the reliability and practicality of production data analysis systems.

[0012] In view of the above problems, this application provides a method for analyzing and processing production data of an automated PVC pipe production line. By performing multi-dimensional analysis and compensation on the current data of the main motor of the extruder, this application effectively distinguishes the current fluctuations caused by external power grid disturbances and internal differences in raw material batches, thereby improving the accuracy and reliability of production data analysis and providing strong support for the stable operation of the PVC pipe production line and product quality control.

[0013] The following specific embodiments will provide a detailed introduction and explanation of the PVC pipe automatic production line production data analysis and processing method provided in this application.

[0014] Reference Figure 1 This application provides a method for analyzing and processing production data from an automated PVC pipe production line, which may include the following steps: S1. Collect real-time operating current data of the extruder main motor, as well as grid voltage and grid frequency data corresponding to the branch where the extruder main motor is located in the workshop shared power grid.

[0015] Among them, the "extruder main motor" refers to the core power equipment that drives the extruder screw to rotate, and the change in its operating current directly reflects the load that the screw bears during the extrusion process.

[0016] "Real-time operating current data" refers to the sequence of main motor current values ​​continuously collected by sensors during the operation of the extruder.

[0017] "Grid voltage data" and "grid frequency data" refer to the real-time monitoring values ​​of the voltage and frequency of the shared power grid in the workshop, which reflect the power quality status of the power grid.

[0018] The core of the method proposed in this application lies in the refined processing of extruder main motor current data and the accurate identification of fluctuation sources. First, it is necessary to collect real-time operating current data of the extruder main motor, as well as grid voltage and frequency data corresponding to the branch of the workshop's shared power grid where the extruder main motor is located. Real-time operating current data can be obtained by deploying current sensors on the three-phase power lines of the extruder main motor. These sensors can be Hall effect sensors or current transformers, which can convert current signals into electrical signals and digitize them at a certain sampling frequency. For example, a high-precision current transformer can be used, whose output signal is sent to a high-speed data acquisition card for sampling after passing through a conditioning circuit. Grid voltage and frequency data can be obtained by deploying power quality monitoring equipment at the main incoming switch of the workshop's shared power grid and in the distribution cabinet of the branch where the extruder main motor is located. These devices can be multi-functional power meters or professional power quality analyzers, which can monitor grid parameters such as voltage, frequency, and harmonics in real time. For example, a power quality analyzer with high-precision voltage and frequency measurement functions can be selected, and the data can be transmitted to the central control system through its communication interface.

[0019] S2. In response to instantaneous fluctuations in grid voltage or grid frequency data, the real-time operating current data is compensated based on a pre-established mapping relationship between grid disturbances and motor current response to obtain first current data. The first current data is used to characterize the extruder main motor current after excluding power quality disturbances corresponding to grid voltage and grid frequency data.

[0020] "Instantaneous fluctuations" refer to significant deviations from the normal range in grid voltage or frequency over a short period of time, usually caused by the start-up and shutdown of other high-power equipment.

[0021] The "mapping relationship between power grid disturbance and motor current response" refers to the intrinsic connection between changes in power grid voltage or frequency and changes in the current of the extruder's main motor. This relationship can be established through experiments or historical data.

[0022] The "first current data" is the current data after power grid disturbance compensation, which more purely reflects the operating status of the extruder itself.

[0023] In some embodiments, in response to instantaneous fluctuations in grid voltage or frequency data, the real-time operating current data is compensated based on a pre-established mapping relationship between grid disturbances and motor current response to obtain first current data. When grid voltage or frequency fluctuates instantaneously, the current reading of the extruder's main motor is affected. To eliminate this effect, the real-time operating current data needs to be compensated. The mapping relationship between grid disturbances and motor current response can be established through historical data or experiments. For example, when the extruder production line is in an unloaded or stable operating state, the corresponding data set of grid voltage changes, grid frequency changes, and extruder main motor current response changes can be synchronously recorded by repeatedly triggering the start-stop operations of other power equipment in the workshop. This data can be used to train a compensation model, such as a model based on regression analysis or a neural network, which can predict the current compensation amount caused by grid disturbances based on real-time grid voltage and frequency fluctuations. Then, the first current data is obtained by subtracting this current compensation amount from the real-time operating current data.

[0024] S3. Determine the reference operating current mode based on the process parameter settings under the current production task of the extruder, and generate a residual signal based on the difference between the first current data and the reference operating current mode.

[0025] The residual signal is used to characterize the abnormal current fluctuation component after excluding the normal operating current component of the extruder itself.

[0026] "Process parameter setpoints" refer to the extruder operating parameters set under specific production tasks to ensure product quality and production efficiency, such as screw speed and barrel temperature.

[0027] "Reference operating current mode" refers to the current operating mode that the main motor of the extruder should have under ideal process parameters and stable raw material conditions. It can be the expected current range or the expected current curve.

[0028] The "residual signal" is the difference between the first current data and the reference operating current pattern. It excludes the normal operating current component and highlights abnormal fluctuations.

[0029] In some embodiments, a reference operating current mode is determined based on the process parameter settings for the current production task of the extruder, and a residual signal is generated based on the difference between the first current data and the reference operating current mode. The extruder's process parameter settings, such as screw speed, barrel temperature, and melt pressure, will vary under different production tasks, and these parameters directly affect the normal operating current of the main motor. The reference operating current mode can be determined through a pre-established correspondence between process parameters and reference current. For example, based on historical production data, the average current value or current curve during stable extruder operation can be recorded under different process parameter settings and used as a reference. Once the process parameter settings for the current production task are obtained, the corresponding expected current range or expected current curve can be obtained from a preset database or calculated through a model, serving as the reference operating current mode. Then, the difference between the first current data and the reference operating current mode is calculated; this difference is the residual signal. The residual signal excludes the normal operating current component of the extruder itself, thus highlighting the abnormal current fluctuation component caused by other factors.

[0030] S4. Perform preset multi-scale decomposition processing on the residual signal, and extract at least one microscopic feature parameter reflecting the trend of change in the physical properties of the raw material from the decomposed signal components.

[0031] Among them, "preset multi-scale decomposition processing" is a signal processing technique used to decompose residual signals into components of different frequency scales in order to analyze the fluctuation characteristics of different time scales.

[0032] "Microscopic characteristic parameters" are quantitative indicators extracted from the decomposed signal components that can reflect changes in the physical properties of raw materials, such as root mean square value and kurtosis value.

[0033] In some embodiments, the residual signal undergoes a pre-defined multi-scale decomposition, and at least one microscopic feature parameter reflecting the trend of changes in the physical properties of the raw material is extracted from the decomposed signal components. The residual signal may contain multiple frequency components, which may correspond to different physical phenomena. To better analyze the minute fluctuations caused by changes in the physical properties of the raw material, the residual signal needs to be decomposed at multiple scales. For example, wavelet decomposition, empirical mode decomposition (EMD), or variational mode decomposition (VMD) can be used to decompose the residual signal into multiple signal components at different frequency scales. These decomposition operations can separate the residual signal according to frequency ranges. Then, from these signal components, target signal components within a target frequency range associated with the plasticizing cycle of the raw material in the extruder are selected. For example, through experiments or experience, it can be determined that the minute fluctuations during the plasticizing process of the raw material are mainly concentrated in a specific frequency range. Amplification gain processing of the target signal components can enhance their characteristics, making them easier to identify. Finally, microscopic feature parameters are extracted from the enhanced signal components. These parameters may be at least one of the following: root mean square value, kurtosis value, spectral energy proportion of the pre-defined frequency range, and trend change rate of the enhanced signal components. These parameters can quantify subtle trends in the physical properties of raw materials.

[0034] S5. Based on the matching results of microscopic characteristic parameters and preset raw material characteristic fingerprints, as well as the fluctuation records of grid voltage data or grid frequency data, determine the fluctuation source category of real-time operating current data, and output the early warning information corresponding to the fluctuation source category.

[0035] Among them, the "preset raw material characteristic fingerprint" is a set of microscopic characteristic parameters that are pre-established and correspond to different raw material characteristic change patterns.

[0036] "Fluctuation Source Category" refers to the specific reasons that cause fluctuations in real-time operating current data, such as differences within raw material batches or external power grid disturbances.

[0037] In some embodiments, based on the matching results of microscopic feature parameters and preset raw material characteristic fingerprints, as well as fluctuation records of grid voltage data or grid frequency data, the source category of fluctuations in real-time operating current data is determined, and warning information corresponding to the fluctuation source category is output. The preset raw material characteristic fingerprint database stores a set of microscopic feature parameters corresponding to various known raw material characteristic change patterns. The currently extracted microscopic feature parameters are matched with patterns in the fingerprint database for similarity, for example, by calculating Euclidean distance, cosine similarity, or correlation coefficient. If the matching result shows that the similarity score between the current microscopic feature parameter and a certain raw material characteristic change pattern exceeds a preset matching threshold, the fluctuation source category can be determined to be internal batch variation of the raw material. Simultaneously, a comprehensive judgment is also needed by combining the fluctuation records of grid voltage data or grid frequency data. If, before compensation processing, the real-time operating current data exhibits fluctuations exceeding a preset fluctuation range, and the grid voltage data or grid frequency data shows records of instantaneous fluctuations, while the microscopic feature parameters remain below a preset stability threshold, the fluctuation source category can be determined to be external grid disturbance. Based on the identified source of the fluctuation, the system will output corresponding early warning information, such as "early warning of internal differences in raw material batches" or "early warning of external power grid disturbance," so that operators can take timely measures.

[0038] In summary, this application effectively solves the problem of unclear identification of current fluctuation sources in the prior art by introducing technical means such as power grid disturbance compensation, residual signal generation, multi-scale decomposition and micro-feature parameter extraction, and multi-factor comprehensive judgment. It significantly improves the accuracy and reliability of production data analysis of PVC pipe automatic production lines, and provides strong technical support for achieving more precise quality control and more efficient production management.

[0039] Specifically, the above-mentioned collection of real-time operating current data of the extruder main motor, as well as the collection of grid voltage and grid frequency data in the workshop shared power grid corresponding to the branch where the extruder main motor is located, can be carried out in the following manner.

[0040] The acquisition of real-time operating current data of the extruder main motor, as well as the acquisition of grid voltage and frequency data in the workshop shared power grid corresponding to the branch where the extruder main motor is located, includes: A current sensor is deployed on the three-phase power line of the main motor of the extruder, and the real-time operating current data is acquired through the current sensor at the first sampling frequency. The first sampling frequency is set to a frequency range that can distinguish the details of the current fluctuation of the main motor of the extruder when the physical properties of the raw material change. Power quality monitoring devices are deployed at the main incoming switch of the shared power grid in the workshop and in the distribution cabinet of the branch where the main motor of the extruder is located. The power quality monitoring devices synchronously acquire the grid voltage data and the grid frequency data at the second sampling frequency. The second sampling frequency is set to a frequency range that can capture instantaneous power quality changes caused by the start-up and shutdown of other power equipment in the workshop.

[0041] Specifically, current sensors are deployed on the three-phase power lines of the extruder's main motor to directly and accurately monitor the real-time operating current of the main motor. These current sensors are configured to acquire real-time operating current data at a first sampling frequency. The setting of this first sampling frequency is crucial; it is set to a frequency range capable of resolving the details of current fluctuations in the extruder's main motor as the physical properties of the raw materials change. This means that the sampling frequency is high enough to capture the minute, rapid fluctuations in the motor load current caused by subtle changes in the physical properties of the raw materials during melting, plasticizing, and mixing, thus providing a high-resolution data foundation for subsequent raw material characteristic analysis.

[0042] Furthermore, to comprehensively monitor the impact of the power grid environment on the extruder's operation, power quality monitoring equipment is deployed at the main incoming switch of the shared power grid in the workshop and in the distribution cabinet of the branch line where the extruder's main motor is located. Through these power quality monitoring devices, grid voltage and frequency data can be acquired simultaneously at a second sampling frequency. This second sampling frequency is set to a range capable of capturing instantaneous power quality changes caused by the start-up and shutdown of other high-power equipment in the workshop. This means that this sampling frequency can effectively record instantaneous disturbances in the power grid, such as voltage drops, surges, and frequency shifts. These disturbances are typically caused by the frequent start-up and shutdown operations of other high-power equipment in the workshop, directly affecting the operating current of the extruder's main motor.

[0043] Through the above technical solution, this application can achieve accurate and high-resolution acquisition of extruder main motor operating current data and power grid power quality data. Specifically, by deploying current sensors on the three-phase power lines of the main motor and employing high-frequency sampling, it ensures the capture of subtle current fluctuations caused by changes in the physical properties of the raw materials, providing sufficient detailed information for subsequent raw material characteristic analysis. Simultaneously, by deploying power quality monitoring equipment in the shared power grid main switch and the motor branch distribution cabinet, and employing sampling frequencies capable of capturing instantaneous changes, it ensures the synchronous and accurate acquisition of power grid voltage and frequency data. This refined data acquisition strategy effectively improves the quality and reliability of the raw data, providing a solid data foundation for subsequent power grid disturbance compensation, residual signal generation, and fluctuation source classification, thereby significantly improving the accuracy and effectiveness of the entire production data analysis and processing method.

[0044] Specifically, the steps for responding to instantaneous fluctuations in grid voltage or frequency data and compensating for real-time operating current data based on a pre-established mapping relationship between grid disturbances and motor current response to obtain the first current data include: comparing grid voltage data with a preset voltage fluctuation threshold and comparing grid frequency data with a preset frequency fluctuation threshold; determining that an instantaneous fluctuation has occurred when the change in grid voltage data exceeds the preset voltage fluctuation threshold or the change in grid frequency data exceeds the preset frequency fluctuation threshold; calling the mapping relationship between grid disturbances and motor current response, and determining the current compensation amount caused by grid disturbances based on the real-time monitored changes in grid voltage or grid frequency data; the mapping relationship between grid disturbances and motor current response is a set of corresponding data records between the grid voltage change, grid frequency change, and extruder main motor current response change corresponding to each operation, which is generated by triggering the start-stop operation of other power equipment in the workshop multiple times when the extruder production line is in an unloaded or stable operating state, and is constructed based on the set of dynamic current response coefficients determined by the corresponding data set; and subtracting the current compensation amount from the real-time operating current data to obtain the first current data.

[0045] The comparison of grid voltage data with preset voltage fluctuation thresholds and grid frequency data with preset frequency fluctuation thresholds aims to establish a benchmark for judging instantaneous grid fluctuations. The preset voltage and frequency fluctuation thresholds can be set according to the power quality standards of the actual production environment and the extruder's sensitivity to grid fluctuations. For example, they can be set to voltage changes exceeding ±5% or frequency changes exceeding ±0.5Hz.

[0046] Furthermore, when the change in grid voltage data exceeds a preset voltage fluctuation threshold, or the change in grid frequency data exceeds a preset frequency fluctuation threshold, the system determines that an instantaneous fluctuation has occurred. This determination mechanism ensures that subsequent compensation processing is only triggered when the grid disturbance reaches a certain level sufficient to affect the operation of the extruder's main motor, thus avoiding over-response to minor, insignificant fluctuations.

[0047] Specifically, the mapping relationship between power grid disturbances and motor current response is the core of this scheme. This mapping relationship is established by artificially triggering the start-stop operations of other power equipment in the workshop when the extruder production line is in an unloaded or stable operating state. During this process, the changes in power grid voltage, power grid frequency, and extruder main motor current response corresponding to each operation are recorded simultaneously. Through a large amount of experimental data, a corresponding data set can be constructed, and a set of dynamic current response coefficients can be determined from it. These coefficients can accurately characterize the expected response of the extruder main motor current under different levels of power grid disturbances. For example, when the power grid voltage drops by X%, the motor current usually increases by Y amperes; when the power grid frequency fluctuates by Z Hz, the motor current will change by W amperes.

[0048] Therefore, when real-time changes in grid voltage or frequency data are detected, the system can invoke a pre-established mapping relationship to accurately calculate the current compensation amount caused by grid disturbances based on the currently monitored fluctuations. This compensation amount represents the difference between the motor current value that would have been reached without grid disturbances and the current real-time operating current data.

[0049] Finally, the determined current compensation amount is subtracted from the real-time operating current data to obtain the first current data. This first current data effectively eliminates the influence of power quality disturbances corresponding to grid voltage and frequency data on the extruder main motor current, thus more accurately reflecting the true operating status of the extruder main motor after eliminating external grid interference.

[0050] Through the above technical solution, this application can accurately identify and effectively compensate for fluctuations caused by external power grid disturbances in the real-time operating current data of the extruder main motor. This significantly improves the accuracy and reliability of the first current data, enabling it to more realistically characterize the operating state of the extruder main motor after eliminating power quality disturbances. Consequently, the accuracy of subsequent analysis steps based on the first current data, such as residual signal generation, micro-feature parameter extraction, and fluctuation source classification, will be greatly improved, avoiding misjudgments or omissions caused by power grid fluctuations. This allows the PVC pipe automated production line data analysis and processing method to more accurately diagnose production problems, improving production efficiency and product quality.

[0051] In some embodiments described above, a reference operating current pattern is determined based on the process parameter settings under the current production task of the extruder, and a residual signal is generated based on the difference between the first current data and the reference operating current pattern. However, in actual production environments, the operating current of the extruder's main motor is affected by a combination of various process parameters. If key process parameters are not accurately identified and utilized to construct the reference operating current pattern, the generated residual signal may contain a large amount of normal operation components, thereby reducing the sensitivity and accuracy of identification to abnormal current fluctuation components.

[0052] In this regard, this application further proposes the following steps: determining the reference operating current mode based on the process parameter settings under the current production task of the extruder, and generating a residual signal based on the difference between the first current data and the reference operating current mode, including: The process parameter setting value is obtained as at least one of the following under the current production task of the extruder: the screw speed setting value, the temperature setting value of each heating zone of the barrel, and the melt pressure setting value. Based on the pre-established correspondence between process parameters and reference current, the expected current range or expected current curve corresponding to the process parameter setting value is determined as the reference operating current mode. Calculate the difference between the first current data and the reference operating current mode, and use the difference as the residual signal.

[0053] Specifically, the process parameter settings refer to the key operating parameters that directly affect the extruder's working status and energy consumption during PVC pipe production. The screw speed setting directly determines the material's conveying speed and shear strength within the extruder; the temperature settings of each heating zone in the barrel control the plasticizing temperature distribution of the material; and the melt pressure setting reflects the resistance at the extrusion die and the material's density. Any change in these parameters will directly or indirectly affect the load current of the extruder's main motor. By obtaining the settings of these key parameters, the expected operating status of the extruder under specific production tasks can be more accurately reflected.

[0054] The pre-established correspondence between process parameters and reference current can be understood as a mapping relationship between the setpoints of process parameters and the current values ​​of the extruder's main motor under normal and stable operating conditions, established through historical data analysis, experimental testing, or physical modeling. For example, when the extruder is in a stable production state, the average operating current or current fluctuation range of the main motor under different combinations of screw speed, barrel temperature, and melt pressure is recorded, thereby constructing a multi-dimensional lookup table or regression model. This correspondence is used to predict the expected normal operating current pattern of the extruder's main motor given the setpoints of the process parameters.

[0055] In practical applications, the reference operating current mode can be the expected current range corresponding to the process parameter setting value, i.e., the upper and lower limits of the main motor current under normal production conditions; or it can be the expected current curve, i.e., the ideal trajectory of the main motor current changing over time. By comparing the first current data with the reference operating current mode, the difference can be accurately calculated. This difference is the residual signal, which aims to characterize the abnormal current fluctuation component after excluding the normal operating current component of the extruder itself, thereby more effectively highlighting current fluctuations caused by abnormal factors such as changes in raw material characteristics.

[0056] By employing the aforementioned technical solution to precisely define the baseline operating current mode using key process parameters such as screw speed, barrel temperature, and melt pressure, the generated residual signal can more accurately exclude the normal operating current components of the extruder itself. This significantly improves the sensitivity and identification accuracy of the residual signal to abnormal current fluctuations. This helps to detect subtle fluctuations caused by differences within raw material batches earlier and more accurately during production, avoiding misjudging current changes caused by normal process adjustments as abnormal. Consequently, it enhances the reliability of production data analysis and processing and the timeliness of early warnings in automated PVC pipe production lines.

[0057] In some preferred embodiments, a specific example is given below. Assume that when producing a PVC pipe of a certain specification, the process engineer sets the screw speed to 50 RPM, the barrel heating zone temperatures to 180°C, 190°C, 195°C, and 190°C respectively, and the melt pressure to 25 MPa. During stable extruder operation, a database of the correspondence between process parameters and reference currents has been established through historical data analysis and experimental verification. When the above process parameter settings are input, the database outputs an expected current curve. For example, under stable conditions, the main motor current should fluctuate within the range of 150A ± 5A, exhibiting a specific periodic small fluctuation pattern. In actual production, the collected first current data is compared with this expected current curve in real time. For example, if the first current data is 160A at a certain moment, and the expected current curve corresponds to a current value of 150A at that moment, the calculated difference is 10A. This 10A difference constitutes part of the residual signal. By continuously calculating these differences, a complete residual signal curve can be generated. This curve will mainly reflect abnormal current fluctuations after excluding the influence of normal process operation, such as current fluctuations caused by changes in the melting characteristics of raw materials.

[0058] In some embodiments described above in this application, a pre-defined multi-scale decomposition process is proposed to be performed on the residual signal, and at least one microscopic feature parameter reflecting the trend of changes in the physical properties of the raw material is extracted from the decomposed signal components. However, in practical applications, the residual signal may contain fluctuation components from multiple sources. If no targeted processing is performed and features are directly extracted from all the decomposed signal components, it may lead to insufficient accuracy in identifying changes in the physical properties of the raw material, or it may be easily interfered with by other non-raw material factors.

[0059] To address this, this application further proposes the following steps for performing a preset multi-scale decomposition on the residual signal and extracting at least one microscopic feature parameter reflecting the changing trend of the raw material's physical properties from the decomposed signal components: Performing a preset multi-scale decomposition operation on the residual signal to obtain multiple signal components at different frequency scales, wherein the preset multi-scale decomposition operation is used to separate the residual signal according to frequency ranges; specifically, the preset multi-scale decomposition operation can be understood as a signal processing technique aimed at decomposing the original residual signal into sub-signals with specific characteristics at different frequencies or time scales. For example, wavelet transform, empirical mode decomposition (EMD), or variational mode decomposition (VMD) methods can be used. Through this decomposition, the fluctuation components at different frequency ranges in the residual signal can be effectively isolated, laying the foundation for subsequent detailed analysis. From the multiple signal components at different frequency scales, a target signal component within a target frequency range associated with the plasticizing cycle of the raw material in the extruder is selected; wherein, the target frequency range associated with the plasticizing cycle of the raw material in the extruder refers to a frequency interval predetermined based on the working principle of the extruder and the physical properties of the raw material. For example, during PVC extrusion, the melting, mixing, and conveying of raw materials generate specific mechanical vibrations and current fluctuations, which are typically concentrated within a specific low- or mid-frequency range. By selecting signal components within this target frequency range, high-frequency noise or low-frequency drift unrelated to changes in raw material properties can be effectively filtered out, thus focusing on the signal that truly reflects the plasticization state of the raw material. Amplitude gain processing is then applied to the target signal component to obtain an enhanced signal component; in practical applications, amplitude gain processing refers to amplifying the selected target signal component. The purpose is that minute changes in the physical properties of the raw material may result in small amplitude fluctuations in the residual signal, making them difficult to detect directly or their characteristics inconspicuous. By applying a preset gain coefficient, the amplitude of these weak signals can be effectively increased, making them more prominent in subsequent feature extraction and analysis, thereby improving sensitivity to subtle changes in the raw material.

[0060] The micro-feature parameters are extracted from the enhanced signal components. The micro-feature parameters include at least one of the following: root mean square value, kurtosis value, spectral energy ratio of a preset frequency range, and trend rate of change of the enhanced signal components.

[0061] Among these, the microscopic characteristic parameters are indicators used to quantify the characteristics of enhanced signal components. The root mean square (RMS) value reflects the effective energy or average power of the signal, and its variation may indicate changes in the viscosity or friction of the raw material. Kurtosis characterizes the sharpness or impact of the signal distribution, which is important for detecting the presence of abnormal particles or agglomeration in the raw material. The spectral energy proportion value of a preset frequency range can reveal the relative intensity of a specific frequency component in the entire signal; for example, changes in the energy proportion of certain frequency bands may be related to the melting state or mixing homogeneity of the raw material. The trend change rate is used to capture long-term, slow changes in the signal, which may reflect the overall differences between batches of raw material or the degree of aging. By extracting these multi-dimensional characteristic parameters, the changing trends of the physical properties of the raw material can be more comprehensively and accurately characterized.

[0062] Through the above technical solution, this application overcomes the limitations of traditional methods in accurately identifying changes in the physical properties of raw materials from complex residual signals. Specifically, by selectively choosing target frequency components related to the plasticizing cycle of the raw materials and performing amplitude gain processing, the detection sensitivity and accuracy of minute changes in the properties of the raw materials are significantly improved. This allows for earlier and more precise detection of differences or anomalies within raw material batches, providing strong technical support for timely adjustments to the production process and quality control, effectively preventing decreased production efficiency and product quality defects caused by raw material issues.

[0063] In some preferred embodiments, a specific example is given below. Assume that during the PVC pipe production process, the initial current data of the extruder's main motor has been obtained and compared with a reference operating current pattern to generate a residual signal. To accurately identify the components in the residual signal reflecting changes in the physical properties of the raw materials, the following steps can be taken: First, wavelet decomposition is performed on the residual signal, for example, by using the Daubechies wavelet function to decompose the signal into multiple layers, thereby obtaining a series of wavelet coefficients at different frequency scales. These wavelet coefficients represent the energy distribution of the residual signal in different frequency ranges.

[0064] Secondly, based on the typical cycle of the plasticizing and melting process of PVC raw materials under the action of the extruder screw, a target frequency range is determined. For example, this range may be related to the harmonic frequency of the extruder screw speed or the characteristic frequency of melt pressure fluctuation. From the decomposed wavelet coefficients, the signal component corresponding to this target frequency range is selected as the target signal component.

[0065] Next, the selected target signal component is subjected to amplitude gain processing, for example, by multiplying it by a preset gain coefficient (such as 2 or 3 times) to enhance any weak fluctuation characteristics that may exist within it.

[0066] Finally, microscopic characteristic parameters are extracted from the enhanced signal components. For example, the root mean square value of the enhanced signal component is calculated to reflect its overall energy level; the kurtosis value is calculated to characterize the sharpness or impact of its signal distribution; the spectral energy proportion of a specific frequency range (e.g., 0.5Hz-2Hz) is calculated to reflect the activity level of that frequency band; and its trend change rate is calculated to capture long-term or slow changes. The combination of these parameters can comprehensively and sensitively reflect subtle changes in the physical properties of the raw materials, such as differences in viscosity, particle size distribution, or moisture content between batches, thus providing crucial data for subsequent determination of the source of fluctuations.

[0067] Specifically, in the steps of performing preset multi-scale decomposition on the residual signal and extracting at least one microscopic feature parameter reflecting the trend of change in the physical properties of the raw material from the decomposed signal components, the preset multi-scale decomposition operation is an operation of performing multi-level decomposition on the residual signal using a preset mother wavelet function.

[0068] Multiscale decomposition is a signal processing technique that aims to break down a complex signal into multiple simpler components at different scales or frequency ranges. This decomposition helps reveal hidden patterns and features in the signal, and is particularly suitable for the analysis of non-stationary signals. In the analysis of PVC pipe production data, the residual signal of the extruder main motor current may contain multiple frequency components, each corresponding to different physical phenomena, such as minute changes in raw materials, equipment vibration, and fluctuations in process parameters. Through multiscale decomposition, these signal components of different frequencies can be separated, thereby enabling more accurate analysis of the microscopic features related to changes in the physical properties of the raw materials. The preset mother wavelet function is the basis for wavelet decomposition, and its selection is crucial to the decomposition effect. For example, Daubechies wavelet, Symlets wavelet, Coiflets wavelet, or Mexican Hat wavelet can be selected based on the characteristics of the residual signal and the desired decomposition effect. Different mother wavelet functions have different support lengths, symmetries, and regularities, making them suitable for capturing different types of signal features. By selecting an appropriate mother wavelet function, the residual signal can be effectively decomposed into a series of wavelet coefficients, which represent the local features of the signal at different scales and locations. Multilevel decomposition refers to the step-by-step decomposition of a signal. Each level of decomposition further decomposes the approximate components of the previous level into lower-frequency approximate components and higher-frequency detail components. Through this iterative decomposition, signal components at multiple different frequency scales can be obtained, thereby enabling refined analysis of the residual signal.

[0069] By employing the aforementioned technical solution and using a preset mother wavelet function to perform multi-level decomposition of the residual signal, a more refined and comprehensive time-frequency analysis of the residual signal can be achieved. Compared to other multi-scale decomposition methods, wavelet decomposition has significant advantages in processing non-stationary signals and capturing local transient features. It can more effectively separate micro-fluctuations related to changes in the physical properties of raw materials from other interferences in the residual signal. This allows the subsequently extracted micro-feature parameters to more accurately reflect the true state changes of the raw materials, thereby improving the accuracy and sensitivity of fluctuation source classification. This helps to promptly identify and resolve raw material problems in the production process, improving product quality and production efficiency.

[0070] In some embodiments described above in this application, a method for determining the source category of fluctuations in real-time operating current data is proposed based on the matching results of microscopic characteristic parameters and preset raw material characteristic fingerprints, as well as fluctuation records of grid voltage data or grid frequency data. Specifically, to more accurately identify and distinguish the specific causes of current fluctuations in the extruder main motor, this application further proposes the following method for determining the source category of fluctuations.

[0071] Based on the matching results of microscopic characteristic parameters and preset raw material characteristic fingerprints, and the fluctuation records of power grid voltage data or power grid frequency data, the fluctuation source categories of real-time operating current data are determined, including: The microscopic feature parameters are matched with multiple raw material characteristic change patterns stored in a preset raw material characteristic fingerprint database based on similarity. When the similarity score between the micro-feature parameter and at least one raw material characteristic change pattern in the preset raw material characteristic fingerprint database exceeds a preset matching threshold, the fluctuation source category of the real-time operating current data is determined to be the internal difference of the raw material batch. When the microscopic characteristic parameters remain below a preset stable threshold, and the real-time operating current data exhibits fluctuations exceeding a preset fluctuation range before compensation processing, and the grid voltage data or the grid frequency data contains records of the instantaneous fluctuations, the fluctuation source category of the real-time operating current data is determined to be external grid disturbance.

[0072] Specifically, when determining the source category of fluctuations in real-time operating current data, the microscopic feature parameters obtained through multi-scale decomposition are first matched with a set of pre-stored raw material characteristic change patterns in a pre-defined raw material characteristic fingerprint database. This pre-defined raw material characteristic fingerprint database can be understood as containing a set of typical microscopic feature parameters from different raw material batches, suppliers, or storage conditions, all associated with specific raw material characteristic change patterns. Similarity matching aims to quantify the degree of closeness between the current microscopic feature parameters and known raw material characteristic patterns.

[0073] Specifically, when the similarity score between the microscopic feature parameters and at least one raw material characteristic change pattern in the preset raw material characteristic fingerprint database exceeds a preset matching threshold, it can be determined that the fluctuation of the real-time operating current data mainly originates from the internal differences within the raw material batch. This indicates that the fluctuation characteristics of the extruder main motor current highly match a certain known raw material characteristic change pattern, such as changes in the melt index, density, and moisture content of the raw material.

[0074] Furthermore, in another scenario, where the microscopic characteristic parameters remain below a preset stability threshold, while the real-time operating current data exhibits fluctuations exceeding a preset fluctuation range before compensation processing, and there are records of instantaneous fluctuations in the grid voltage or frequency data, then the source of the real-time operating current data fluctuations can be determined to be external grid disturbances. The preset stability threshold is used to determine whether the microscopic characteristic parameters significantly deviate from the normal range; if they remain stable, it indicates that the raw material characteristics have not changed significantly. The fluctuations in the real-time operating current data before compensation processing and the existence of records of instantaneous grid fluctuations directly point to grid quality issues as the root cause of the current fluctuations.

[0075] This application's solution, through the establishment of a clear judgment logic, can effectively distinguish the two main sources of current fluctuations in the extruder's main motor: internal differences in raw material batches and external power grid disturbances. By performing similarity matching between the microscopic feature parameters extracted from the residual signal and a pre-set raw material characteristic fingerprint database, current fluctuation patterns related to changes in raw material characteristics can be directly identified. This is based on the principle that different raw material characteristics cause subtle changes in motor load during extrusion; these changes are captured and quantified into microscopic feature parameters by the multi-scale decomposition of the residual signal. When these parameters highly match the known raw material fingerprint, the problem can be accurately attributed to the raw material.

[0076] On the other hand, when the microscopic characteristic parameters remain stable, indicating that the raw material properties have not changed significantly, the system will then check the fluctuations in the original current data before compensation processing, as well as the instantaneous fluctuation records of the grid voltage or frequency. Since the first current data has already eliminated the influence of grid disturbances, if the original current data fluctuated before compensation, and the grid also has instantaneous fluctuation records, it can be clearly determined that the fluctuation was caused by external grid disturbances. This dual judgment mechanism utilizes the differences in signal characteristics and external environment records of different fluctuation sources, thereby achieving precise location of the fluctuation source.

[0077] Through the above technical solution, this application can accurately identify and classify the sources of current fluctuations in the main motor of the extruder in an automated PVC pipe production line. Specifically, by introducing a similarity matching mechanism between microscopic feature parameters and a preset raw material characteristic fingerprint database, current fluctuations caused by differences within raw material batches can be effectively identified, thus providing data support for raw material quality control during the production process. Simultaneously, by combining the stability assessment of microscopic feature parameters with records of power grid voltage or frequency fluctuations, external power grid disturbances can be accurately distinguished as sources of current fluctuations, avoiding misjudging power grid problems as process or raw material issues. This refined classification of fluctuation sources helps production managers quickly locate the root cause of problems, take targeted measures, improve production line stability and product quality, and reduce unnecessary downtime and debugging time.

[0078] In some embodiments described above in this application, a method is proposed to determine the source category of fluctuations in real-time operating current data based on the matching results of microscopic feature parameters and preset raw material characteristic fingerprints. However, in its implementation, if the establishment of the preset raw material characteristic fingerprint database lacks systematicity and accuracy, it may lead to insufficient reliability of the matching results, thereby affecting the accurate identification of fluctuation sources such as internal differences within raw material batches.

[0079] In this regard, this application further proposes that the preset raw material characteristic fingerprint database be established in the following way: While the extruder is running, record the historical residual signals when the known properties of the raw materials change; Perform the same decomposition operation as the above-mentioned preset multi-scale decomposition process on the historical residual signal, and extract the historical micro-feature parameter sets corresponding to each of the various raw material characteristic change patterns; The set of historical microscopic feature parameters is associated with and stored with the corresponding raw material characteristic change types to form the preset raw material characteristic fingerprint database.

[0080] Specifically, recording historical residual signals when known raw material properties change during extruder operation refers to the process of synchronously collecting real-time operating current data of the extruder's main motor when known changes occur in the physical properties of the raw material (such as melt index, density, additive content, and recycled material ratio). This data is then compensated, compared with a baseline model, and used to generate corresponding residual signals. These residual signals characterize the abnormal fluctuations in the extruder's main motor current under specific raw material property changes. "Known changes in raw material properties" can be understood as events explicitly recorded by production personnel or the quality control system, such as raw material batch changes, supplier changes, formula adjustments, or raw material quality fluctuations.

[0081] Specifically, the historical residual signal undergoes the same decomposition operation as the aforementioned preset multi-scale decomposition process. From the decomposed signal components, historical micro-feature parameter sets corresponding to each of the various raw material characteristic change patterns are extracted. This aims to ensure that the feature parameters used to establish the fingerprint database are consistent and comparable with the feature parameters extracted during actual online monitoring. The historical micro-feature parameter sets may include at least one of the following: root mean square value, kurtosis value, spectral energy proportion value of the preset frequency range, and trend change rate of the enhanced signal component. These parameters can effectively characterize the current fluctuation characteristics caused by changes in the physical properties of the raw materials.

[0082] In practical applications, the historical micro-feature parameter set is associated with and stored in relation to the corresponding raw material characteristic change types to form the preset raw material characteristic fingerprint database. This means binding the historical micro-feature parameter set, after decomposition and feature extraction, with the specific raw material characteristic change types that caused these feature parameter changes (e.g., "new batch of raw material A", "recycled material ratio increased by 5%", "additive B content decreased", etc.) and storing it in a queryable database. This fingerprint database serves as a reference benchmark for subsequent real-time data matching.

[0083] Through the above technical solution, this application can establish an accurate, reliable, and representative pre-defined raw material characteristic fingerprint database. This significantly improves the accuracy and reliability of judging the source of current fluctuations in the extruder main motor, especially when identifying fluctuations caused by differences within raw material batches. This solution avoids misjudgments or omissions caused by an incomplete or inaccurate fingerprint database, thereby enabling the production line to respond to raw material quality issues more promptly and accurately, optimize production process control, reduce scrap rates, and improve product quality stability.

[0084] In some preferred embodiments, a specific example is given below. Suppose a PVC pipe production line needs to establish a characteristic fingerprint database to identify differences within raw material batches during production. First, with the extruder operating stably, a batch of PVC raw materials with a known melt flow index (MFI) slightly higher than the standard value is introduced. During production line operation, real-time operating current data of the extruder's main motor is continuously collected, and compensation processing and residual signal generation are performed according to the method described above. When a specific fluctuation pattern is observed in the residual signal, these residual signals are recorded and marked as "high MFI raw material batch". Subsequently, the same multi-scale decomposition operation as online monitoring is performed on these historical residual signals, and micro-feature parameters such as root mean square value, kurtosis value, and spectral energy ratio value in a specific frequency range are extracted to form a set of historical micro-feature parameters. Then, this parameter set is associated with and stored as the raw material characteristic change type "high MFI raw material batch". Similarly, a batch of PVC raw materials with a known density slightly lower than the standard value can be introduced, and the above process is repeated to establish a characteristic fingerprint for "low-density raw material batch". By systematically recording, decomposing, and extracting features from various known changes in raw material properties, and storing these changes in association with corresponding change types, a comprehensive and accurate pre-defined raw material property fingerprint database can be formed. When unknown current fluctuations occur during subsequent production, matching the extracted microscopic feature parameters with this fingerprint database in real time can quickly and accurately identify whether the fluctuations are caused by internal differences within a known batch of raw materials.

[0085] In some embodiments described above in this application, the real-time operating current data of the extruder's main motor and the voltage and frequency data of the workshop's shared power grid are collected, and the sampling frequency is initially set. However, in practical applications, if the sampling frequencies for different types of data are not carefully designed and coordinated, data acquisition efficiency may be low, or the subtle characteristics unique to different fluctuation sources may not be effectively captured, thus affecting the accuracy and real-time performance of subsequent analysis. For example, if the sampling frequency used to monitor motor current is insufficient to capture the rapid dynamic changes in raw material characteristics, or if the sampling frequency used to monitor power grid data cannot effectively capture instantaneous power quality disturbances, it may lead to misjudgments or missed judgments.

[0086] In response, this application further proposes that the method described above, which involves collecting real-time operating current data of the extruder main motor and collecting grid voltage and frequency data corresponding to the branch where the extruder main motor is located in the workshop shared power grid, includes: Real-time operating current data of the main motor of the extruder is collected at a first sampling frequency, and grid voltage data and grid frequency data corresponding to the branch where the main motor of the extruder is located in the workshop shared power grid are collected at a second sampling frequency; wherein, the first sampling frequency is higher than the second sampling frequency, the first sampling frequency is in the kilohertz range, and the second sampling frequency is in the kilohertz range.

[0087] Specifically, the first sampling frequency refers to the frequency used to acquire real-time operating current data of the extruder's main motor, and it is set to the kilohertz range, such as 2kHz, 5kHz, or even higher. This high sampling frequency aims to ensure sufficient capture of subtle fluctuations and high-frequency components in the extruder's main motor current caused by minute changes in the physical properties of the raw materials (such as melt state, viscosity, and uniformity). These subtle fluctuations are often early indicators of batch-to-batch differences or raw material quality deterioration. The second sampling frequency refers to the frequency used to synchronously acquire voltage and frequency data from the workshop's shared power grid, and it is also set to the kilohertz range, such as 1kHz or 2kHz. Although its range is the same as the first sampling frequency, the second sampling frequency is set lower. This design aims to balance the real-time nature of data acquisition with the amount of data, ensuring effective capture of instantaneous power quality disturbances caused by the start-up and shutdown of other high-power equipment and sudden load changes in the workshop, while avoiding the acquisition of excessive redundant data. By setting the second sampling frequency to the kilohertz range, sufficient resolution for transient changes in the power grid, such as voltage dips, swells, and frequency shifts, can be ensured.

[0088] Through the above technical solution, this application can achieve efficient and accurate acquisition of extruder main motor operating current data and power grid power quality data. Setting the first sampling frequency higher than the second sampling frequency, and keeping both in the kilohertz range, ensures sufficient resolution for subtle fluctuations in motor current reflecting changes in the physical properties of raw materials, while also effectively capturing instantaneous disturbances in the power grid. This differentiated high-frequency sampling strategy significantly improves the quality and information content of the raw data, laying a solid foundation for subsequent power grid disturbance compensation, residual signal generation, and microscopic feature parameter extraction. Therefore, this application can more accurately and reliably identify whether fluctuations in real-time operating current data originate from internal differences in raw material batches or external power grid disturbances, thus providing a more precise basis for fault diagnosis and process optimization in PVC pipe automated production lines. It effectively avoids misjudgments or information loss caused by sampling frequency mismatch, improving the robustness and effectiveness of the entire production data analysis and processing method.

[0089] In some preferred embodiments, this application is implemented as follows: Assuming a PVC pipe automated production line, the real-time operating current data of the extruder's main motor is collected via current sensors deployed on the three-phase power lines. To capture subtle current fluctuations caused by changes in the physical properties of the raw material, such as its melt state and viscosity, during extrusion, the first sampling frequency is set to 5 kHz. This means that 5000 current data points will be collected per second, sufficient to distinguish high-frequency current components within the hundreds of hertz range. Simultaneously, voltage and frequency data of the workshop's shared power grid are collected via power quality monitoring equipment deployed in the main incoming switch and distribution cabinet. To effectively capture transient power quality disturbances such as voltage dips, spikes, or frequency shifts caused by the start-up and shutdown of other high-power equipment (such as injection molding machines and air compressors) in the workshop, the second sampling frequency is set to 2 kHz. Although lower than the first sampling frequency, a sampling rate of 2 kHz still provides sufficient time resolution to accurately record the occurrence and duration of these transient events. With this configuration, when fluctuations occur in the extruder's main motor current, the system first analyzes its internal characteristics using high-resolution motor current data. Simultaneously, it combines this with relatively low but still sufficient grid data to capture transients, performing grid disturbance compensation. For example, if high-frequency, low-amplitude continuous fluctuations are detected in the motor current data, while the grid data remains stable over the same time period, it can be more confidently determined that the fluctuation originates from changes in raw material characteristics. Conversely, if fluctuations in the motor current occur synchronously with instantaneous large changes in grid voltage or frequency, they can be accurately attributed to external grid disturbances. This differentiated and high-precision sampling strategy enables the system to more effectively separate and identify interference from different sources, thereby improving the accuracy and reliability of production data analysis.

[0090] In automated PVC pipe production lines, analyzing and processing the data generated during production is essential to ensure product quality and production efficiency. Especially when using the high-frequency current information from the extruder's main motor as an indirect measurement method to infer minute changes in the physical properties of raw materials in real time, accurately identifying the signal source becomes crucial. In actual production, current fluctuations in the extruder's main motor can be caused by various factors. On one hand, minute differences within raw material batches, such as fluctuations in molecular weight distribution or plasticizer absorption performance, can affect the melting and plasticizing behavior of the material during extrusion, leading to changes in the torque required for screw shearing and material conveying, ultimately reflected in the current fluctuations of the drive motor. This "batch variation" is information that raw material suppliers cannot provide in quality inspection reports and is beyond the current data acquisition system's scope, yet it significantly impacts product quality. On the other hand, the start-up and shutdown of other high-power equipment in the shared power grid, such as the start-up of a large crusher, can cause instantaneous fluctuations in grid voltage or frequency, i.e., power quality disturbances. These external grid disturbances can also be captured by the current sensor of the extruder's main motor, manifesting as drastic jumps in current readings. Existing data analysis systems often struggle to distinguish between current fluctuations caused by external power grid disturbances and genuine current fluctuations caused by changes in the physical properties of raw materials. When a system cannot effectively differentiate between these two highly similar current signal fluctuations, misjudgments are prone to occur. For example, current fluctuations caused by external power grid disturbances may be misjudged as sudden changes in raw material properties, leading to unnecessary downtime for inspection, resulting in decreased production efficiency and economic losses. Alternatively, genuine changes in process conditions caused by batch-to-batch variations in raw materials may be misjudged as normal fluctuations, missing genuine quality defect warnings and resulting in batches of substandard products. This overlap in signal characteristics severely reduces the reliability and practicality of production data analysis systems.

[0091] In some embodiments, this application proposes a production data analysis and processing system for an automated PVC pipe production line, comprising: a data acquisition unit for acquiring real-time operating current data of the extruder main motor, and acquiring grid voltage data and grid frequency data corresponding to the branch where the extruder main motor is located in the workshop shared power grid; an acquisition unit for, in response to instantaneous fluctuations in grid voltage data or grid frequency data, compensating the real-time operating current data according to a pre-established mapping relationship between grid disturbances and motor current response to obtain first current data, the first current data being used to characterize the extruder main motor current after excluding power quality disturbances corresponding to grid voltage data and grid frequency data; and a generation unit for, based on the extruder... The process parameter settings under the current production task determine the reference operating current mode, and generate a residual signal based on the difference between the first current data and the reference operating current mode. The residual signal is used to characterize the abnormal current fluctuation component after excluding the normal operating current component of the extruder itself. The processing unit is used to perform preset multi-scale decomposition processing on the residual signal and extract at least one microscopic feature parameter reflecting the trend of change of the physical properties of the raw material from the decomposed signal components. The output unit is used to determine the fluctuation source category of the real-time operating current data based on the matching result of the microscopic feature parameter and the preset raw material characteristic fingerprint, as well as the fluctuation record of the grid voltage data or grid frequency data, and output the warning information corresponding to the fluctuation source category.

[0092] The above are merely embodiments of this application and are not intended to limit the scope of protection of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of protection of this application.

Claims

1. A method for analyzing and processing production data from an automated PVC pipe production line, characterized in that, include: Collect real-time operating current data of the main motor of the extruder, as well as grid voltage and grid frequency data in the workshop shared power grid corresponding to the branch where the main motor of the extruder is located; In response to instantaneous fluctuations in the grid voltage data or the grid frequency data, the real-time operating current data is compensated according to a pre-established mapping relationship between grid disturbances and motor current response to obtain first current data. The first current data is used to characterize the extruder main motor current after excluding the power quality disturbances corresponding to the grid voltage data and the grid frequency data. Based on the process parameter settings under the current production task of the extruder, a reference operating current mode is determined, and a residual signal is generated based on the difference between the first current data and the reference operating current mode. The residual signal is used to characterize the abnormal current fluctuation component after excluding the normal operating current component of the extruder itself. The residual signal is subjected to a preset multi-scale decomposition process, and at least one microscopic feature parameter reflecting the trend of change in the physical properties of the raw material is extracted from the decomposed signal components. Based on the matching results of the microscopic feature parameters and the preset raw material characteristic fingerprint database, as well as the fluctuation records of the power grid voltage data or the power grid frequency data, the fluctuation source category of the real-time operating current data is determined, and the early warning information corresponding to the fluctuation source category is output. The step of performing a preset multi-scale decomposition process on the residual signal and extracting at least one microscopic feature parameter reflecting the changing trend of the physical properties of the raw material from the decomposed signal components includes: A preset multi-scale decomposition operation is performed on the residual signal to obtain multiple signal components with different frequency scales. The preset multi-scale decomposition operation is used to separate the residual signal according to frequency range. From the multiple signal components of different frequency scales, select the target signal component within the target frequency range that is associated with the plasticizing cycle of the raw material in the extruder. The target signal component is subjected to amplitude gain processing to obtain an enhanced signal component; The micro-feature parameters are extracted from the enhanced signal components. The micro-feature parameters include at least one of the following: root mean square value, kurtosis value, spectral energy ratio of a preset frequency range, and trend rate of change of the enhanced signal components. The preset raw material characteristic fingerprint database is established in the following way: While the extruder is running, record the historical residual signals when the known properties of the raw materials change; Perform the same decomposition operation as the preset multi-scale decomposition process on the historical residual signal, and extract the historical micro-feature parameter sets corresponding to each of the various raw material characteristic change patterns; The historical micro-feature parameter set is associated with the corresponding raw material characteristic change type and stored to form the preset raw material characteristic fingerprint database. The methods for obtaining the first current data include: The mapping relationship between the power grid disturbance and the motor current response is invoked, and the current compensation amount caused by the power grid disturbance is determined based on the changes in the power grid voltage data or the changes in the power grid frequency data monitored in real time. The mapping relationship between the power grid disturbance and the motor current response is formed by repeatedly triggering the start and stop operations of other power equipment in the workshop under no-load or stable operating conditions, and synchronously recording the corresponding data set of the power grid voltage change, power grid frequency change and extruder main motor current response change for each operation. The dynamic current response coefficient set is determined based on the corresponding data set. The first current data is obtained by subtracting the current compensation amount from the real-time operating current data.

2. The method according to claim 1, characterized in that, The acquisition of real-time operating current data of the extruder main motor, as well as the acquisition of grid voltage and frequency data in the workshop shared power grid corresponding to the branch where the extruder main motor is located, includes: A current sensor is deployed on the three-phase power line of the main motor of the extruder, and the real-time operating current data is acquired through the current sensor at a first sampling frequency. The first sampling frequency is set to a frequency range that can distinguish the details of the current fluctuation of the main motor of the extruder when the physical properties of the raw material change. Power quality monitoring devices are deployed at the main incoming switch of the shared power grid in the workshop and in the distribution cabinet of the branch where the main motor of the extruder is located. The power quality monitoring devices synchronously acquire the grid voltage data and the grid frequency data at a second sampling frequency. The second sampling frequency is set to a frequency range that can capture instantaneous power quality changes caused by the start-up and shutdown of other power equipment in the workshop.

3. The method according to claim 1, characterized in that, In response to instantaneous fluctuations in the grid voltage data or grid frequency data, the real-time operating current data is compensated based on a pre-established mapping relationship between grid disturbances and motor current response to obtain first current data, including: The grid voltage data is compared with a preset voltage fluctuation threshold, and the grid frequency data is compared with a preset frequency fluctuation threshold; When the change in the grid voltage data exceeds the preset voltage fluctuation threshold, or the change in the grid frequency data exceeds the preset frequency fluctuation threshold, the instantaneous fluctuation is determined to have occurred.

4. The method according to claim 1, characterized in that, The step of determining a reference operating current mode based on the process parameter settings under the current production task of the extruder, and generating a residual signal based on the difference between the first current data and the reference operating current mode, includes: The process parameter setting value is obtained as at least one of the following under the current production task of the extruder: the screw speed setting value, the temperature setting value of each heating zone of the barrel, and the melt pressure setting value. Based on the pre-established correspondence between process parameters and reference current, the expected current range or expected current curve corresponding to the process parameter setting value is determined as the reference operating current mode. Calculate the difference between the first current data and the reference operating current mode, and use the difference as the residual signal.

5. The method according to claim 1, characterized in that, The preset multi-scale decomposition operation is an operation that uses a preset mother wavelet function to perform multi-level decomposition on the residual signal.

6. The method according to claim 1, characterized in that, The step of determining the source category of fluctuations in the real-time operating current data based on the matching results of the microscopic feature parameters and the preset raw material characteristic fingerprint database, as well as the fluctuation records of the power grid voltage data or the power grid frequency data, includes: The microscopic feature parameters are matched with multiple raw material characteristic change patterns stored in a preset raw material characteristic fingerprint database based on similarity. When the similarity score between the micro-feature parameter and at least one raw material characteristic change pattern in the preset raw material characteristic fingerprint database exceeds a preset matching threshold, the fluctuation source category of the real-time operating current data is determined to be the internal difference of the raw material batch. When the microscopic characteristic parameters remain below a preset stable threshold, and the real-time operating current data exhibits fluctuations exceeding a preset fluctuation range before compensation processing, and the grid voltage data or the grid frequency data contains records of the instantaneous fluctuations, the fluctuation source category of the real-time operating current data is determined to be external grid disturbance.

7. The method according to any one of claims 1 to 6, characterized in that, The acquisition of real-time operating current data of the extruder main motor, as well as the acquisition of grid voltage and frequency data in the workshop shared power grid corresponding to the branch where the extruder main motor is located, includes: Real-time operating current data of the main motor of the extruder is collected at a first sampling frequency, and grid voltage data and grid frequency data corresponding to the branch where the main motor of the extruder is located in the workshop shared power grid are collected at a second sampling frequency; wherein, the first sampling frequency is higher than the second sampling frequency, the first sampling frequency is in the kilohertz range, and the second sampling frequency is in the kilohertz range.

8. A production data analysis and processing system for an automated PVC pipe production line, characterized in that, include: The acquisition unit is used to acquire real-time operating current data of the main motor of the extruder, as well as grid voltage data and grid frequency data in the workshop shared power grid corresponding to the branch where the main motor of the extruder is located; The acquisition unit is used to respond to instantaneous fluctuations in the grid voltage data or the grid frequency data, and to perform compensation processing on the real-time operating current data according to a pre-established mapping relationship between grid disturbances and motor current response to obtain first current data. The first current data is used to characterize the extruder main motor current after excluding the power quality disturbances corresponding to the grid voltage data and the grid frequency data. The generation unit is used to determine the reference operating current mode based on the process parameter setting value under the current production task of the extruder, and generate a residual signal based on the difference between the first current data and the reference operating current mode. The residual signal is used to characterize the abnormal current fluctuation component after excluding the normal operating current component of the extruder itself. The processing unit is used to perform preset multi-scale decomposition processing on the residual signal and extract at least one microscopic feature parameter reflecting the trend of change of the physical properties of the raw material from the decomposed signal components. The output unit is used to determine the fluctuation source category of the real-time operating current data based on the matching result of the micro-feature parameters and the preset raw material characteristic fingerprint database, as well as the fluctuation record of the power grid voltage data or the power grid frequency data, and output the early warning information corresponding to the fluctuation source category. The step of performing a preset multi-scale decomposition process on the residual signal and extracting at least one microscopic feature parameter reflecting the changing trend of the physical properties of the raw material from the decomposed signal components includes: A preset multi-scale decomposition operation is performed on the residual signal to obtain multiple signal components with different frequency scales. The preset multi-scale decomposition operation is used to separate the residual signal according to frequency range. From the multiple signal components of different frequency scales, select the target signal component within the target frequency range that is associated with the plasticizing cycle of the raw material in the extruder. The target signal component is subjected to amplitude gain processing to obtain an enhanced signal component; The micro-feature parameters are extracted from the enhanced signal components. The micro-feature parameters include at least one of the following: root mean square value, kurtosis value, spectral energy ratio of a preset frequency range, and trend rate of change of the enhanced signal components. The preset raw material characteristic fingerprint database is established in the following way: While the extruder is running, record the historical residual signals when the known properties of the raw materials change; Perform the same decomposition operation as the preset multi-scale decomposition process on the historical residual signal, and extract the historical micro-feature parameter sets corresponding to each of the various raw material characteristic change patterns; The historical micro-feature parameter set is associated with the corresponding raw material characteristic change type and stored to form the preset raw material characteristic fingerprint database. The methods for obtaining the first current data include: The mapping relationship between the power grid disturbance and the motor current response is invoked, and the current compensation amount caused by the power grid disturbance is determined based on the changes in the power grid voltage data or the changes in the power grid frequency data monitored in real time. The mapping relationship between the power grid disturbance and the motor current response is formed by repeatedly triggering the start and stop operations of other power equipment in the workshop under no-load or stable operating conditions, and synchronously recording the corresponding data set of the power grid voltage change, power grid frequency change and extruder main motor current response change for each operation. The dynamic current response coefficient set is determined based on the corresponding data set. The first current data is obtained by subtracting the current compensation amount from the real-time operating current data.

Images