Enameled wire production and processing system

Abstract

The invention relates to the technical field of enameled wire production and processing, and discloses an enameled wire production and processing system which comprises a pay-off device, an annealing device, a painting device, a drying device, a cooling device, a take-up device, a detection device and a waste gas treatment device, and is characterized in that a sealing monitoring system is arranged in the waste gas treatment device; the sealing performance monitoring system comprises a fault model construction module which is used for obtaining historical monitoring data and historical fault data of a sealing performance monitoring point and constructing a fault prediction model; and the monitoring point rationality judgment module is used for analyzing fault early warning data of the sealing monitoring points based on the constructed fault prediction model. By judging the early warning effectiveness of the monitoring points, evaluating the actual operation performance of the monitoring points, accurately sending out fault early warning and distinguishing reasonable and unreasonable monitoring points according to reasonable judgment values, a basis is provided for optimizing the layout of the monitoring points, the reliability and effectiveness of a sealing monitoring system are improved, and safe operation of the sealing system is guaranteed.

Description

Technical Field

[0001] The present invention relates to the technical field of enameled wire production and processing, and specifically relates to a production and processing system for enameled wire. Background Art

[0002] Enameled wire is a main variety of winding wire, which consists of a conductor and an insulating layer. After the bare wire is annealed and softened, it is coated with paint and baked multiple times. The processes of painting and baking will generate waste gas containing volatile organic compounds, which will affect the health of workers. Therefore, these waste gases need to be purified by waste gas treatment equipment to ensure that the emissions meet environmental protection standards and avoid harm to the environment and human health. Therefore, waste gas treatment is an indispensable part of enameled wire production; However, during the process of processing enameled wire, when using waste gas treatment equipment to treat waste gas, since the connecting pipe heads are movably connected to the first pipe body and the second pipe body respectively, and the outer surface of the connecting pipe head is in sliding fit with their inner surfaces, during frequent movement and rotation, and the filter plate will move up and down, the sealing performance at the connection may decrease. Once the seal is not tight, waste gas may leak, which will not only reduce the treatment efficiency, but also may cause the waste gas concentration in the workshop to rise, endangering the health of workers; At the same time, during the process of monitoring the sealing performance, the layout of the sealing monitoring points may be unreasonable, and the sealing failure cannot be accurately and timely monitored, thus affecting the reliability and effectiveness of the sealing monitoring system. Summary of the Invention

[0003] The purpose of the present invention is to provide a production and processing system for enameled wire to solve at least one of the above-mentioned prior art problems.

[0004] The present invention provides a production and processing system for enameled wire, including a wire feeding device, an annealing device, a painting device, a drying device, a cooling device, a wire winding device, a detection device and a waste gas treatment device. It is characterized in that a sealing performance monitoring system is provided in the waste gas treatment device, and the sealing performance monitoring system includes: A fault model construction module: obtaining the historical monitoring data and historical fault data of the sealing monitoring points, and constructing a fault prediction model; A monitoring point rationality judgment module: based on the constructed fault prediction model, analyzing the fault warning data of the sealing monitoring points, processing to obtain a rational judgment value, and judging the rationality of the setting of the sealing monitoring points; A monitoring point adjustment module: based on the unreasonable setting of the sealing monitoring points, analyzing and obtaining the adjustment priority level of the unreasonable monitoring.

[0005] As a further solution of the present invention: the construction process of the fault prediction model: Obtain the historical monitoring data and historical fault data of all seal tightness monitoring points, and use a neural network model to train a fault prediction model.

[0006] As a further solution of the present invention: The process of judging the rationality of the setting of the seal tightness monitoring points is as follows: Obtain a rational judgment value, and mark the seal tightness monitoring points with a value greater than the rationality judgment threshold as unreasonable monitoring points.

[0007] As a further solution of the present invention: The process of obtaining the rational judgment value is as follows: Based on the fault prediction model, obtain the predicted fault occurrence time, and combine with the real-time fault data for data processing to obtain the proportion of abnormal points, the proportion of consecutive points, the proportion of unreasonable early warning numbers, and the proportion of invalid warning numbers; Perform a weighted calculation on the proportion of abnormal points, the proportion of consecutive points, the proportion of unreasonable early warning numbers, and the proportion of invalid warning numbers to obtain a rational judgment value; As a further solution of the present invention: The process of obtaining the proportion of invalid warning numbers is as follows: Preset an analysis period, and obtain real-time monitoring data and real-time fault data within the analysis period; Input the real-time data of each seal tightness monitoring point into the trained fault prediction model to obtain the predicted fault data corresponding to each monitoring point; If the real-time fault occurrence time is earlier than or equal to the predicted fault occurrence time, it is marked as an invalid warning; Obtain the number of invalid warnings, and calculate the ratio with the total number of faults to obtain the proportion of invalid warning numbers.

[0008] As a further solution of the present invention: The process of obtaining the proportion of unreasonable early warning numbers is as follows: If the real-time fault occurrence time is earlier than or equal to the predicted fault occurrence time, and if the real-time fault occurrence time is later than the predicted fault occurrence time, it is marked as a valid warning; Based on the valid warnings, calculate the difference between the real-time fault occurrence time and the predicted fault occurrence time to obtain the warning time deviation value, and construct a warning time deviation sequence; Extract the warning time deviation values greater than the warning time deviation limit value, count their number, and calculate the ratio with the number of data points in the warning time deviation value data sequence to obtain the proportion of unreasonable early warning numbers.

[0009] As a further solution of the present invention: The process of obtaining the proportion of abnormal points and the proportion of consecutive points is as follows: Through an anomaly detection algorithm, obtain an anomaly detection score; If the anomaly detection score is greater than the anomaly threshold, mark the corresponding data point as an abnormal point; Count the number of abnormal points, and calculate the ratio with the number of data points in the data sequence of the early warning time deviation value to obtain the proportion of the number of abnormal points. Integrate consecutive abnormal points to obtain abnormal point combinations, respectively count their consecutive numbers for mean processing, and then calculate the ratio with the number of abnormal points to obtain the proportion of consecutive numbers.

[0010] As a further solution of the present invention: the process of obtaining the abnormal detection score is as follows: The abnormal detection algorithm includes the Isolation Forest algorithm; Analyze the early warning time deviation value sequence through the Isolation Forest algorithm to obtain the abnormal detection score of each data point.

[0011] As a further solution of the present invention: the process of analyzing and obtaining the adjustment priority level of unreasonable monitoring is as follows: Obtain all unreasonable monitoring points and obtain their corresponding positions and rationality judgment values; Determine the range of the key area, respectively measure the distance values from the unreasonable monitoring points to the center of each key area, and perform weighted summation calculation on all the distance values to obtain the position influence value; Cluster the rationality judgment values and position influence values of each monitoring point through a clustering algorithm; According to the clustering results, divide the unreasonable monitoring points in different clusters into different adjustment levels.

[0012] As a further solution of the present invention: the process of dividing into different adjustment levels is as follows: For each cluster after clustering, calculate their quartiles respectively for the rationality judgment value and position influence value of each cluster after clustering; For the rationality judgment value, they are respectively 、 、 , and for the position influence value, they are respectively 、 、 ; For each cluster, if the rationality judgment value ≥ and the position influence value ≥ , then divide this cluster into the high-priority adjustment category; if the rationality judgment value < and the position influence value < , then divide this cluster into the low-priority adjustment category; otherwise, divide this cluster into the medium-priority adjustment category.

[0013] The beneficial effects of the present invention: 1. The present invention constructs a model by obtaining historical monitoring and fault data, verifies the rationality of the monitoring point positions, traces the fault monitoring points, reduces manual troubleshooting, improves work efficiency, and can also dynamically layout the monitoring point positions. 2. The present invention determines the effectiveness of the monitoring point warnings, evaluates its actual operating performance, accurately issues fault warnings, distinguishes reasonable and unreasonable monitoring points according to the rationality judgment value, provides a basis for optimizing the monitoring point layout, improves the reliability and effectiveness of the seal monitoring system, and ensures the safe operation of the seal system. 3. The present invention divides and adjusts the levels according to the unreasonable situations of the monitoring points, takes measures for different levels, further optimizes the monitoring point layout, improves the monitoring accuracy and reliability, effectively reduces waste gas leakage, protects the health of the staff, and improves the safety and stability of the waste gas treatment link in the enameled wire production and processing process. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0015] Figure 1 It is a process architecture diagram for the analysis of the layout of the seal monitoring points in a production and processing system of enameled wire according to the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0016] In order to enable those skilled in the art to better understand the solution of the present invention, the following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the drawings in the embodiments of the present invention. Obviously, the described embodiments are only some of the embodiments of the present invention, rather than all of them. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0017] Embodiment 1 A production and processing system of enameled wire provided in an embodiment of the present invention includes: an exhaust gas treatment device for enameled wire processing, and its use process is as follows: Exhaust gas treatment: The exhaust gas passes through the coarse filtration of the pushing blade and the fine filtration of the filter plate in sequence, and then reacts with ozone under the irradiation of the ultraviolet lamp to be oxidized and decomposed into harmless gases and discharged. The waste gas generated from the processing of enameled wire is introduced into the first pipe body. The waste gas passes through the through holes on the pushing blades, and the large particulate dust in the waste gas is blocked by the pushing blades and drops onto the inner wall of the first pipe body. Then the waste gas passes through the filter plate, and the small particulate dust in the waste gas is filtered out by the filter plate and adheres to the filter plate. The filtered waste gas passes through the second pipe body and enters the third pipe body. Ozone is added to the air inlet, and the ozone enters the third pipe body. The ultraviolet lamp emits ultraviolet rays. Under the action of ultraviolet rays, the waste gas is oxidized by ozone into harmless gas and discharged from the outlet at one end of the third pipe body far away from the second pipe body; Large particulate cleaning: The separation pipe body is rotated by the cylinder and the motor drives the pipe body to rotate. The pushing blades are used to push the deposited dust to the inclined collection tank; Start the push-pull cylinder to drive the connecting pipe head to move, so that the connecting pipe head is separated from the second pipe body. Start the first rotating motor to drive the second pipe body to rotate through the third gear and the fourth gear. Start the second rotating motor to drive the first pipe body to rotate through the first gear and the second gear, and drive the pushing blades to rotate. The pushing blades push the large particulate dust inside the first pipe body to move away from the second pipe body, so that the large particulate dust is discharged from the first pipe body and falls downward. The large particulate dust passes through the first opening and falls onto the inclined surface of the inclined block, and slides along the inclined surface into the collection tank; Filter plate maintenance: The lifting mechanism lowers the filter plate, and the scraper and the vibration motor cooperate to remove the adhered dust. After the pressing cylinder is unlocked, the filter plate can be removed for maintenance; Start the lifting motor, and the lifting motor drives the lifting belt to rotate, so that the two lifting belts of the lifting mechanism rotate in opposite directions. The two lifting belts drive the lifting end plate to descend and drive the filter plate to descend. Start the moving cylinder, and the moving cylinder drives the scraper to move, so that one end of the scraper is in close contact with one side of the filter plate. As the filter plate descends, the small particulate dust on the surface of the filter plate is scraped off. At the same time, start the vibration motor, and the vibration motor drives the filter plate to vibrate, so that the remaining dust on the filter plate is shaken off. The small particulate dust passes through the second opening downward and falls into the collection tank. Pull out the collection tank through the push-pull handle to process the dust in the collection tank. Start the two pressing cylinders in the lifting mechanism, and the pressing cylinders drive the moving blocks to move. The moving blocks drive the crawler to move through the extrusion rods, so that the two lifting belts in the lifting mechanism move away from each other, and the lifting end plate is separated from the two lifting belts, and the filter plate is removed for maintenance.

[0018] Embodiment 2 Based on the above embodiment, as Figure 1 shown, a production and processing system for enameled wire provided by an embodiment of the present invention further includes: monitoring the sealing performance of the waste gas treatment device, thereby realizing the monitoring of the sealing performance of the waste gas treatment device, timely discovering the problem of decreased sealing performance, and reducing the leakage of waste gas; And the analysis of the layout of the sealing monitoring points: Fault model construction module: Obtain the historical monitoring data and historical fault data of the sealing monitoring points, and construct a fault prediction model through machine learning algorithms; In some embodiments, obtain historical monitoring data from each sealing monitoring point of the waste gas treatment device, where the historical monitoring data includes but is not limited to: gas concentration, pressure, temperature, and flow rate; It should be noted that the obtained historical monitoring data covers a relatively long time period, so as to capture the operating status of the waste gas treatment device under different working conditions; Obtain the historical fault data of the waste gas treatment device, where the historical fault data includes but is not limited to: the time, type, and severity of the fault; The fault types include but are not limited to: seal leakage, equipment damage, and pipeline blockage; Exemplarily, when a seal leakage fault occurs, record the specific date and time of the fault occurrence, as well as the approximate location and degree of the leakage; Check the obtained data, and remove missing values, outliers, and duplicate data; Exemplarily, if the pressure data of a certain monitoring point shows a sudden abnormally high value, these outliers need to be removed; Normalize different types of monitoring data so that they have the same scale range. Among them, the normalization method can use min-max normalization or Z-score normalization, which is specifically selected by those skilled in the art according to experience; Exemplarily, for gas concentration and pressure data, use min-max normalization to scale the value ranges to between [0,1]; Extract features according to the historical monitoring data and historical fault data; Exemplarily, calculate the change rate of gas concentration, the fluctuation amplitude of pressure, etc. as new features; at the same time, use the fault occurrence situation as a label to construct a training data set; Select a model, where the model selection includes but is not limited to: neural network models and support vector machines (SVM); According to the selected model type, define the structure and parameters of the model, and randomly initialize the parameters of the model; Exemplarily, for the multi-layer perceptron network in the neural network model, define the number of neurons in the input layer, hidden layer, and output layer, as well as the type of activation function; Divide the preprocessed data set into a training set, a validation set, and a test set; Among them, the training set accounts for 70%-80% of the total data, the validation set accounts for 10%-15%, and the test set accounts for 10%-15%. The training set is used for parameter learning of the model, the validation set is used to adjust the hyperparameters of the model, and the test set is used to evaluate the final performance of the model; The selected model is trained using the training set, and the parameters of the model are continuously adjusted through an optimization algorithm to minimize the error between the prediction result of the model and the true label; During the training process, the validation set is used to verify the performance of the model to avoid overfitting; The trained model is evaluated using the test set; Exemplarily, for the fault prediction problem, accuracy is used to measure the accuracy of the model in predicting the occurrence of faults, and recall is used to measure the fault detection ability of the model; Based on the evaluation result of the model, the model is optimized to obtain a fault prediction model; The core purpose of building the fault prediction model is to realize the early warning and location of the seal fault of the waste gas treatment device through machine learning algorithms, so as to verify the rationality of the monitoring point positions; And it is possible to trace the monitoring points of the faults, thereby reducing manual troubleshooting and increasing work efficiency; Through the analysis of the model, the contribution degree of each seal monitoring point to the fault prediction is quantified, and based on the contribution degree, the positions of the seal monitoring points are dynamically arranged; Monitoring point rationality judgment module: Based on the built fault prediction model, analyze the fault early warning data of the seal monitoring points and judge the rationality of the settings of the seal monitoring points; In some embodiments, a preset analysis period is set, and real-time monitoring data and real-time fault data are obtained within the analysis period; The real-time data of each seal monitoring point is input into the trained fault prediction model to obtain the predicted fault data corresponding to each monitoring point; For each seal monitoring point; Compare and judge the real-time fault occurrence time and the predicted fault occurrence time. If the real-time fault occurrence time is earlier than or equal to the predicted fault occurrence time, it is marked as an invalid early warning. If the real-time fault occurrence time is later than the predicted fault occurrence time, it is marked as a valid early warning; Obtain the number of invalid early warnings and calculate the ratio with the total number of faults to obtain the ratio of invalid early warnings; Based on the valid early warnings, calculate the difference between the real-time fault occurrence time and the predicted fault occurrence time to obtain the early warning time deviation value, and construct an early warning time deviation sequence; The early warning time deviation value intuitively reflects the accuracy and advance degree of the monitoring point early warning time, and is a key indicator for evaluating the performance of the monitoring point. The sequence constructed based on this value can be used to analyze the early warning deviation situation of the monitoring point at different times, providing a data basis for subsequent anomaly detection and rationality judgment; Analyze the warning time deviation value sequence through an anomaly detection algorithm to obtain the anomaly detection score for each data point, and obtain the increasing part of the warning time deviation value sequence. The specific process is as follows: Preprocess the data in the warning time deviation value sequence, including data cleaning and data standardization; Use the Isolation Forest algorithm in the anomaly detection algorithm for analysis. Randomly divide the data space to construct multiple isolation trees, and then judge whether a data point is an anomaly point according to the path length of the data point reaching the leaf node in the tree. The shorter the path, the more likely the data point is an anomaly point; The reason for choosing the Isolation Forest algorithm is that the data of the warning time deviation value sequence may be affected by various factors, showing high-dimensional and complex distribution characteristics. The Isolation Forest algorithm does not need to assume the distribution form of the data and can effectively capture the normal patterns and anomaly points of the data in the high-dimensional space; Randomly extract values from the preprocessed warning time deviation sequence to construct multiple isolation trees to form an isolation forest; The construction process of each tree is to randomly select a feature and a splitting point to divide the data into two parts, and repeat this process until each node contains only one sample or reaches the preset tree depth; Through Isolation Forest training, learn the normal distribution pattern of the warning time deviation value, that is, the average path length of normal data points in the tree; For the remaining warning time deviation values, input them into the trained Isolation Forest respectively, calculate the average value of the path lengths in each isolation tree respectively. The shorter the path length, the higher the corresponding anomaly detection score, indicating that the degree of deviation of this data point from the normal pattern is greater; It should be noted that the severity of anomalies at different detection points can be distinguished by the level of the anomaly score; The anomaly detection score can be obtained by normalizing the path length. The path length ranges from [0,1], and the closer it is to 1, the more it represents an anomaly point; Set an anomaly threshold. If the anomaly detection score is greater than the anomaly threshold, mark the corresponding warning time deviation value data point as an anomaly point; otherwise, mark the corresponding warning time deviation value data point as a non-anomaly point; Count the number of anomaly points and calculate the ratio with the number of data points in the warning time deviation value data sequence to obtain the proportion of the number of anomaly points; Integrate the continuous anomaly points to obtain anomaly point combinations, and respectively count their continuous numbers for averaging, and then calculate the ratio with the number of anomaly points to obtain the proportion of continuous numbers; Set a warning time deviation limit value, which is set according to work experience and is used to evaluate the advance degree of the warning time; Extract the warning time deviation values greater than the warning time deviation limit value, count their quantities, calculate the ratio with the number of data points in the warning time deviation value data sequence, and obtain the ratio of the number of unreasonable early warnings; Perform weighted calculation on the proportion of abnormal points, the proportion of consecutive points, the ratio of the number of unreasonable early warnings, and the ratio of invalid warnings to obtain a rationality judgment value; Set a rationality judgment threshold, mark the seal monitoring points greater than the rationality judgment threshold as unreasonable monitoring points, and mark the seal monitoring points less than or equal to the rationality judgment threshold as reasonable monitoring points; The monitoring point rationality judgment module can judge the warning effectiveness of the seal monitoring points, thereby evaluating the performance of the monitoring points in actual operation, and further the accuracy of issuing fault warnings; Mark the seal monitoring points according to the rationality judgment value, distinguish reasonable and unreasonable monitoring points, provide a basis for optimizing the layout of the monitoring points, and then adjust the positions of the unreasonable monitoring points to improve the reliability and effectiveness of the entire seal monitoring system and ensure the safe operation of the seal system; Monitoring point adjustment module: Based on the unreasonable setting of the seal monitoring points, analyze and obtain the adjustment priority levels of the unreasonable monitoring, and perform corresponding adjustments; In some embodiments, obtain all unreasonable monitoring points and their corresponding positions; For each unreasonable monitoring point, obtain its corresponding rationality judgment value, so that the adjustment priority of the unreasonable monitoring point can be determined through the rationality judgment value; Determine the scope of the key area, and the division of this key area is summarized and set by those skilled in the art according to the connection points of the waste gas treatment device combined with past work experience; Exemplarily, a region with a radius of 50 centered on the connecting pipe head is divided into a key area; Measure the distance values from the unreasonable monitoring points to the center of each key area respectively, and perform weighted summation calculation on all the distance values to obtain a position influence value; Among them, when calculating the position influence value using the distance values from the unreasonable points to each key area, the weight coefficients are set according to the high and low degrees of the probability of failure. For example, due to frequent activities of the connecting pipe head, the probability of seal leakage failure is relatively high, and its weight is set to 0.7, while the relative failure probability of the filter plate is relatively low, and the weight is set to 0.3; Perform dimensionless processing on the rationality judgment value and position influence value data of all unreasonable monitoring points. Among them, common dimensionless methods include but are not limited to: min-max normalization; Use a clustering algorithm for clustering analysis, and the K-means clustering algorithm can be selected; Select an appropriate K value through methods such as the elbow method and the silhouette coefficient method; Randomly select K points in the data space as the initial centroids; Calculate the distance values from each data point to each centroid (for example, use the Euclidean distance to calculate the distance values), and assign the data points to the cluster where the nearest centroid is located; The Euclidean distance formula is: , where and are the coordinates of two data points respectively, x represents the reasonable judgment value, y represents the position influence value, and the above data are all dimensionless values; Recalculate the mean of the data points within each cluster and use it as the new centroid. Repeat the steps of assigning data points and updating the centroid until the centroid no longer changes or changes very little, and the clustering process converges; According to the clustering results, divide the monitoring points in different clusters into different adjustment levels; The division of adjustment levels includes: for each cluster after clustering, calculate the quartiles of their reasonable judgment values and position influence values respectively. The quartiles are the values that divide the data into four equal parts, namely the lower quartile (Q1), the median (Q2), and the upper quartile (Q3); For the reasonable judgment value, they are respectively 、 、 , and for the position influence value, they are respectively 、 、 ; For each cluster, if the reasonable judgment value ≥ and the position influence value ≥ , then divide this cluster into the high-priority adjustment category and it needs to be adjusted urgently; For such unreasonable monitoring points, both the reasonable judgment value and the position influence value are relatively high, indicating poor monitoring effect and far from the key area. It is necessary to give priority to migrating them to the key area. For example, taking the connection pipe head as the key area, if the monitoring point is currently far from the connection pipe head, it can be migrated to within 10 cm of the connection pipe head (specifically set according to the actual situation) to improve the monitoring ability for faults such as seal leakage; In addition to position migration, in order to improve the accuracy and reliability of monitoring, it is necessary to upgrade the monitoring equipment. It can be replaced with a higher-precision and more sensitive sensor. For example, upgrade the gas concentration sensor with a detection accuracy of ±5 ppm to a sensor with ±1 ppm. At the same time, optimize the data acquisition frequency and transmission method of the monitoring equipment, increase the data acquisition frequency, from the original once every 10 minutes to once every 3 - 5 minutes, to ensure that monitoring data can be obtained in a timely manner; If the reasonable judgment value < and the position influence value < , then divide this cluster into the low-priority adjustment category and the adjustment requirement is relatively low; For such unreasonable monitoring points, continue to monitor the operating status of the monitoring points to promptly detect potential problems; Otherwise, divide this cluster into the medium-priority adjustment category, and the adjustment requirement is at a moderate level; For such unreasonable monitoring points, the reasonable judgment value and the position influence value are in the middle range. According to the position influence value, appropriately approach the key area; Conduct a comprehensive inspection of the monitoring equipment, including the accuracy and stability of the sensors, as well as the operating status of the data processing unit and the transmission line. Calibrate the sensors to ensure the accuracy of their measurements. For example, calibrate the gas concentration sensor with standard gas, check whether there is data loss or incorrect processing in the data processing unit, and repair the discovered problems; The technical solution of this embodiment is as follows: First, obtain the historical monitoring data and historical fault data of the seal monitoring points, construct a fault prediction model. Based on this model, obtain the real-time monitoring data and real-time fault data. After processing such as comparison and judgment and anomaly detection, calculate indicators such as the ratio of invalid early warnings and the ratio of the number of abnormal points, and obtain the reasonable judgment value through weighting. Use this to judge the rationality of the monitoring point settings, obtain the positions and reasonable judgment values of the unreasonable monitoring points, determine the key area, calculate the position influence value, divide the unreasonable monitoring points into different adjustment levels, and take corresponding adjustment measures for different levels; Thus, the effective monitoring of the seal of the waste gas treatment device is realized, the problem of decreased sealing performance is discovered, waste gas leakage is reduced. By constructing a fault prediction model, the rationality of the monitoring point positions can be verified, the effectiveness of the monitoring point early warnings can be judged, the accuracy of the fault early warnings can be improved, the monitoring points are adjusted according to the reasonable judgment results, the layout of the monitoring points is optimized, the reliability and effectiveness of the entire seal monitoring system are improved, and the safe operation of the seal system is guaranteed.

[0019] In the above embodiments, the descriptions of each embodiment have their own focuses. For the parts not detailed or recorded in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0020] The above formulas are all dimensionless and take their numerical values for calculation. The formula is obtained by collecting a large amount of data and performing software simulation to obtain a formula that is closest to the actual situation. The preset parameters in the formula are set by those skilled in the art according to the actual situation.

[0021] The above has described an embodiment of the present invention in detail, but the content described is only the preferred embodiment of the present invention and cannot be considered as limiting the scope of implementation of the present invention. All equivalent changes and improvements made within the scope of the application of the present invention should still fall within the scope covered by the patent of the present invention.

Claims

1. A production and processing system for enameled wire, comprising a wire pay-off device, an annealing device, a painting device, a drying device, a cooling device, a wire take-up device, a detection device and an exhaust gas treatment device, characterized in that, A seal monitoring system is provided in the waste gas treatment device, and the seal monitoring system includes: Fault model construction module: Obtain the historical monitoring data and historical fault data of the seal monitoring points, and construct a fault prediction model; Monitoring point rationality judgment module: Based on the constructed fault prediction model, analyze the fault warning data of the seal monitoring points, process to obtain a rational judgment value, and judge the rationality of the setting of the seal monitoring points; Monitoring point adjustment module: Based on the unreasonable setting of the seal monitoring points, analyze and obtain the adjustment priority level of the unreasonable monitoring.

2. The production and processing system of an enameled wire according to claim 1, characterized in that, The construction process of the fault prediction model: Obtain the historical monitoring data and historical fault data of all seal monitoring points, and use a neural network model to train the fault prediction model.

3. A production and processing system for enameled wire according to claim 1, characterized in that, The process of judging the rationality of the setting of the seal monitoring points is: Obtain the rational judgment value, and mark the seal monitoring points with a value greater than the rationality judgment threshold as unreasonable monitoring points.

4. The production and processing system of an enameled wire according to claim 1, characterized in that, The process of obtaining the rational judgment value is: Based on the fault prediction model, obtain the predicted fault occurrence time, and combine with the real-time fault data for data processing to obtain the proportion of the number of abnormal points, the proportion of the continuous number, the proportion of the number of unreasonable early warnings, and the proportion of invalid warnings; Perform weighted calculation on the proportion of the number of abnormal points, the proportion of the continuous number, the proportion of the number of unreasonable early warnings, and the proportion of invalid warnings to obtain the rational judgment value.

5. A production and processing system for enameled wire according to claim 4, characterized in that The process of obtaining the proportion of invalid warnings is: Preset an analysis period, and obtain real-time monitoring data and real-time fault data within the analysis period; Input the real-time data of each seal monitoring point into the trained fault prediction model to obtain the predicted fault data corresponding to each monitoring point; If the real-time fault occurrence time is earlier than or equal to the predicted fault occurrence time, mark it as an invalid warning; Obtain the number of invalid warnings, and calculate the ratio with the total number of faults to obtain the proportion of invalid warnings.

6. The production and processing system of an enameled wire according to claim 4, characterized in that, The process of obtaining the proportion of the number of unreasonable early warnings is: If the real-time fault occurrence time is earlier than or equal to the predicted fault occurrence time, and if the real-time fault occurrence time is later than the predicted fault occurrence time, mark it as a valid warning; Based on the valid warning, calculate the difference between the real-time fault occurrence time and the predicted fault occurrence time to obtain the warning time deviation value, and construct a warning time deviation sequence; Extract the warning time deviation values greater than the warning time deviation limit value, count their number, and calculate the ratio with the number of data points in the warning time deviation value data sequence to obtain the proportion of the number of unreasonable early warnings.

7. A production and processing system for enameled wires according to claim 4, characterized in that, The process of obtaining the proportion of the number of abnormal points and the proportion of the continuous number is: Obtain the anomaly detection score through the anomaly detection algorithm; If the anomaly detection score is greater than the anomaly threshold, mark the corresponding data point as an abnormal point; Count the number of abnormal points, and calculate the ratio with the number of data points in the warning time deviation value data sequence to obtain the proportion of the number of abnormal points; Integrate the continuous abnormal points to obtain an abnormal point combination, respectively count their continuous numbers and perform mean processing, and then calculate the ratio with the number of abnormal points to obtain the proportion of the continuous number.

8. The production and processing system of an enameled wire according to claim 7, characterized in that, The process of obtaining the anomaly detection score is: The anomaly detection algorithm includes the isolation forest algorithm; Analyze the early warning time deviation value sequence through the Isolation Forest algorithm to obtain the anomaly detection score of each data point.

9. The production and processing system of an enameled wire according to claim 1, characterized in that, The process of obtaining the adjustment priority level of unreasonable monitoring is as follows: Obtain all unreasonable monitoring points and obtain the corresponding positions and rationality judgment values; Determine the scope of the key area, measure the distance values from the unreasonable monitoring points to the center of each key area respectively, and perform a weighted sum calculation on all the distance values to obtain the position influence value; Cluster the rationality judgment values and position influence values of each monitoring point through a clustering algorithm; According to the clustering results, divide the unreasonable monitoring points in different clusters into different adjustment levels.

10. The production and processing system of an enameled wire according to claim 9, characterized in that, The process of dividing into different adjustment levels is as follows: For each cluster after clustering, calculate their quartiles for the rationality judgment values and position influence values of each cluster after clustering. For the reasonable judgment values, they are respectively , , . For the position influence values, they are respectively , , ; For each cluster, if the reasonable judgment value ≥ and the position influence value ≥ , then this cluster is classified as a high-priority adjustment class; if the reasonable judgment value < and the position influence value < , then this cluster is classified as a low-priority adjustment class; otherwise, this cluster is classified as a medium-priority adjustment class.

Images