Spinning production line operation monitoring method and device based on multi-time-series anomaly detection model

Abstract

The invention provides a spinning production line operation monitoring method and device based on a multi-time-series anomaly detection model. Various data of production equipment and auxiliary equipment in the spinning production process are collected in real time; the collected data is subjected to cleaning, denoising and standardized preprocessing; adopting the preprocessed data to construct a multi-time sequence and carrying out real-time monitoring; and carrying out anomaly detection and analysis by using a multi-time sequence anomaly detection model.

Description

Technical Field

[0001] The present invention relates to the field of textile production and detection technology, and in particular to a method and device for monitoring the operation of a spinning production line based on a multi-time series anomaly detection model. Background Art

[0002] As the global market's requirements for textile quality and production efficiency continue to increase, traditional production monitoring methods can no longer meet the needs of modern textile production. Therefore, developing an efficient technology that can monitor the operation status of the spinning production line in real time and detect and handle production anomalies in a timely manner is of great significance to improving production efficiency, reducing costs and enhancing international competitiveness.

[0003] As the core link of the textile industry, the production efficiency and product quality of spinning directly affect the utilization efficiency of upstream raw materials and the market competitiveness of downstream products. With the advancement of science and technology and the improvement of industrial automation level, intelligent monitoring of spinning production lines has become the key to improving production efficiency. Traditional monitoring systems mainly rely on regular inspections and manual monitoring, which are not only inefficient, but also difficult to achieve precise control of the entire production process. In addition, due to the complexity of the spinning process, such as the influence of various factors such as temperature and humidity, equipment speed, and raw material quality, real-time and comprehensive monitoring of the production line is particularly important. Summary of the invention

[0004] In view of the defects and shortcomings of the prior art, the present invention considers providing an intelligent monitoring system based on a multi-time series anomaly detection model, which can comprehensively monitor and manage the spinning production line and detect production anomalies in real time, which is of great practical significance for ensuring product quality and improving production efficiency.

[0005] The present invention proposes a spinning production line operation monitoring method and device based on a multi-time series anomaly detection model, which adopts autoencoder and long short-term memory (LSTM) network technology to improve the monitoring intelligence and automation level of the spinning production line. Through these technologies, the present invention can perform deep learning and pattern recognition on the data in the production process, identify potential anomalies in real time and accurately, and ensure the continuity and stability of the production process.

[0006] From the perspective of manufacturing companies, the implementation of this technology will significantly improve production efficiency and reduce operating costs. By reducing equipment downtime and improving the efficiency of raw material use, companies can achieve higher production capacity and lower production waste. In addition, improved production quality and stability will also enhance the competitiveness of corporate products in the market and customer satisfaction.

[0007] From an industrial perspective, the application of the present invention helps promote technological progress and industrial automation and intelligence in the textile industry, and has the potential to set new production standards for the entire industry, thereby improving the overall technical level and production efficiency of the industry.

[0008] From a macroeconomic perspective, the development and application of this technology will help improve the level of intelligence and international competitiveness of the manufacturing industry. In addition, by optimizing resource use and reducing energy consumption and material waste in the production process, the present invention also meets the requirements of sustainable development and helps promote national policy goals of environmental protection and efficient resource utilization. Overall, the present invention not only improves the technical standards of the spinning industry, but also provides a reference and demonstration for the development of the entire manufacturing industry.

[0009] The present invention specifically adopts the following technical solutions:

[0010] A method for monitoring the operation of a spinning production line based on a multi-time series anomaly detection model collects various data including production equipment and auxiliary equipment in the spinning production process in real time; performs cleaning, denoising and standardization preprocessing on the collected data; uses the preprocessed data to construct a multi-time series and performs real-time monitoring; and uses the multi-time series anomaly detection model to perform anomaly detection and analysis.

[0011] Furthermore, the production line monitoring platform is used to realize real-time data and signal collection through sensors and signal collection devices, including: collecting the running speed, power consumption, vibration signal, feed amount, and impurity removal rate of the cotton opener; collecting the feed speed, cleanliness, energy consumption data, impurity content, and mechanical load of the cotton cleaner; collecting the fiber arrangement, combing speed, failure rate, combing quality, and fiber length distribution of the carding machine; collecting the merging accuracy, yarn uniformity, operating parameters, output speed, and yarn tension of the drawing frame; collecting the spinning speed, roving quality, and tension of the roving frame. , roving diameter, yarn breakage rate; for the spinning machine, the data collected include yarn fineness, spinning speed, yarn breakage rate, yarn tension, and yarn quality; for the automatic winder, the data collected include winding speed, package yarn quality, tension change, breakage rate, and winding efficiency; for auxiliary equipment, for the temperature control equipment, the data collected include production environment temperature, equipment internal temperature, temperature change rate, and temperature control equipment power consumption; for the humidity control equipment, the data collected include workshop humidity, yarn moisture content, humidity change rate, and humidity control equipment power consumption; for the dust removal equipment, the data collected include work efficiency, dust concentration, filtration effect, equipment wind speed, and dust removal efficiency.

[0012] Furthermore, during the preprocessing process: for data missing due to sensor failure or network connection problems, linear interpolation or spline interpolation methods are used to fill in the missing data; since the timestamps of data collected by different devices may be inconsistent, resulting in data asynchrony, interpolation or resampling methods are used to align the data to a unified time axis; in order to ensure the clock synchronization of all devices, the network time protocol is used for calibration to avoid data errors caused by clock asynchrony.

[0013] Furthermore, the construction of multiple time series and real-time monitoring is specifically to display the system status and setting parameters in real time through an operating platform with a graphical user interface, so as to conduct detailed monitoring of the overall operating status of the production line.

[0014] Furthermore, the anomaly detection and analysis using the multi-time series anomaly detection model specifically uses an autoencoder, reconstructs new time series data using a trained autoencoder model, and calculates the error between the input and the reconstructed output to detect anomalies;

[0015] The specific steps are as follows: First, determine the architecture of the autoencoder: select the number of neurons in the input layer, the hidden layer including the encoding layer and the decoding layer, and the output layer according to the number of production equipment and the amount of data; the number of neurons in the input layer is the same as the number of features of the time series, corresponding to the number of data points in each time step; the number of neurons in the encoding layer decreases layer by layer to compress the dimension of the input data; the number of neurons in the decoding layer increases layer by layer to restore the dimension of the input layer, corresponding to the hidden layer;

[0016] Afterwards, by calculating the error between the input and the reconstructed output, when the error exceeds the preset threshold, it is judged as an abnormal situation, thereby realizing real-time monitoring and intelligent decision-making of the spinning production line; the specific calculation process is as follows: the encoder maps the input data X to the latent space Z, and the decoder reconstructs the representation Z of the latent space back to the original data x′, where the encoder is:

[0017] Z=h(X)

[0018] The decoder is:

[0019] X'=h'(Z)

[0020] Where h and h′ are functions of the encoder and decoder respectively; the mean square error is used as the measure of the reconstruction error:

[0021]

[0022] Among them, x i and x′ iThey are the data of the i-th timestamp and its corresponding reconstructed version; the encoder and decoder of LSTM are:

[0023] p t =LSTM(X t ,p t-1 )

[0024] X′ t =LSTM(p t ,p t-1 )

[0025] p t and X t are the output and input information / state of time step t respectively; the reconstruction error of each time step is calculated to detect anomalies; a threshold ∈ is set, and if the reconstruction error exceeds the threshold, it is considered that there is an anomaly in the time step;

[0026]

[0027] And, a spinning production line operation monitoring device based on a multi-time series anomaly detection model, comprising:

[0028] Production line monitoring platform, used to collect various data of production equipment and auxiliary equipment in the spinning production process in real time;

[0029] Data preprocessing module, used for cleaning, denoising and standardizing the collected data;

[0030] Multi-time series monitoring module, used to construct multi-time series using pre-processed data and perform real-time monitoring;

[0031] Analysis module for anomaly detection and analysis using multiple time series anomaly detection models.

[0032] Furthermore, it also includes an alarm module, which is used to send out an alarm signal according to the trigger or the detection result of the analysis module.

[0033] Furthermore, a control system connected to the spinning production line is used to perform production control and adjustment according to the analysis results of the analysis module.

[0034] An electronic device comprises a memory, a processor and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the steps of a spinning production line operation monitoring method based on a multi-time series anomaly detection model as described above are implemented.

[0035] A non-transitory computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of a spinning production line operation monitoring method based on a multi-time series anomaly detection model as described above.

[0036] Current monitoring systems mainly rely on traditional rule-based algorithms or simple statistical methods to identify anomalies, which often lead to high false positive and false negative rates due to the lack of in-depth understanding of complex data patterns. Compared with the prior art, the present invention and its preferred solution can perform deep learning and pattern recognition on the data in the production process by introducing autoencoders and long short-term memory (LSTM) network technology, and accurately predict and identify potential anomalies in real time, so as to provide early warning and automatic adjustment before problems occur, and ensure the continuity and stability of the production process.

[0037] First, the present invention effectively solves the problem of rare or difficult-to-mark abnormal samples in actual production environments through the unsupervised learning mode of the autoencoder. This method can automatically learn the normal behavior patterns of production data, and then accurately identify abnormal behaviors that deviate from these patterns, significantly improving the reliability and practicality of abnormality detection. Secondly, the present invention not only monitors the operating status of the spinning machinery itself, but also comprehensively considers environmental factors closely related to the production process, such as data of temperature control equipment, humidity control equipment, and dust removal equipment. This comprehensive data integration and analysis ensures that the impact of environmental factors on product quality is effectively controlled, further enhancing the stability of the production line and the consistency of products. At the same time, when the system detects potential risks, it can automatically adjust production or start emergency response procedures, reduce manual intervention, and improve response speed and processing efficiency. The application of these technologies not only improves the accuracy and predictive ability of abnormality detection, but also ensures the optimization of the production environment, significantly improving the efficiency and product quality of the entire production line. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] The present invention is further described in detail below with reference to the accompanying drawings and specific embodiments:

[0039] Figure 1 The overall framework and flow chart of the embodiment of the present invention;

[0040] Figure 2 The present invention is a flowchart of a spinning production line operation monitoring system based on a multi-time series anomaly detection model. DETAILED DESCRIPTION

[0041] In order to make the features and advantages of the present invention more clearly understood, the following embodiments are specifically described in detail with reference to the accompanying drawings:

[0042] It should be noted that the following detailed descriptions are exemplary and are intended to provide further explanation of the present application. Unless otherwise specified, all technical and scientific terms used herein have the same meanings as those commonly understood by those skilled in the art to which the present application belongs.

[0043] It should be noted that the terms used herein are only for describing specific embodiments and are not intended to limit the exemplary embodiments according to the present application. As used herein, unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. In addition, it should be understood that when the terms "comprise" and / or "include" are used in this specification, it indicates the presence of features, steps, operations, devices, components and / or combinations thereof.

[0044] The embodiment of the present invention provides a method and device for monitoring the operation of a spinning production line based on a multi-time series anomaly detection model, which is used to improve production efficiency, ensure product quality, reduce equipment failure rate, optimize resource allocation, reduce production costs, and enhance production safety. Figure 1 As shown, its main process includes the following steps:

[0045] Relying on the production line monitoring platform to collect data and signals in real time, various data in the spinning production process are collected in real time;

[0046] Use the data preprocessing module to preprocess the data, clean, denoise and standardize the collected data;

[0047] Through the multi-time series monitoring module, pre-processed data is used to construct multiple time series and conduct real-time monitoring;

[0048] Through the intelligent analysis and decision-making module, anomaly detection and analysis are performed using a multi-time series anomaly detection model;

[0049] Through the alarm module, an alarm is issued in time when an abnormality is detected;

[0050] Through the control system module, the production line is automatically adjusted and controlled according to the analysis results;

[0051] The present invention monitors and analyzes various parameters in the production process through a multi-time series anomaly detection model, meets the needs of efficient, stable and intelligent production, avoids production interruptions, product quality degradation and resource waste caused by equipment failure or production anomalies, and improves the overall operating efficiency and reliability of the spinning production line.

[0052] As a preferred solution of the embodiment of the present invention, the production line monitoring platform is used to realize real-time data and signal collection, involving a variety of production equipment and auxiliary equipment to ensure comprehensive and accurate acquisition of various key parameters of the spinning production line operation. The collected data are divided into two categories according to the equipment type: production equipment and auxiliary equipment. Specifically, the cotton opener collects the running speed, power consumption, vibration signal, feed amount, impurity removal rate, etc.; the cotton cleaner collects the feed speed, cleanliness, energy consumption data, impurity content, mechanical load, etc.; the carding machine collects the fiber arrangement, combing speed, failure rate, combing quality, fiber length distribution, etc.; the drawing machine collects the merging accuracy, strip uniformity, operating parameters, output speed, yarn tension, etc.; the roving machine collects the spinning speed, roving quality, tension, roving diameter, yarn breakage rate, etc.; the spinning machine collects the yarn fineness, spinning speed, yarn breakage rate, yarn tension, yarn quality, etc.; the automatic winder collects the winding speed, package yarn quality, tension change, broken rate, winding efficiency, etc. In terms of auxiliary equipment, the temperature control equipment collects the production environment temperature, internal equipment temperature, temperature change rate, power consumption of temperature control equipment, etc.; the humidity control equipment collects the workshop humidity, yarn moisture content, humidity change rate, power consumption of humidity control equipment, etc.; the dust removal equipment collects the work efficiency, dust concentration, filtration effect, equipment wind speed, dust removal efficiency, etc. The operating data of all these equipment are transmitted to the monitoring system in real time through sensors and signal acquisition devices, providing basic data support for subsequent data preprocessing, real-time monitoring, intelligent analysis and decision-making.

[0053] As a preferred solution of the embodiment of the present invention, the data preprocessing module is responsible for cleaning, denoising and standardizing the collected raw data to ensure the accuracy and reliability of subsequent analysis. First, for data missing due to sensor failure, network connection problems, etc., linear interpolation, spline interpolation and other methods are used to fill the missing data, thereby ensuring the integrity of the data. Secondly, since the timestamps of data collected by different devices may be inconsistent, resulting in data asynchrony, interpolation or resampling methods are used to align the data to a unified time axis. This step ensures the consistency of all data in time. In addition, in order to ensure the clock synchronization of all devices, the network time protocol (NTP) is used for calibration to avoid data errors caused by clock asynchrony. The data preprocessing module also needs to process the noise that may be contained in the sensor data, because the noise will affect the subsequent processing results. To this end, low-pass filtering, high-pass filtering and other methods are used to remove noise and improve the purity and reliability of the data. The data is standardized to make data from different sources comparable, ensuring that accurate results can be obtained in subsequent analysis. All these steps work together to ensure the integrity, consistency and accuracy of the data, and provide reliable basic data support for real-time monitoring, intelligent analysis and decision-making.

[0054] like Figure 2As shown, this embodiment constructs a multi-time series as the input of the real-time monitoring module by preprocessing the collected raw data. As a preferred solution of the embodiment of the present invention, the multi-time series monitoring module accepts the preprocessed multi-time series data, provides an operating platform with a graphical user interface (GUI), and allows operators to view the system status, set parameters and operate the production line in real time. The module contains advanced control logic, which can not only realize the visual display of real-time data, but also perform detailed monitoring of the overall operation status of the production line. The operator can view the operation status, production parameters and historical data of each device through the GUI interface, and adjust the production parameters or take corrective measures as needed. The visualization function of the overall operation status of the production line includes displaying the real-time operation data, production indicators, fault information and trend analysis charts of each device, which helps the operator to fully understand the operation of the production line. The real-time monitoring module not only improves the transparency and controllability of the production line, but also provides operators with powerful tools to optimize the production process, improve production efficiency and ensure product quality.

[0055] As a preferred solution of the embodiment of the present invention, the intelligent analysis and decision module uses a multi-time series anomaly detection model to perform anomaly detection and analysis. In this embodiment, an unsupervised learning neural network, Autoencoder, is used. The basic idea is to reconstruct new time series data using a trained autoencoder model and calculate the error between the input and the reconstructed output to detect anomalies. The specific steps are as follows:

[0056] First, determining the architecture of the autoencoder is one of the key steps in multi-time series anomaly detection. According to the number of production equipment and the amount of data, select the appropriate number of neurons in the input layer, hidden layer (encoding layer and decoding layer), and output layer. The number of neurons in the input layer is usually the same as the number of features of the time series (that is, the number of data points at each time step). The number of neurons in the encoding layer is generally reduced layer by layer to compress the dimension of the input data. The first encoding layer is usually set to half or a quarter of the number of neurons in the input layer, the second encoding layer is set to half or less of the number of neurons in the first encoding layer, and the third encoding layer (if further compression is required) is set to half of the number of neurons in the second encoding layer. The number of neurons in the decoding layer is usually increased layer by layer to restore the dimension of the input layer, corresponding to the hidden layer. The number of neurons in the first decoding layer is the same as the last encoding layer, the second decoding layer is the same as the second encoding layer, and the third decoding layer is the same as the first encoding layer. The number of neurons in the output layer is the same as the input layer to ensure that the reconstructed output is consistent with the dimension of the input data. Through the above architecture design, the intelligent analysis and decision module can efficiently process and analyze multi-time series data. The architecture can be adjusted according to specific application scenarios and data characteristics to optimize the performance of the autoencoder and the anomaly detection effect. Ultimately, by calculating the error between the input and the reconstructed output, when the error exceeds the preset threshold, it can be judged as an abnormal situation, thereby achieving real-time monitoring and intelligent decision-making of the spinning production line, improving production efficiency and product quality, and reducing equipment failure rate.

[0057] The specific calculation process is as follows: the encoder maps the input data X to the latent space Z, and the decoder reconstructs the representation Z of the latent space back to the original data X′, where the encoder is:

[0058] Z=h(X)

[0059] The decoder is:

[0060] X'=h'(Z)

[0061] Where h and hv are functions of the encoder and decoder respectively. This embodiment uses mean square error (MSE) as a measure of reconstruction error:

[0062]

[0063] Here x i and x′ i They are the data of the i-th timestamp and its corresponding reconstructed version. When processing time series data, the long short-term memory network (LSTM) autoencoder is often used to capture time dependencies. The encoder and decoder of LSTM are:

[0064] p t =LSTM(Xt ,p t-1 )

[0065] X' t =LSTM(p t ,p t-1 )

[0066] p t and X t are the output and input information / state at time step t, respectively. By calculating the reconstruction error at each time step, anomalies can be detected:

[0067]

[0068] A threshold ∈ is set. If the reconstruction error exceeds the threshold, it is considered that there is an anomaly at this time step.

[0069] As a preferred solution of the embodiment of the present invention, the alarm module is used to detect and send out alarm signals. The alarm module is divided into two categories, one is manually triggered, including but not limited to employees discovering abnormal situations, such as machine equipment failures caused by human or external factors, unexpected shutdowns of production lines, raw material problems, etc., and the other is the system automatically triggering the alarm. When the intelligent analysis and decision-making module detects an abnormal situation through data anomaly detection, machine learning algorithm analysis, prediction model comparison and other means, a signal is sent to the alarm module. The alarm is sounded and information is sent to the mobile terminal management system, such as a mobile phone application or a computer terminal. The alarm module will record all alarm events and their related data for subsequent analysis and optimization. These data can help enterprises identify common problems and improve production processes. It automatically interacts with the system control module and automatically executes predetermined emergency response measures, such as shutdown, switching to backup equipment, etc., to reduce losses and downtime.

[0070] As a preferred solution of the embodiment of the present invention, the control system module is used to coordinate and control various equipment and systems of the spinning production line. The control system module manages production equipment: cotton openers, cotton cleaners, carding machines, drawing machines, roving machines, spinning machines, automatic winding machines; auxiliary equipment: temperature control equipment, humidity control equipment, dust removal equipment, etc. It interacts with the intelligent alarm module automatically, and when an abnormality occurs, it immediately responds and executes predetermined emergency measures, such as shutting down, adjusting operating parameters, switching to backup equipment, etc. It interacts and integrates data with MES, ERP and other systems to achieve real-time interaction and management of production systems; it supports remote monitoring and operation of production lines, and adjusts production parameters and operating equipment through remote control devices to ensure stable and efficient operation of the production process.

[0071] As a specific example, the production steps of a spinning production line, the corresponding equipment and the collected data are as follows.

[0072]

[0073]

[0074]

[0075]

[0076] It should be pointed out that in addition to the data of equipment related to the production process, the present invention also collects data of auxiliary equipment, such as: humidity control equipment (humidity data: humidity levels in various areas of the workshop, equipment status: operating status, fault alarm, energy consumption data: power consumption); temperature control equipment (temperature data: temperature levels in various areas of the workshop, equipment status: operating status, fault alarm, energy consumption data: power consumption); dust removal equipment (air quality: dust concentration; equipment status: operating status, fault alarm; energy consumption data: power consumption).

[0077] The present invention adopts multi-time series data processing technology, which can synchronously analyze the time series data of multiple production line variables (such as machine operating parameters, environmental factors, etc.) related to the spinning process. The correlation between time series data is used to more accurately identify potential anomalies, and a new anomaly detection mechanism for a spinning production line is provided. In the field of spinning, the autoencoder in the present invention is used to effectively perform feature extraction and data dimension reduction, while the LSTM network is used to process and predict long-term dependencies in time series data. This combination not only optimizes the accuracy of anomaly detection, but also enhances the system's understanding depth of time series data, enabling it to more effectively process complex production line data. An automatic response mechanism is implemented when abnormal behavior is detected through an intelligent decision support system, such as automatically adjusting production parameters or starting emergency procedures. This automated response reduces the need for manual intervention and improves production efficiency and safety. The present invention not only focuses on the operating data of the machine equipment itself, but also includes real-time analysis of environmental monitoring equipment data such as temperature control, humidity control and dust removal equipment. This all-round monitoring mode is a major innovation of the present invention, which optimizes product quality and production process by comprehensively analyzing production environment factors. The combined effect of these key technical points provides an advanced and efficient operation monitoring solution for spinning production lines. The realization of these technical points ensures the significant advantages of the present invention in improving production efficiency, reducing costs, and improving product quality, while also reflecting its technological innovation in the intelligent upgrading of the textile industry.

[0078] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that include computer-usable program code.

[0079] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0080] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0081] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0082] The present invention is not limited to the above-mentioned optimal implementation mode. Anyone can derive other various forms of a spinning production line operation monitoring method and device based on a multi-time series anomaly detection model under the inspiration of the present invention. All equal changes and modifications made within the scope of the patent application of the present invention should fall within the scope of the present invention.

Claims

1. A method for monitoring the operation of a spinning production line based on a multi-time series anomaly detection model, characterized in that: Collect various data of production equipment and auxiliary equipment in the spinning production process in real time; perform cleaning, denoising and standardization preprocessing on the collected data; use the preprocessed data to construct multiple time series and conduct real-time monitoring; And use the multi-time series anomaly detection model for anomaly detection and analysis.

2. The method for monitoring the operation of a spinning production line based on a multi-time series anomaly detection model according to claim 1, characterized in that: The production line monitoring platform is used to collect data and signals in real time through sensors and signal acquisition devices, including: collecting running speed, power consumption, vibration signal, feed amount, and impurity removal rate of the cotton opener; collecting feed speed, cleanliness, energy consumption data, impurity content, and mechanical load of the cotton cleaner; collecting fiber arrangement, combing speed, failure rate, combing quality, and fiber length distribution of the carding machine; collecting merging accuracy, strip uniformity, operating parameters, output speed, and yarn tension of the drawing frame; collecting spinning speed, roving quality, tension, and coarseness of the roving frame. Yarn diameter, yarn breakage rate; for spinning machines, data on yarn fineness, spinning speed, yarn breakage rate, yarn tension, and yarn quality; for automatic winders, data on winding speed, package yarn quality, tension change, breakage rate, and winding efficiency; for auxiliary equipment, data on temperature control equipment include production environment temperature, equipment internal temperature, temperature change rate, and power consumption of temperature control equipment; data on humidity control equipment include workshop humidity, yarn moisture content, humidity change rate, and power consumption of humidity control equipment; for dust removal equipment, data on working efficiency, dust concentration, filtration effect, equipment wind speed, and dust removal efficiency.

3. The method for monitoring the operation of a spinning production line based on a multi-time series anomaly detection model according to claim 1, characterized in that: During the preprocessing process: for data missing due to sensor failure or network connection problems, linear interpolation and spline interpolation methods are used to fill the missing data; since the timestamps of data collected by different devices may be inconsistent, resulting in data asynchrony, interpolation or resampling methods are used to align the data to a unified time axis; in order to ensure the clock synchronization of all devices, the network time protocol is used for calibration to avoid data errors caused by clock asynchrony.

4. The method for monitoring the operation of a spinning production line based on a multi-time series anomaly detection model according to claim 1, characterized in that: The construction of multiple time series and real-time monitoring is specifically to display the system status and setting parameters in real time through an operating platform with a graphical user interface, so as to conduct detailed monitoring of the overall operating status of the production line.

5. The method for monitoring the operation of a spinning production line based on a multi-time series anomaly detection model according to claim 1, characterized in that: The anomaly detection and analysis using the multi-time series anomaly detection model specifically uses an autoencoder, which reconstructs new time series data using a trained autoencoder model and calculates the error between the input and the reconstructed output to detect anomalies; The specific steps are as follows: First, determine the architecture of the autoencoder: select the number of neurons in the input layer, the hidden layer including the encoding layer and the decoding layer, and the output layer according to the number of production equipment and the amount of data; the number of neurons in the input layer is the same as the number of features of the time series, corresponding to the number of data points in each time step; the number of neurons in the encoding layer decreases layer by layer to compress the dimension of the input data; the number of neurons in the decoding layer increases layer by layer to restore the dimension of the input layer, corresponding to the hidden layer; Afterwards, by calculating the error between the input and the reconstructed output, when the error exceeds the preset threshold, it is judged as an abnormal situation, thereby realizing real-time monitoring and intelligent decision-making of the spinning production line; the specific calculation process is as follows: the encoder maps the input data X to the latent space Z, and the decoder reconstructs the representation Z of the latent space back to the original data X′, where the encoder is: Z=h(X) The decoder is: X'=h'(Z) Where h and h′ are functions of the encoder and decoder respectively; the mean square error is used as the measure of the reconstruction error: Among them, x i and x′ i They are the data of the i-th timestamp and its corresponding reconstructed version; the encoder and decoder of LSTM are: p t =LSTM(X t ,p t-1 ) X' t =LSTM(p t ,p t-1 ) p t and X t are the output and input information / state of time step t respectively; the reconstruction error of each time step is calculated to detect anomalies; a threshold ∈ is set, and if the reconstruction error exceeds the threshold, it is considered that there is an anomaly in the time step; 6. A spinning production line operation monitoring device based on a multi-time series anomaly detection model, characterized in that: include: Production line monitoring platform, used to collect various data of production equipment and auxiliary equipment in the spinning production process in real time; Data preprocessing module, used for cleaning, denoising and standardizing the collected data; Multi-time series monitoring module, used to construct multi-time series using pre-processed data and perform real-time monitoring; Analysis module for anomaly detection and analysis using multiple time series anomaly detection models.

7. The spinning production line operation monitoring device based on a multi-time series anomaly detection model according to claim 6 is characterized in that: It also includes an alarm module, which is used to send out an alarm signal according to the trigger or the detection result of the analysis module.

8. The spinning production line operation monitoring device based on a multi-time series anomaly detection model according to claim 6 is characterized in that: A control system connected to the spinning production line is used to perform production control and adjustment according to the analysis results of the analysis module.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, the steps of a spinning production line operation monitoring method based on a multi-time series anomaly detection model as described in any one of claims 1 to 5 are implemented.

10. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of a method for monitoring the operation of a spinning production line based on a multi-time series anomaly detection model as described in any one of claims 1 to 5 are implemented.

Images