A liquid cooling plate production parameter management method and system based on the Internet of Things

Abstract

The present invention relates to the field of data processing technology, and more specifically, to a method and system for managing production parameters of a liquid cooling plate based on the Internet of Things, the method comprising: obtaining sliding data of each stamping slide block at each sampling moment during multiple complete stamping processes of the liquid cooling plate to construct sliding data points at each sampling moment; performing COF detection on the current sliding data point, and during the detection process, using the average noise level of the current sliding data point and the previous sliding data points, correcting the trajectory cost of the corresponding two sliding data points, the correction value is negatively correlated with the average noise level, and based on the corrected trajectory cost, obtaining the abnormal detection result of the COF detection of the current sliding data point to adjust the production parameters of the liquid cooling plate. The present invention can improve the accuracy of the abnormal detection result and realize the accurate monitoring of the abnormality of the stamping equipment.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology. More specifically, the present invention relates to a method and system for managing production parameters of a liquid cooling plate based on the Internet of Things. Background Art

[0002] As an efficient heat dissipation device, liquid cooling plates have been put into practical use in electronic equipment, new energy vehicles and other fields. The production process usually adopts a stamping process, which requires multiple stamping sliders to cooperate with each other to achieve forming. However, during the stamping process, if the parameters of the stamping slider are not set properly, it may not only lead to forming failure, but also damage the stamping equipment. Therefore, in the stamping process of liquid cooling plates, the management of the parameters of each stamping slider is very important, especially the detection of abnormal data.

[0003] The COF (Connectivity-Based Outlier Factor) algorithm is an anomaly detection method, the core of which is to evaluate the degree of anomaly by calculating the local connectivity of data points. The specific process is: for each point in the data set, first determine its k nearest neighbors, then build the shortest path based on these nearest neighbors, and calculate the local average link distance of the point. Next, calculate the COF value of the point, that is, the ratio of its local average link distance to the average of the local average link distances of all points in the neighborhood. The higher the COF value, the greater the difference in connectivity between the point and the neighborhood, and the more likely it is an anomaly. By considering the connectivity between data points, the algorithm effectively improves the accuracy of anomaly detection, especially for scenarios with low-density and linearly correlated data, and overcomes the defect that the LOF algorithm cannot effectively measure sequence data and low-density data objects.

[0004] However, since the COF algorithm itself is sensitive to noise and is easily disturbed during the link construction process, both the noise data and the data with actual anomalies have large COF values, resulting in the abnormal data detected by the algorithm being likely to be noise data, thereby affecting the accuracy of the anomaly detection results and making it impossible to accurately monitor abnormal situations based on the detection results. Summary of the invention

[0005] In order to solve the problem that when using the COF algorithm to detect abnormal data in the production parameters of liquid cooling plates, the abnormal data cannot be accurately identified due to the interference of noise data, the present invention provides a liquid cooling plate production parameter management method and system based on the Internet of Things.

[0006] According to a first aspect of the present invention, a method for managing production parameters of a liquid cooling plate based on the Internet of Things is provided, comprising:

[0007] In multiple complete stamping processes of the liquid cooling plate, the sliding data of each stamping slider at each sampling time is obtained to construct the sliding data points at each sampling time, and the sliding data points include the sampling order, the sliding progress of each stamping slider at the corresponding sampling time, and the sliding speed;

[0008] Perform COF detection on the current sliding data point. During the detection process, use the average noise level of the current sliding data point and the previous sliding data points to correct the trajectory cost of the corresponding two sliding data points. The correction value is negatively correlated with the average noise level. Based on the corrected trajectory cost, obtain the abnormal detection result of COF detection on the current sliding data point to adjust the production parameters of the liquid cooling plate.

[0009] The method for obtaining the noise level of any sliding data point includes:

[0010] The sliding data points with the same sampling order are regarded as the same-sequence data points, and the sliding data points with sampling order differences less than a preset value are regarded as the same-window data points; the difference between any sliding data point and the same-sequence data points and the same-window data points is calculated to obtain a first difference and a second difference, and the first difference is corrected by using the second difference through gamma transformation to obtain the noise level of any sliding data point.

[0011] When the COF algorithm is used to detect abnormal data, the present invention can avoid noise data having a large COF value, so that the abnormal data detected by the algorithm is the data of the actual abnormality of the stamping equipment, reducing the interference of noise data on the abnormality detection using the algorithm, ensuring the accuracy of the abnormality detection results, and realizing accurate monitoring of abnormal conditions of the stamping equipment based on the abnormality detection results with higher accuracy.

[0012] Preferably, the method for obtaining the first difference includes:

[0013] Calculate the difference between the average value of each data in any sliding data point except the sampling order and the corresponding data in the data point with the same order, and take the maximum value, which is used as the first difference corresponding to any sliding data point and the data point with the same order.

[0014] The present invention can accurately measure the possibility of data with abnormal values ​​existing in each sliding data point.

[0015] Preferably, the method for obtaining the second difference includes:

[0016] Calculate the difference between the average value of each data in any sliding data point except the sampling order and the corresponding data in the data point in the same window, and take the average value to obtain the second difference corresponding to the any sliding data point and the data point in the same window.

[0017] The present invention can accurately evaluate the possibility of the existence of data with continuous abnormal values ​​in each sliding data point, thereby effectively distinguishing real abnormal data from noise data.

[0018] Preferably, the noise level of any sliding data point satisfies the following relationship:

[0019] ;

[0020] In the formula, is the noise level of any sliding data point; , are the first Data and individual data; is the data point with the same sequence as any sliding data point. The average value of the data; is the data point in the same window of any sliding data point. The average value of the data; is the number of stamping slides; is the absolute value symbol; A function that returns the maximum value.

[0021] The present invention can accurately measure the possibility that each sliding data point is noise data.

[0022] Preferably, the average noise level of the current sliding data point and the previous sliding data points is used to correct the trajectory cost of the corresponding two sliding data points to satisfy the following relationship:

[0023] ;

[0024] In the formula, , are the current sliding data point and the previous The trajectory cost of the sliding data point before and after correction; The current sliding data point and the The average noise level of sliding data points; is a natural exponential function.

[0025] The present invention can increase the distance between sliding data points with a larger noise level, thereby preventing the sliding data points with a larger noise level from being linked to the local shortest path of the current sliding data point by the COF algorithm, thereby reducing the impact of noise data on the anomaly detection result.

[0026] Preferably, using the average noise level of the current sliding data point and the previous sliding data points to correct the trajectory cost corresponding to the two sliding data points also includes:

[0027] If the average noise level of the current sliding data point and any previous sliding data point is less than the preset level threshold, the trajectory costs of the two sliding data points before and after correction are the same; if it is greater than or equal to the level threshold, the correction value of the trajectory cost of the two sliding data points satisfies the following relationship:

[0028] ;

[0029] In the formula, , are the current sliding data point and the previous The trajectory cost of the sliding data point before and after correction; The current sliding data point and the The average noise level of sliding data points; is a natural exponential function.

[0030] Preferably, based on the corrected trajectory cost, obtaining the anomaly detection result of the current sliding data point includes:

[0031] Based on the corrected trajectory cost between the current sliding data point and each previous sliding data point, a local outlier factor based on connectivity of the current sliding data point is obtained, so as to determine the anomaly detection result of the current sliding data point according to the comparison result of the local outlier factor and a preset threshold.

[0032] Preferably, determining the abnormality detection result of the current sliding data point according to the comparison result of the local outlier factor and the preset threshold value includes:

[0033] If the local outlier factor is greater than the preset abnormal threshold, the current sliding data point is determined to be an abnormal data point; if it is less than or equal to the abnormal threshold, the current sliding data point is determined to be a normal data point.

[0034] Preferably, before constructing the sliding data points at each sampling moment, the method further includes:

[0035] The sliding speed of each punching slide at each sampling moment is normalized to construct a sliding data point at each sampling moment based on the normalized value of the sliding speed.

[0036] According to a second aspect of the present invention, a liquid cooling plate production parameter management system based on the Internet of Things is provided. The system includes a memory and a processor. A computer program is stored in the memory. The processor executes the computer program to implement the steps of the first aspect of the present invention.

[0037] The present invention has the following effects:

[0038] 1. The method of constructing sliding data points in the present invention can comprehensively reflect the coordinated state between the stamping slide blocks during the liquid cooling plate stamping from multiple dimensions, so that when the abnormal stamping equipment is identified based on the abnormal detection results of the sliding data points, the slight changes of the equipment can be captured, thereby ensuring the accuracy of the abnormal detection results.

[0039] 2. When the present invention uses the COF algorithm to perform anomaly detection on sliding data points, the trajectory cost between the corresponding two data points is corrected by the average noise level between the sliding data points, which can avoid the noise data having a large COF value, thereby causing the noise data to be detected as abnormal data. In addition, when determining the noise level of each sliding data point, the present invention can effectively distinguish the interference of real abnormal data on noise data identification, thereby ensuring the accuracy of the determined noise level, and further ensuring the accuracy of the anomaly detection result. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] By reading the following detailed description with reference to the accompanying drawings, the above and other objects, features and advantages of the exemplary embodiments of the present invention will become readily understood. In the accompanying drawings, several embodiments of the present invention are shown in an exemplary and non-limiting manner, and the same or corresponding reference numerals represent the same or corresponding parts, wherein:

[0041] Figure 1 The present invention is a flowchart of a method for managing production parameters of a liquid cooling plate based on the Internet of Things. DETAILED DESCRIPTION

[0042] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0043] The specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0044] Reference Figure 1 , a method for managing production parameters of a liquid cooling plate based on the Internet of Things, including steps S1 to S3, specifically as follows:

[0045] S1: In multiple complete stamping processes of the liquid cooling plate, the sliding data of each stamping slider at each sampling moment is obtained to construct the sliding data points at each sampling moment. The sliding data points include the sampling order, the sliding progress of each stamping slider at the corresponding sampling moment, and the sliding speed.

[0046] It should be noted that, due to the complex structure of the liquid cooling plate, multiple stamping sliders need to cooperate with each other during the stamping process. If any stamping slider is abnormal, the produced liquid cooling plate will be defective. Therefore, the present invention collects the sliding data of all stamping sliders.

[0047] In an exemplary embodiment of the present invention, the sampling order refers to the numbering of each sampling moment in a complete stamping process in chronological order. For example, the sampling order of the first sampling moment in each complete stamping process is 1, the sampling order of the second sampling moment is 2, and so on, so that the sampling order of each sampling moment in each complete stamping process can be obtained. This embodiment does not specifically limit the number of selected complete stamping processes.

[0048] Sliding progress refers to the ratio of the displacement of any stamping slider as of any sampling moment to the total displacement of the stamping slider in the complete stamping process. For example, when the total displacement of any stamping slider in the complete stamping process is 100, and the displacement of the stamping slider as of a certain sampling moment is 50, then the sliding progress of the stamping slider at the corresponding sampling moment is .

[0049] In an exemplary embodiment of the present invention, before constructing the sliding data points at each sampling moment, the following steps are also included:

[0050] The sliding speed of each punching slide at each sampling moment is normalized to construct a sliding data point at each sampling moment based on the normalized value of the sliding speed.

[0051] Alternatively, a linear normalization function can be used, such as The function normalizes the sliding speed of each stamping slider at each sampling moment, and the sliding speed of the corresponding stamping slider can also be normalized by using the cumulative sum of the sliding speeds of each stamping slider at all sampling moments in a complete stamping process. This embodiment does not specifically limit the selected normalization method.

[0052] Next, the process of constructing the sliding data points at each sampling moment is described in detail:

[0053] First, the first The sampling order of the sampling moments is recorded as ; The number of punch slides is recorded as ; The sliding progress of each stamping slide at the sampling time is recorded as ,in, , as well as The first punching slide, the second punching slide and the The sliding progress of each stamping slider at the sampling moment; the normalized values ​​of the sliding speed of each stamping slider at the sampling moment are recorded as ,in, , as well as The first punching slide, the second punching slide and the The normalized value of the sliding velocity of the punch slider at the sampling moment.

[0054] Then, the sampling order , the sliding progress of each stamping slide at the sampling time And the normalized value of the sliding speed of each punch slider at the sampling time The sequence composed of , recorded as the first The sliding data points at each sampling moment are obtained, thereby obtaining the sliding data points at each sampling moment in each complete stamping process. This embodiment does not specifically limit the arrangement of each data in the sliding data point.

[0055] S2: Perform COF detection on the current sliding data point. During the detection process, the average noise level of the current sliding data point and the previous sliding data points is used to correct the trajectory cost of the corresponding two sliding data points. The correction value is negatively correlated with the average noise level.

[0056] It should be noted that when using the COF (Connectivity-Based Outlier Factor) detection algorithm for anomaly detection, the presence of noise data will interfere with the neighborhood search process of the algorithm, causing the noise data to have a larger COF value, thereby affecting the detection results. Therefore, the present invention improves the traditional COF algorithm, and the specific improvements are as follows: using the average noise level of the current sliding data point and any previous sliding data point, the trajectory cost between the corresponding two data points is corrected, and the COF value of the current sliding data point is determined based on the corrected value of the trajectory cost between the current sliding data point and each previous sliding data point, thereby avoiding the noise data having a large COF value, ensuring the accuracy of the anomaly detection results, and realizing accurate monitoring of abnormal conditions of stamping equipment. Among them, trajectory cost is a professional term in the COF algorithm, which can be understood as the Euclidean distance between data points.

[0057] Among them, the present invention only corrects the trajectory cost between each data point in the COF algorithm, and does not change the calculation method of other data in the algorithm.

[0058] Specifically, the noise level of any sliding data point can be determined by the following steps:

[0059] Step 1: The sliding data points with the same sampling order are regarded as the same-order data points, and the sliding data points with sampling order differences less than a preset value are regarded as the same-window data points. The difference between any sliding data point and the same-order data points and the same-window data points is calculated to obtain the first difference and the second difference.

[0060] It should be noted that noise data usually appears as random, short-term deviations and has mutation, that is, the data fluctuations have no obvious time correlation, while the data when the stamping slider is abnormal often appears as continuous and significant deviations from normal values, and when the stamping slider is abnormal, the data obtained by different sensors may show abnormal changes at the same time. Therefore, the present invention uses this feature to distinguish between noise data and abnormal data by evaluating the performance of each data in each sliding data point, thereby determining the noise level of each sliding data point. Among them, the same sequence data points and the same window data points of any sliding data point are in different complete stamping processes with any sliding data point.

[0061] The first difference refers to the data that can reflect the degree of deviation between each sliding data point and the data points of the same sequence, which is used to evaluate the possibility of abnormal values ​​in each sliding data point. data points, and the When the overall level of the data is different, it means that the first data, and the There is a significant deviation from the normal level of the data, which indicates that the first The value of a data point is likely to be abnormal, and the corresponding sliding data point is likely to be noise data or abnormal data.

[0062] The second difference refers to the data that can reflect the degree of deviation between each sliding data point and the data points in the same window. It is used to evaluate the possibility that each data in each sliding data point continuously deviates from the normal level, so as to distinguish abnormal data from noise data. data, and the first data point in the same window When the overall level of the data is different, it means that the first data, within a certain continuous range, There is a large deviation from the normal level of the data, that is, the first There is a high possibility that the data continuously deviates from the normal level. At this time, if there is data with abnormal values ​​in the sliding data point, the cause of the abnormality is likely to be an abnormality in the stamping equipment, so that the cause of the abnormality in the sliding data point can be effectively distinguished.

[0063] Optionally, the preset value may be set to 7, so that the same-window data points of each sliding data point may be determined based on the preset value. This embodiment does not specifically limit the size of the preset value.

[0064] In an exemplary embodiment of the present invention, the determination of the first difference may be achieved by the following steps:

[0065] Calculate the difference between each data in any sliding data point and the average value of the corresponding data in the data point with the same sequence, take the maximum value, and use the maximum value as the first difference corresponding to any sliding data point and the data point with the same sequence.

[0066] Optionally, the absolute value of the difference between any data in any sliding data point and the average value of corresponding data in the data point with the same sequence can be used as the difference between the any data and the average value, or the square difference between the any data and the average value can be used as the difference between the any data and the average value. This embodiment does not specifically limit the method for determining the difference between data.

[0067] In another embodiment, the maximum value of the difference between each data in any sliding data and the median of the corresponding data in the same-sequence data point may be used as the first difference between the any sliding data and the same-sequence data point.

[0068] In an exemplary embodiment of the present invention, the determination of the second difference may be achieved by the following steps:

[0069] Calculate the difference between the average value of each data in any sliding data point except the sampling order and the corresponding data in the data point in the same window, and take the average value to obtain the second difference corresponding to the any sliding data point and the data point in the same window.

[0070] It should be noted that since the sampling order is irrelevant to the abnormal situation of the data, there is no need to consider the influence of the sampling order when calculating the first difference between the sliding data point and the data point with the same sequence; and when calculating the second difference between the sliding data point and the data point with the same window, in order to maintain the consistency of the data dimension, the present invention does not consider the influence of the sampling order.

[0071] In another embodiment, the average difference between each data in any sliding data and the median of the corresponding data in the data point in the same window may also be used as the second difference between the any sliding data and the data point in the same sequence.

[0072] Step 2: Using the second difference to correct the first difference through gamma transformation, the noise level of any sliding data point is obtained.

[0073] Specifically, the noise level of any sliding data point satisfies the following relationship:

[0074] ;

[0075] In the formula, is the noise level of any sliding data point; , are the first Data and individual data; is the data point with the same sequence as any sliding data point. The average value of the data; is the data point in the same window of any sliding data point. The average value of the data; is the number of stamping slides; is the absolute value symbol; A function that returns the maximum value.

[0076] in, It reflects the first difference between any sliding data point and the corresponding data point of the same sequence. The larger the value is, the more likely it is that the sliding data point has abnormal values, and the corresponding noise level of the sliding data point is relatively large.

[0077] It reflects the second difference between any sliding data point and the data point in the same window. The larger the value is, the more likely it is that the sliding data point contains abnormal data that continuously deviates from the normal level, which further indicates that the cause of the abnormal value of the sliding data point is more likely to be an abnormality in the stamping equipment rather than noise data. Therefore, when using the first difference between the sliding data point and the data point in the same sequence to determine the noise level of the sliding data point, it is necessary to make a large downward correction to the noise level to eliminate the interference of abnormal data on noise data identification when an abnormality occurs in the stamping equipment, thereby ensuring the accuracy of the noise level determination.

[0078] It should be noted that due to The value is between 0 and 1. The value of is also within 0-1, so that when using ,right When making corrections, The larger the value, the more A larger degree of downward correction is performed, thereby reducing the interference of abnormal data generated when the stamping equipment is abnormal on the recognition of noise data.

[0079] Furthermore, after determining the noise level of each sliding data point, the average noise level of the current sliding data point and the previous sliding data points can be used to correct the trajectory cost of the corresponding two sliding data points, thereby reducing the impact of noise data on the anomaly detection results determined based on the COF algorithm.

[0080] Specifically, the corrected trajectory cost of the current sliding data point and any previous sliding data point satisfies the following relationship:

[0081] ;

[0082] In the formula, , are the current sliding data point and the previous The trajectory cost of the sliding data point before and after correction; The current sliding data point and the The average noise level of sliding data points; is a natural exponential function, where the natural exponential function refers to a function with a natural constant An exponential function with base .

[0083] Optional, when When it is larger, it means that the current sliding data point is different from the previous There is a high possibility of noise data in the sliding data points. By reducing the trajectory cost corresponding to the two sliding data points, it is possible to avoid the noise data having a large COF value, so that when the COF value is used to screen abnormal sliding data points, the interference of noise data on abnormal data detection can be avoided.

[0084] In another exemplary embodiment of the present invention, the correction of the trajectory cost of the current sliding data point and any previous sliding data point can also be achieved by the following steps:

[0085] If the average noise level of the current sliding data point and any previous sliding data point is less than the preset level threshold, the trajectory costs of the two sliding data points before and after correction are the same; if it is greater than or equal to the level threshold, the correction value of the trajectory cost of the two sliding data points satisfies the following relationship:

[0086] ;

[0087] In the formula, , are the current sliding data point and the previous The trajectory cost of the sliding data point before and after correction; The current sliding data point and the The average noise level of sliding data points; is a natural exponential function.

[0088] Optionally, the preset degree threshold can be set to 0.8. If the average noise degree of the current sliding data point and any previous sliding data point is less than 0.8, it means that it is less likely that noise data exists in the corresponding two sliding data points. At this time, the connectivity between the corresponding two sliding data points is mainly affected by the real data distribution, and there is no need to make too many adjustments to the trajectory cost between the corresponding two sliding data points, so that the real connection relationship between the original values ​​can be retained. This embodiment does not specifically limit the size of the degree threshold.

[0089] S3: Based on the corrected trajectory cost, obtain an abnormal detection result of COF detection on the current sliding data point to adjust the production parameters of the liquid cooling plate.

[0090] In an exemplary embodiment of the present invention, the abnormality detection result of the current sliding data point can be determined by the following steps:

[0091] Based on the corrected trajectory cost between the current sliding data point and each previous sliding data point, a local outlier factor based on connectivity of the current sliding data point is obtained, so as to determine the anomaly detection result of the current sliding data point according to the comparison result of the local outlier factor and a preset threshold.

[0092] It should be noted that the process of determining the COF value based on the connectivity of the corresponding data point based on the distance between the data points is a prior art and is not described in detail in this embodiment. In this embodiment, when obtaining the COF value of the current sliding data point, the k in the k distance neighborhood is selected as 30. This embodiment does not specifically limit the value of k.

[0093] Optionally, the preset threshold may be set to 0.7, so that the abnormality of the current sliding data point is determined according to the comparison result of the COF value of the current sliding data point with the preset threshold. This embodiment does not specifically limit the value of the preset threshold.

[0094] In an exemplary embodiment of the present invention, the determination of the current sliding data point can be achieved by the following steps:

[0095] If the local outlier factor is greater than the preset abnormal threshold, the current sliding data point is determined to be an abnormal data point; if it is less than or equal to the abnormal threshold, the current sliding data point is determined to be a normal data point.

[0096] Furthermore, when the current sliding data point is determined to be an abnormal data point, an alarm is issued to promptly remind the operator that there is an abnormality in the stamping equipment at the current moment, thereby enabling real-time monitoring of the abnormality of the stamping equipment based on the abnormality of the production parameters of the liquid cooling plate.

[0097] The present invention also provides a liquid cooling plate production parameter management method system based on the Internet of Things. The system includes a memory and a processor, and a computer program is stored in the memory. The computer program integrates the functions of a liquid cooling plate production parameter management method based on the Internet of Things. When the computer program is executed, a liquid cooling plate production parameter management method based on the Internet of Things can improve the accuracy of abnormality detection results and achieve accurate monitoring of stamping equipment abnormalities.

[0098] In the description of this specification, "plurality" or "several" means at least two, such as two, three or more, etc., unless otherwise clearly and specifically defined.

[0099] Although this specification has shown and described a number of embodiments of the present invention, it will be apparent to those skilled in the art that such embodiments are provided by way of example only. Those skilled in the art will conceive of many modifications, changes and alternatives without departing from the ideas and spirit of the present invention. It should be understood that in the practice of the present invention, various alternatives to the embodiments of the present invention described herein may be employed.

Claims

1. A method for managing production parameters of liquid cooling plates based on the Internet of Things. It is characterized in that include: In multiple complete stamping processes of the liquid cooling plate, the sliding data of each stamping slider at each sampling time is obtained to construct the sliding data points at each sampling time, wherein the sliding data points include the sampling order, the sliding progress and the sliding speed of each stamping slider at the corresponding sampling time; Perform COF detection on the current sliding data point. During the detection process, use the average noise level of the current sliding data point and the previous sliding data points to correct the trajectory cost of the corresponding two sliding data points. The correction value is negatively correlated with the average noise level. Based on the corrected trajectory cost, obtain the abnormal detection result of COF detection on the current sliding data point to adjust the production parameters of the liquid cooling plate. The method for obtaining the noise level of any sliding data point includes: The sliding data points with the same sampling order are regarded as the same-order data points, and the sliding data points with sampling order differences less than the preset value are regarded as the same-window data points; Calculate the difference between any sliding data point and the data points in the same sequence and the data points in the same window to obtain a first difference and a second difference, and use the second difference to correct the first difference through gamma transformation to obtain the noise level of any sliding data point; The calculation formula for the noise level of any sliding data point is: ； In the formula, is the noise level of any sliding data point; , are the first Data and individual data; is the data point with the same sequence as any sliding data point. The average value of the data; is the data point in the same window of any sliding data point. The average value of the data; is the number of stamping slides; is the absolute value symbol; A function that returns the maximum value.

2. According to the method for managing production parameters of liquid cooling plates based on the Internet of Things according to claim 1, It is characterized in that The method for obtaining the first difference includes: Calculate the difference between the average value of each data in any sliding data point except the sampling order and the corresponding data in the data point with the same order, and take the maximum value, and use the maximum value as the first difference corresponding to any sliding data point and the data point with the same order.

3. According to the method for managing production parameters of liquid cooling plates based on the Internet of Things in claim 1, It is characterized in that The method for obtaining the second difference includes: Calculate the difference between the average value of each data in any sliding data point except the sampling order and the corresponding data in the data point in the same window, and take the average value to obtain the second difference corresponding to the any sliding data point and the data point in the same window.

4. According to the method for managing production parameters of liquid cooling plates based on the Internet of Things in claim 1, It is characterized in that The average noise level of the current sliding data point and the previous sliding data points is used to correct the trajectory cost of the corresponding two sliding data points to satisfy the following relationship: ； In the formula, , are the current sliding data point and the previous The trajectory cost of the sliding data point before and after correction; The current sliding data point and the The average noise level of sliding data points; is a natural exponential function.

5. According to the method for managing production parameters of liquid cooling plates based on the Internet of Things in claim 1, It is characterized in that The method of using the average noise level of the current sliding data point and the previous sliding data points to correct the trajectory cost of the corresponding two sliding data points also includes: If the average noise level of the current sliding data point and any previous sliding data point is less than the preset level threshold, the trajectory costs of the two sliding data points before and after correction are the same; if it is greater than or equal to the level threshold, the correction value of the trajectory cost of the two sliding data points satisfies the following relationship: ； In the formula, , are the current sliding data point and the previous The trajectory cost of the sliding data point before and after correction; The current sliding data point and the The average noise level of sliding data points; is a natural exponential function.

6. According to the method for managing production parameters of liquid cooling plates based on the Internet of Things in claim 1, It is characterized in that The method of obtaining the abnormality detection result of the current sliding data point based on the corrected trajectory cost includes: Based on the corrected trajectory cost between the current sliding data point and each previous sliding data point, a local outlier factor based on connectivity of the current sliding data point is obtained, so as to determine the abnormality detection result of the current sliding data point according to the comparison result of the local outlier factor and a preset threshold.

7. According to claim 6, a method for managing production parameters of liquid cooling plates based on the Internet of Things, It is characterized in that The determining, according to the comparison result of the local outlier factor and the preset threshold, the abnormality detection result of the current sliding data point includes: If the local outlier factor is greater than a preset abnormal threshold, the current sliding data point is determined to be an abnormal data point; if it is less than or equal to the abnormal threshold, the current sliding data point is determined to be a normal data point.

8. According to the method for managing production parameters of liquid cooling plates based on the Internet of Things in claim 1, It is characterized in that Before constructing the sliding data points at each sampling moment, the method further includes: The sliding speed of each punching slide at each sampling moment is normalized to construct a sliding data point at each sampling moment based on the normalized value of the sliding speed.

9. A liquid cooling plate production parameter management system based on the Internet of Things, It is characterized in that The liquid cooling plate production parameter management system based on the Internet of Things includes a memory and a processor, the memory stores a computer program, and the processor executes the computer program to implement the steps of the method as described in any one of claims 1 to 8.

Images