Paper machine net part load abnormity monitoring system

Abstract

The invention provides a paper machine net part load abnormity monitoring system, and relates to the technical field of papermaking equipment monitoring, and the system comprises a data collection module which is used for collecting operation data of a paper machine net part, the operation data comprises net part driving motor current data, net part tension data, net part vibration data, net part temperature data, slurry flow data and slurry concentration data, and the data acquisition module comprises a current sensor, a tension sensor, a vibration sensor, a temperature sensor, a flow sensor and a concentration sensor; through multi-dimensional key parameter acquisition and coverage of operation of a net part driving motor, an equipment state and material characteristics, the limitation that the load is indirectly judged only depending on motor current traditionally is broken through, abnormal missed judgment caused by neglecting of key factors such as tension and slurry concentration is avoided, and comprehensive monitoring of the net part load state is realized.

Description

Technical Field

[0001] This invention relates to the field of papermaking equipment monitoring technology, and in particular to a paper machine wire section load abnormality monitoring system. Background Technology

[0002] The wire section of a paper machine is a core component, responsible for pulp dewatering and wet paper web formation. Its load status directly affects paper quality, production efficiency, and equipment lifespan. Existing wire section load monitoring systems have the following problems: The monitoring dimension is too narrow: the load is judged indirectly by the current of the drive motor in the mesh section, ignoring key influencing factors such as tension, vibration and slurry concentration. For example, an increase in slurry concentration will lead to an increase in dewatering load, but the motor current may not change significantly, making it easy to miss abnormalities. Weak data processing capabilities: Collected data is simply displayed / stored without cleaning or standardization. For example, the data dimensions of different sensors vary greatly, and direct analysis will lead to errors. Furthermore, there is a lack of in-depth analysis models, making it impossible to accurately calculate real-time load. Untimely early warning: It relies solely on fixed thresholds for judgment without considering load change trends. For example, when the load slowly exceeds the threshold, no early warning is given, and the early warning method is limited to the site, making it impossible for managers to know remotely and in a timely manner. Poor data storage reliability: It uses a single database for storage, which makes it easy to lose data in case of failure. In addition, there is no index management, and the efficiency of querying historical data is low. Summary of the Invention

[0003] In view of the above-mentioned shortcomings of the prior art, the present invention provides a paper machine wire section load abnormality monitoring system.

[0004] This invention provides a paper machine wire section load anomaly monitoring system, specifically including: a data acquisition module, a data preprocessing module, a load analysis module, an anomaly judgment module, an early warning module, a data storage module, and a remote monitoring module; The data acquisition module is used to collect the operating data of the wire section of the paper machine. The operating data includes the current data of the wire section drive motor, the wire section tension data, the wire section vibration data, the wire section temperature data, the pulp flow rate data, and the pulp concentration data. The data acquisition module includes a current sensor, a tension sensor, a vibration sensor, a temperature sensor, a flow sensor, and a concentration sensor. The data preprocessing module is connected to the data acquisition module and is used to preprocess the running data acquired by the data acquisition module. The preprocessing includes data cleaning, data standardization, and data noise reduction. The load analysis module is connected to the data preprocessing module and is used to acquire preprocessed operating data. Combined with preset historical load data of the paper machine wire section, a real-time load analysis model of the paper machine wire section is constructed to calculate the real-time load value of the paper machine wire section. The anomaly judgment module is connected to the load analysis module and is used to set the normal range threshold of the paper machine wire section load, including static threshold and dynamic threshold. The real-time load value calculated by the load analysis module is compared with the normal range threshold. At the same time, combined with the changing trend of the real-time load analysis model, it is used to judge whether the load of the paper machine wire section is abnormal. The early warning module is connected to the anomaly judgment module and is used to issue corresponding early warning signals according to the anomaly level, such as mild / moderate / severe, when the anomaly judgment module determines that the paper machine wire section load is abnormal. The early warning signals include audible and visual warnings, SMS warnings, and system pop-up warnings. The data storage module is connected to the data acquisition module, data preprocessing module, load analysis module and anomaly judgment module respectively, and is used to store the acquired raw operating data, preprocessed operating data, real-time load value, historical load data and anomaly judgment results; The remote monitoring module is connected to the load analysis module, the anomaly judgment module, and the data storage module, respectively, and is used to enable remote users to view the real-time load value, operating data, and anomaly judgment results of the paper machine wire section. It also supports remote users to adjust the warning threshold.

[0005] Optionally, the data acquisition module includes a current sensor, a tension sensor, a vibration sensor, a temperature sensor, a flow sensor, and a concentration sensor; The current sensor is installed on the power supply line of the drive motor in the mesh section and is used to collect the current data of the drive motor. The tension sensor is installed at the tension detection point of the mesh section, such as the tension roller at the mesh inlet and the tension roller at the vacuum roll outlet, to collect tension data of the mesh section; The vibration sensors are installed in key parts of the mesh frame, such as the mesh crossbeam, vacuum box support, and drying cylinder bearing seat, to collect vibration data of the mesh. The temperature sensors are installed on the mesh bearing, the surface of the drying cylinder, and the outer wall of the slurry flow channel to collect temperature data at the corresponding locations. The flow sensor is installed on the slurry conveying pipeline to collect slurry flow data; The concentration sensors are installed at the outlet of the slurry mixing tank and the slurry inlet of the mesh section to collect slurry concentration data.

[0006] Optionally, the data cleaning in the data preprocessing module includes removing missing values, outliers, and duplicate values; For missing values: if the missing proportion is less than the preset proportion threshold, use linear interpolation or mean imputation to fill the missing values; if the missing proportion is greater than or equal to the preset proportion threshold, discard the data and record the discarding situation. For outliers: use Grubbs' test or box plot method to identify them, and use nearest neighbor replacement method or data smoothing method to handle them according to the degree of deviation; For duplicate values: delete the duplicate data entries directly, keeping only 1 valid data.

[0007] Optionally, the data standardization in the data preprocessing module adopts the Z-score standardization method, with the following formula: Z represents the standardized data, X represents the original data, μ represents the mean of the original data, and σ represents the standard deviation of the original data.

[0008] Optionally, the data denoising in the data preprocessing module employs wavelet transform denoising, with the following specific steps: (S11) Perform wavelet decomposition on the running data, select the appropriate wavelet basis function and the number of decomposition levels, and obtain multi-scale wavelet coefficients; (S12) Process the wavelet coefficients according to the preset threshold rule to suppress the coefficients corresponding to noise; (S13) Reconstruct the processed wavelet coefficients to obtain the denoised data.

[0009] Optionally, the load analysis module uses a BP neural network model to construct a real-time load analysis model; Input layer: 6 neurons, corresponding to the preprocessed data of "current, tension, vibration, temperature, flow rate, and concentration"; Hidden layers: at least 2 layers, with the number of neurons in each layer set to 12-16 depending on the input dimension and accuracy requirements; Output layer: 1 neuron, corresponding to the real-time load value of the paper machine wire section; Model training: Gradient descent is used, with historical load data and corresponding operational data as samples, to adjust network weights and thresholds so that the prediction error is less than or equal to the preset error threshold.

[0010] Optionally, the load normal range threshold of the anomaly judgment module 4 includes: Static threshold: determined based on paper machine wire section design parameters, rated load, and historical normal operation data, such as ±10% of rated load; Dynamic threshold: Based on real-time operating conditions, the production speed, slurry type, and product specifications are automatically adjusted. The adjustment algorithm uses an operating condition parameter-threshold mapping table and pre-stores multiple threshold ranges corresponding to operating conditions. Anomaly detection logic: When the real-time load value exceeds the normal range threshold and the real-time load analysis model shows that the load is "continuously deviating from the normal trend", it is determined to be an abnormal load.

[0011] Optionally, the correspondence between the anomaly level and the warning method of the early warning module is as follows: Mild anomaly: Real-time load value exceeds the threshold by less than 5%, and the load deviation trend is slow → only audible and visual warning; Moderate anomaly: 5% ≤ real-time load value exceeds the threshold by less than 10%, and the load deviation trend is moderate → audible and visual alarm + SMS alarm, sending an alarm SMS to the mobile phone of the on-duty manager, containing the abnormal parameters; Severe anomaly: Real-time load value exceeds the threshold by ≥10%, and the load deviation trend is rapid → audible and visual warning + SMS warning + system pop-up warning, remote monitoring module pop-up prompt, and push anomaly handling suggestions.

[0012] Optionally, the data storage module adopts a distributed database, consisting of a master database and 3-5 slave databases; Master database: Receives and stores all data, and synchronizes data to slave databases; Secondary database: Backs up the primary database data and automatically switches to a new primary database if the primary database fails; Data management: Indexes are created based on data type (e.g., raw / preprocessed / load value / abnormal results), acquisition time (divided by hour), and data source, supporting multi-dimensional queries, such as tension sensor tension data for the past 24 hours.

[0013] Optionally, the remote monitoring module includes a web client and a mobile client; Web client: Accessed via browser, features include "real-time data viewing, historical data query, export of data from the past year to Excel, and generation of daily / weekly / monthly load statistics reports"; Mobile client: Mobile APP, supporting Android / iOS, with functions including "receiving warning information, pushing notifications to desktop, real-time data preview, and viewing of abnormal records", and automatically popping up a warning prompt box when there are serious abnormalities; User access control is divided into three levels: "administrator full-function operation, technician data viewing + threshold adjustment, and operator real-time data viewing". Account passwords are stored in encrypted form.

[0014] The beneficial effects are as follows: By collecting key parameters from multiple dimensions, the system covers the tension, vibration, temperature, and material characteristics of the drive motor and equipment status, such as slurry flow rate and concentration. This breaks through the limitations of traditional methods that rely solely on motor current to indirectly determine the load, and avoids omissions due to neglecting key factors such as tension and slurry concentration. This enables comprehensive monitoring of the load status of the network.

[0015] A full-process data processing mechanism is adopted to clean the collected data, remove missing, abnormal, and duplicate values, standardize the data, and suppress interference signals. Combined with intelligent models, such as BP neural network deep analysis, it replaces the traditional simple data storage / display method, which greatly improves the accuracy of real-time load calculation and provides reliable data support for accurate judgment of load status.

[0016] Abandoning the traditional fixed threshold judgment mode, a dynamic threshold is adopted to adapt to real-time operating conditions and judge anomalies in combination with load change trends. At the same time, a graded early warning mode is designed, including on-site sound and light, remote text message, and system pop-up window. This overcomes the shortcomings of being limited to on-site early warning, allowing managers to know about anomalies in a timely manner without being on-site, significantly shortening the anomaly response time and avoiding the expansion of problems due to delayed early warning.

[0017] It adopts a distributed database architecture, with a master database and multiple slave databases, replacing the traditional single database storage. It can quickly switch slave databases when the master database fails, avoiding data loss. At the same time, through multi-dimensional indexing, it manages data by data type, collection time, and source, solving the problem of low historical data query efficiency caused by the lack of indexes in traditional systems. It supports long-term data storage and efficient querying, and provides complete data traceability for equipment maintenance and load optimization. Attached Figure Description

[0018] Figure 1 A schematic diagram of the overall process structure of an embodiment of the present invention is shown; Figure 2 This diagram illustrates the connection structure of the data acquisition module in an embodiment of the present invention. Figure 3 This diagram illustrates the working structure of the data preprocessing module in an embodiment of the present invention. Figure 4 This diagram illustrates the workflow structure of the load analysis module in an embodiment of the present invention. Figure 5 A schematic diagram of the working state structure of the early warning module in an embodiment of the present invention is shown; Figure 6 This diagram illustrates the working structure of the data storage module in an embodiment of the present invention. Figure 7 A schematic diagram of the connection structure of the data storage module in an embodiment of the present invention is shown; Figure 8 A schematic diagram of the workflow structure of the remote monitoring module in an embodiment of the present invention is shown.

[0019] List of reference numerals in the attached diagram: 1. Data acquisition module; 2. Data preprocessing module; 3. Load analysis module; 4. Anomaly detection module; 5. Early warning module; 6. Data storage module; 7. Remote monitoring module; 101. Current sensor; 102. Tension sensor; 103. Vibration sensor; 104. Temperature sensor; 105. Flow sensor; 106. Concentration sensor. Detailed Implementation

[0020] To make the objectives, solutions, and advantages of the technical solutions of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings of specific embodiments of the present invention.

[0021] Example 1: Please refer to the accompanying drawings in the instruction manual. Figures 1 to 8 As shown: This invention proposes a paper machine wire section load anomaly monitoring system, comprising: a data acquisition module 1, a data preprocessing module 2, a load analysis module 3, an anomaly judgment module 4, an early warning module 5, a data storage module 6, and a remote monitoring module 7; The data acquisition module 1 is used to collect the operating data of the wire section of the paper machine. The operating data includes the current data of the wire section drive motor, the wire section tension data, the wire section vibration data, the wire section temperature data, the pulp flow rate data, and the pulp concentration data. The data acquisition module 1 includes a current sensor 101, a tension sensor 102, a vibration sensor 103, a temperature sensor 104, a flow sensor 105, and a concentration sensor 106. The data preprocessing module 2 is connected to the data acquisition module 1 and is used to preprocess the running data acquired by the data acquisition module 1. The preprocessing includes data cleaning, data standardization and data noise reduction. The load analysis module 3 is connected to the data preprocessing module 2 and is used to obtain the preprocessed operating data. Combined with the preset historical load data of the paper machine wire section, a real-time load analysis model of the paper machine wire section is constructed to calculate the real-time load value of the paper machine wire section. The anomaly judgment module 4 is connected to the load analysis module 3 and is used to set the normal range threshold of the paper machine wire section load, including static threshold and dynamic threshold. It compares the real-time load value calculated by the load analysis module 3 with the normal range threshold and, in conjunction with the changing trend of the real-time load analysis model, judges whether the paper machine wire section load is abnormal. The early warning module 5 is connected to the anomaly judgment module 4. When the anomaly judgment module 4 determines that the paper machine wire section load is abnormal, it issues a corresponding early warning signal according to the anomaly level (mild / moderate / severe). The early warning signals include audible and visual warnings, SMS warnings, and system pop-up warnings. The data storage module 6 is connected to the data acquisition module 1, the data preprocessing module 2, the load analysis module 3, and the anomaly judgment module 4 respectively, and is used to store the acquired raw operating data, preprocessed operating data, real-time load values, historical load data, and anomaly judgment results. The remote monitoring module 7 is connected to the load analysis module 3, the anomaly judgment module 4, and the data storage module 6, respectively, to enable remote users to view the real-time load value, operating data, and anomaly judgment results of the paper machine wire section, and also supports remote users to adjust the warning threshold.

[0022] The data acquisition module 1 includes a current sensor 101, a tension sensor 102, a vibration sensor 103, a temperature sensor 104, a flow sensor 105, and a concentration sensor 106. The current sensor 101 is installed on the power supply line of the drive motor in the mesh section to collect the current data of the drive motor. Tension sensor 102 is installed at tension detection points in the mesh section, such as the tension roller at the mesh inlet and the tension roller at the vacuum roll outlet, to collect tension data of the mesh section; Vibration sensor 103 is installed in key parts of the mesh frame, such as mesh crossbeams, vacuum box support seats, and drying cylinder bearing seats, to collect vibration data of the mesh. Temperature sensor 104 is installed on the mesh bearing, the surface of the drying cylinder and the outer wall of the slurry flow channel to collect temperature data at the corresponding locations. The flow sensor 105 is installed on the slurry conveying pipeline to collect slurry flow data; Concentration sensor 106 is installed at the outlet of the slurry mixing tank and the slurry inlet of the mesh section to collect slurry concentration data.

[0023] Among them, data cleaning in data preprocessing module 2 includes removing missing values, outliers, and duplicate values; For missing values: if the missing proportion is less than the preset proportion threshold, use linear interpolation or mean imputation to fill the missing values; if the missing proportion is greater than or equal to the preset proportion threshold, discard the data and record the discarding situation. For outliers: use Grubbs' test or box plot method to identify them, and use nearest neighbor replacement method or data smoothing method to handle them according to the degree of deviation; For duplicate values: delete the duplicate data entries directly, keeping only 1 valid data.

[0024] In particular, the data standardization in data preprocessing module 2 adopts the Z-score standardization method, with the following formula: Z represents the standardized data, X represents the original data, μ represents the mean of the original data, and σ represents the standard deviation of the original data.

[0025] In particular, the data denoising in data preprocessing module 2 adopts wavelet transform denoising method, and the specific steps are as follows: (S11) Perform wavelet decomposition on the running data, select the appropriate wavelet basis function and the number of decomposition levels, and obtain multi-scale wavelet coefficients; (S12) Process the wavelet coefficients according to the preset threshold rule to suppress the coefficients corresponding to noise; (S13) Reconstruct the processed wavelet coefficients to obtain the denoised data.

[0026] Among them, load analysis module 3 uses a BP neural network model to construct a real-time load analysis model; Input layer: 6 neurons, corresponding to the preprocessed data of "current, tension, vibration, temperature, flow rate, and concentration"; Hidden layers: at least 2 layers, with the number of neurons in each layer set to 12-16 depending on the input dimension and accuracy requirements; Output layer: 1 neuron, corresponding to the real-time load value of the paper machine wire section; Model training: Gradient descent is used, with historical load data and corresponding operational data as samples, to adjust network weights and thresholds so that the prediction error is less than or equal to the preset error threshold.

[0027] The load normal range threshold of the anomaly detection module 4 includes: Static threshold: determined based on paper machine wire section design parameters, rated load, and historical normal operation data, such as ±10% of rated load; Dynamic threshold: Automatically adjusted based on real-time production speed, slurry type, and product specifications. The adjustment algorithm uses a "working condition parameter-threshold mapping table". Anomaly detection logic: When the real-time load value exceeds the normal range threshold and the real-time load analysis model shows that the load is "continuously deviating from the normal trend", it is determined to be an abnormal load.

[0028] The correspondence between the anomaly level and the warning method in the early warning module 5 is as follows: Mild anomaly: The real-time load value exceeds the threshold by less than 5%, and the load deviation trend is slow → Only audible and visual warning is issued, and the on-site audible and visual alarm is activated once every 30 seconds. Moderate anomaly: 5% ≤ real-time load value exceeds the threshold by less than 10%, and the load deviation trend is moderate → audible and visual alarm + SMS alarm, sending an alarm SMS to the mobile phone of the on-duty manager, containing the abnormal parameters; Severe anomaly: Real-time load value exceeds the threshold by ≥10%, and the load deviation trend is rapid → audible and visual warning + SMS warning + system pop-up warning.

[0029] Among them, data storage module 6 adopts a distributed database; Master database: Receives and stores all data, and synchronizes data to slave databases; From the database: Backs up the main database data, and automatically switches to the new main database when the main database fails, with a switching time of ≤10s; Data management: Creates indexes by "data type, acquisition time, and data source", and supports multi-dimensional queries, such as "tension data of tension sensor 102 in the past 24 hours".

[0030] The remote monitoring module 7 includes a web client and a mobile client; Web client: Accessed via browser, features include "real-time data viewing, historical data query, and report generation"; Mobile client: Mobile APP supports Android / iOS, features include "early warning information reception, real-time data preview, and abnormal record viewing", an early warning prompt will automatically pop up when there is a severe abnormality, which needs to be closed manually; User access control: There are three levels of access: "administrator, technician, and operator". Account passwords are stored in encrypted form, such as SHA-256 encryption.

[0031] The specific usage and function of this embodiment are as follows: Step 1: Install data acquisition module 1: Current sensor 101 (model ACS758): Installed on the three-phase power supply line of the drive motor (model Y355M-4), and fixed by a snap-on method to avoid the influence of line vibration; Tension sensor 102 (model HBMU9C): Two sensors are symmetrically installed on both sides of the tension roller (diameter 200mm) at the entrance of the mesh, and are fixed to the tension roller bracket with bolts. The distance between the sensor and the roller surface is 5mm. Vibration sensor 103 (model PCB352C33): One is installed in the middle of the crossbeam (material Q345), one in the vacuum box support, and one in the drying cylinder bearing seat. It is fixed by a magnetic base for easy disassembly and calibration later. Temperature sensor 104 (model PT100A): 2 on the end cover of the drive motor bearing, 3 on the surface of the drying cylinder, evenly distributed, and 2 on the outer wall of the slurry flow channel, using threaded installation, penetrating 5mm into the installation surface; Flow sensor 105 (model E+HPromag50W): Installed on the main slurry conveying pipeline, with the installation position ≥ 5 times the pipe diameter away from the pipe bend to avoid the influence of fluid disturbance; Concentration sensor 106 (model Optech NIR-2000): One is installed at the outlet of the slurry mixing tank and one at the slurry inlet main of the mesh section. It is flange-mounted and coaxial with the pipeline.

[0032] Step 2: Deploy other modules: Data preprocessing module 2 (Siemens S7-1200): Installed in the electrical control cabinet of the network department, temperature 0-40℃, humidity ≤85%, connected to each sensor via RS485 bus, bus length ≤100m; Load analysis module 3: Installed in the central control room, ≤500m away from the network department, and connected to data preprocessing module 2 via Ethernet. The network cable is Category 5e shielded cable. Data storage module 6 uses 3 industrial servers: deployed in the enterprise computer room at a temperature of 18-25℃, equipped with UPS power supply, and connected to the industrial computer in the central control room via fiber optic cable; Remote monitoring module 7: The web client is deployed on the central control room server for on-site viewing, while the mobile client is downloaded and installed by administrators through the company's internal app store.

[0033] System debugging Step 1, Sensor Calibration: Calibrate each sensor. For example, apply a standard current signal of 0-500A to current sensor 101 and record the output value to ensure that the error is ≤ ±0.2%FS. The tension sensor 102 is calibrated using the "standard weight loading method". A weight of 0-50kN is loaded, and the output tension value is recorded. The error is ≤±0.5%FS.

[0034] The second step is model training: Import historical data from the past year into data storage module 6, including normal operation data and known abnormal data, and train a BP neural network model in load analysis module 3. Adjust model parameters, such as the number of neurons in the hidden layer and the learning rate, to make the model's accuracy in identifying historical abnormal data ≥95%, such as 98% accuracy in identifying "abnormal load caused by increased slurry concentration".

[0035] Step 3, threshold setting: The static threshold is set at 120kW ± 10%, which is 108-132kW; The dynamic threshold is pre-stored for the range corresponding to three sets of operating conditions: Vehicle speed 1000m / min, slurry concentration 2%: 105-129kW; Vehicle speed 1200m / min, slurry concentration 2.5%: 108-132kW; Vehicle speed 1400m / min, slurry concentration 3%: 111-135kW.

[0036] Running Example Taking "abnormal increase in slurry concentration leading to abnormal load on the wire mesh" as an example, the system operation process is as follows: Step 1, Data Acquisition: Concentration sensor 106 detected that the slurry concentration increased from 2.5% to 4%, while flow sensor 105 detected that the slurry flow rate remained stable at 50 m³ / s. 3 / h, the current sensor 101 collects the drive motor current from 300A to 350A; the second step, data preprocessing: the data preprocessing module 2 cleans and standardizes the above raw data, converts the concentration of 4% and the current of 350A into Z values, removes noise from the concentration data caused by the slurry flow, and transmits it to the load analysis module 3. The third step is load calculation: Input the "concentration Z value, current Z value, tension / vibration / temperature / flow rate Z value" into the BP model to calculate the real-time load value of 140kW. The fourth step is to determine the anomaly: The current operating conditions are "vehicle speed 1200m / min, slurry concentration 4%", the dynamic threshold is adjusted to 111-135kW, the real-time load value of 140kW exceeds the threshold by 5kW, the amplitude is 3.7%, and for 5 consecutive data points, the deviation from the threshold direction changes within 500ms, which is determined to be "slight anomaly". Step 5, Warning Trigger: Warning module 5 activates the audible and visual alarm, the red light flashes once every 30 seconds, and at the same time, a yellow "minor anomaly" indicator is displayed on the web client and mobile client of remote monitoring module 7; Step 6, Anomaly Handling: The on-duty technician checks the abnormal parameters through the remote monitoring module 7 and finds that the slurry concentration has increased abnormally. He notifies the on-site operator to adjust the slurry dilution water valve. After the concentration drops to 2.5%, the real-time load value drops to 125kW. The anomaly judgment module 4 determines that the load is normal, the early warning module 5 stops the audible and visual alarm, and the system returns to normal operation. Step 7, Data storage: The entire process of "original sensor data, preprocessed data, real-time load value 140kW→125kW, and anomaly judgment result of slight anomaly→normal" is stored in data storage module 6, generating anomaly records, including the anomaly cause "increased slurry concentration" and the handling measure "adjusting the dilution water valve".

Claims

1. A paper machine wire section load anomaly monitoring system, characterized in that, include: The system includes a data acquisition module (1), a data preprocessing module (2), a load analysis module (3), an anomaly detection module (4), an early warning module (5), a data storage module (6), and a remote monitoring module (7). The data acquisition module (1) is used to collect the operating data of the wire section of the paper machine. The operating data includes the current data of the wire section drive motor, the wire section tension data, the wire section vibration data, the wire section temperature data, the pulp flow rate data, and the pulp concentration data. The data acquisition module (1) includes a current sensor (101), a tension sensor (102), a vibration sensor (103), a temperature sensor (104), a flow sensor (105), and a concentration sensor (106). The data preprocessing module (2) is connected to the data acquisition module (1) and is used to preprocess the running data acquired by the data acquisition module (1). The preprocessing includes data cleaning, data standardization and data noise reduction. The load analysis module (3) is connected to the data preprocessing module (2) and is used to obtain the preprocessed operating data, combine it with the preset historical load data of the paper machine wire section, construct the real-time load analysis model of the paper machine wire section, and calculate the real-time load value of the paper machine wire section. The anomaly judgment module (4) is connected to the load analysis module (3) and is used to set the normal range threshold of the paper machine wire section load. The real-time load value calculated by the load analysis module (3) is compared with the normal range threshold. At the same time, the change trend of the real-time load analysis model is combined to determine whether the paper machine wire section load is abnormal. The early warning module (5) is connected to the anomaly judgment module (4) and is used to issue a corresponding early warning signal according to the anomaly level when the anomaly judgment module (4) judges that the paper machine wire section load is abnormal. The early warning signal includes sound and light warning, SMS warning and system pop-up warning. The data storage module (6) is connected to the data acquisition module (1), the data preprocessing module (2), the load analysis module (3) and the anomaly judgment module (4) respectively, and is used to store the acquired raw operating data, preprocessed operating data, real-time load value, historical load data and anomaly judgment result; The remote monitoring module (7) is connected to the load analysis module (3), the anomaly judgment module (4) and the data storage module (6) respectively, and is used to enable remote users to view the real-time load value, operation data and anomaly judgment results of the paper machine wire section, and at the same time support remote users to adjust the warning threshold.

2. The paper machine wire section load anomaly monitoring system as described in claim 1, characterized in that, The data acquisition module (1) includes a current sensor (101), a tension sensor (102), a vibration sensor (103), a temperature sensor (104), a flow sensor (105), and a concentration sensor (106). The current sensor (101) is installed on the power supply line of the drive motor and is used to collect the current data of the drive motor. The tension sensor (102) is installed at the tension detection point of the net section and is used to collect the tension data of the net section; The vibration sensor (103) is installed in a key part of the mesh frame and is used to collect vibration data of the mesh. The temperature sensor (104) is installed on the mesh bearing, the surface of the drying cylinder, and the outer wall of the slurry flow channel to collect temperature data at the corresponding locations; The flow sensor (105) is installed on the slurry conveying pipeline to collect slurry flow data; The concentration sensor (106) is installed at the outlet of the slurry mixing tank and the slurry inlet of the mesh section to collect slurry concentration data.

3. The paper machine wire section load anomaly monitoring system as described in claim 1, characterized in that, The data cleaning in the data preprocessing module (2) includes removing missing values, outliers, and duplicate values; For missing values: if the missing proportion is less than the preset proportion threshold, use linear interpolation or mean imputation to fill the missing values; if the missing proportion is greater than or equal to the preset proportion threshold, discard the data and record the discarding situation. For outliers: use Grubbs' test or box plot method to identify them, and use nearest neighbor replacement method or data smoothing method to handle them according to the degree of deviation. For duplicate values: delete the duplicate data entries directly, keeping only 1 valid data.

4. The paper machine wire section load anomaly monitoring system as described in claim 1, characterized in that, The data standardization in the data preprocessing module (2) adopts the Z-score standardization method, and the formula is: Z represents the standardized data, X represents the original data, μ represents the mean of the original data, and σ represents the standard deviation of the original data.

5. The paper machine wire section load anomaly monitoring system as described in claim 1, characterized in that, The data denoising in the data preprocessing module (2) adopts wavelet transform denoising method, and the specific steps are as follows: (S11) Perform wavelet decomposition on the running data, select the appropriate wavelet basis function and the number of decomposition levels, and obtain multi-scale wavelet coefficients; (S12) Process the wavelet coefficients according to the preset threshold rule to suppress the coefficients corresponding to noise; (S13) Reconstruct the processed wavelet coefficients to obtain the denoised data.

6. The paper machine wire section load anomaly monitoring system as described in claim 1, characterized in that, The load analysis module (3) uses a BP neural network model to construct a real-time load analysis model; Input layer: 6 neurons, corresponding to the preprocessed data of "current, tension, vibration, temperature, flow rate, and concentration"; Hidden layers: at least 2 layers, with the number of neurons in each layer determined by the input dimension and accuracy requirements; Output layer: 1 neuron, corresponding to the real-time load value of the paper machine wire section; Model training: Gradient descent is used, with historical load data and corresponding operational data as samples, to adjust network weights and thresholds so that the prediction error is less than or equal to the preset error threshold.

7. The paper machine wire section load anomaly monitoring system as described in claim 1, characterized in that, The load normal range threshold of the anomaly judgment module (4) includes: Static threshold: determined based on paper machine wire section design parameters, rated load, and historical normal operation data. Dynamic threshold: Automatically adjusted based on real-time operating conditions, using an "operating condition parameter-threshold mapping table" as the adjustment algorithm; Anomaly detection logic: When the real-time load value exceeds the normal range threshold and the real-time load analysis model shows that the load is "continuously deviating from the normal trend", it is determined to be an abnormal load.

8. The paper machine wire section load anomaly monitoring system as described in claim 1, characterized in that, The correspondence between the anomaly level and the warning method of the early warning module (5) is as follows: Mild anomaly: Real-time load value exceeds the threshold by less than 5%, and the load deviation trend is slow → only audible and visual warning; Moderate anomaly: 5% ≤ real-time load value exceeds the threshold by less than 10%, and the load deviation trend is moderate → audible and visual warning + SMS warning; Severe anomaly: Real-time load value exceeds the threshold by ≥10%, and the load deviation trend is rapid → audible and visual warning + SMS warning + system pop-up warning.

9. The paper machine wire section load anomaly monitoring system as described in claim 1, characterized in that, The data storage module (6) adopts a distributed database master database and 3-5 slave databases; Master database: Receives and stores all data, and synchronizes data to slave databases; From the database: Back up the data of the primary database, and automatically switch to the new primary database when the primary database fails. Switching time ≤10s; Data management: Indexes are created based on "data type (raw / preprocessed / load value / abnormal results), acquisition time (partitioned by hour), and data source (sensor / module)," supporting multi-dimensional queries.

10. The paper machine wire section load anomaly monitoring system as described in claim 1, characterized in that, The remote monitoring module (7) includes a web client and a mobile client; Web client: Accessed via a browser, features include "real-time data viewing of sensor data, real-time load values, historical data querying, and report generation"; Mobile client: A mobile app with functions including "receiving early warning information, previewing real-time data, and viewing anomaly records". It automatically pops up an early warning prompt when there are severe anomalies. User access control: There are three levels of access: "administrator, technician, and operator". Account passwords are stored in encrypted form.

Images