Coating equipment fault real-time early warning method

Abstract

The invention relates to the technical field of data processing, in particular to a coating equipment fault real-time early warning method, which comprises the following steps of: acquiring pressure data and vibration data, acquiring instability of each pressure data according to surrounding data of each pressure data, and determining whether each pressure data is unstable or not according to the instability of each pressure data. Obtaining a characteristic value of each piece of pressure data, obtaining the optimal degree of each clustering result according to the characteristic value of each piece of pressure data and the correlation of the abnormal degree of each piece of pressure data in each clustering result, obtaining the optimal number of midpoints, obtaining the optimal clustering result, and obtaining the optimal clustering result according to the optimal clustering result. The final anomaly degree of each piece of pressure data is obtained, then the fault of the coating equipment is judged, and the anomaly detection accuracy is improved.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and more specifically, to a method for real-time early warning of coating equipment failure. Background Art

[0002] In modern industrial production, improving production efficiency is one of the important goals pursued by enterprises. As a key equipment on the production line, the operating status of the coating equipment directly affects the efficiency of the entire production line. Therefore, a real-time early warning method is needed to promptly detect and handle equipment failures to reduce downtime and production losses. During the operation of the coating machine, the pressure between the coating roller and the support roller directly affects the uniform distribution of the coating and the coating quality. Therefore, the failure of the coating equipment can be effectively detected by performing real-time abnormal detection of the pressure data between the coating roller and the support roller.

[0003] The patent application document with the current publication number CN117473278A discloses a method for improving FMEA based on fuzzy analytic hierarchy process, fuzzy VIKOR method and PAM clustering, which improves the shortcomings in the three aspects of failure mode level evaluation, risk factor weight determination and failure mode importance ranking in traditional FMEA. First, triangular fuzzy numbers are used to make the evaluation of failure modes more in line with reality. Secondly, the weight of risk factors is improved by using comprehensive weights. Then, fuzzy VIKOR method is used to improve FMEA, and the comprehensive evaluation value of group benefit value and personal regret value is combined as the basis for sorting. Finally, PAM clustering algorithm is used to cluster failure modes according to failure mode RPN value and comprehensive evaluation value.

[0004] The PAM (Partitioning Around Medoid) algorithm is a data clustering method. By clustering pressure data, abnormal data in the data sequence can be effectively identified, and real-time early warning of coating equipment failure can be better completed. However, the number of midpoints in the PAM algorithm needs to be set manually. Since the pressure data between the coating roller and the support roller varies during the operation of the coating machine, if the number of midpoints is too small, the clustering result may be too general and cannot accurately reflect the internal structure of the pressure data; if the number of midpoints is too large, the clustering result may be too fragmented, increasing the difficulty of subsequent abnormality detection. Summary of the invention

[0005] In order to solve the problem that the number of midpoints set manually in the PAM algorithm is inaccurate due to the various changes in pressure data, the present invention proposes a real-time early warning method for coating equipment failure, which includes the following steps: Collect pressure data and vibration data; take the average of the absolute values ​​of the differences between the fitted values ​​and the actual values ​​of all surrounding data of each pressure data as the instability of each pressure data; take the product of the instability of each pressure data and the Pearson correlation coefficient of the instability sequence of each pressure data and the instability sequence of the vibration data corresponding to each pressure data as the characteristic value of each pressure data; The number of initial midpoints is preset as K, and the iteration step length of the number of midpoints is preset as 1. The pressure data is clustered according to the initial number of midpoints to obtain the first clustering result; the preference degree of the first clustering result is obtained according to the first clustering result and the characteristic value of each pressure data; K is added by one as the second number of midpoints, and the second clustering result is obtained, and the preference degree of the second clustering result is obtained according to the method for obtaining the preference degree of the first clustering result; according to the preference degree from the first clustering result to the second clustering result, it is judged whether to continue to iterate the number of initial midpoints; if the iteration is stopped, the second clustering result is the best clustering result; if the iteration is continued, it is stopped until the best clustering result is obtained; According to the best clustering result, the final abnormality of each pressure data is obtained, and then the failure of the coating equipment is judged.

[0006] The innovation of the present invention lies in obtaining the characteristic value of each pressure data according to the change characteristics of the pressure data and its surrounding data, which can reflect the abnormal situation of each pressure data, and then clustering the pressure data each time by iterating the number of midpoints of the PAM algorithm, and obtaining the preferred degree of each clustering result according to the characteristic value of each pressure data to obtain the best clustering result, and finally obtaining the final abnormality degree of each pressure data according to the best clustering result, and then judging the fault of the coating equipment, avoiding the problem of instability and uncertainty of clustering results caused by the artificial setting of the number of midpoints in traditional algorithms, which often relies on experience and subjective judgment, thereby improving the quality and accuracy of clustering, and then improving the efficiency of abnormality detection.

[0007] Preferably, the acquisition of all surrounding data of each pressure data includes: The number of neighborhood data is preset as M, and the i-th pressure data, the M pressure data before the i-th pressure data, and the M pressure data after the i-th pressure data are used as the surrounding data of the i-th pressure data.

[0008] It is convenient to obtain the characteristic value of each pressure data according to all surrounding data of each pressure data.

[0009] Preferably, obtaining the fitting values ​​of all surrounding data of each pressure data includes: The surrounding data of the ith pressure data are fitted using the least square method to obtain the fitting value of each surrounding data of the ith pressure data.

[0010] Preferably, the acquisition of the instability sequence of each pressure data and the instability sequence of vibration data corresponding to each pressure data comprises: Sequentially record the instability of each surrounding data of the i-th pressure data as the instability sequence of the i-th pressure data; The number of neighborhood data is preset as M, and the i-th vibration data, the M vibration data before the i-th vibration data, and the M vibration data after the i-th vibration data are taken as the surrounding data of the i-th vibration data; the instability of each surrounding data of the vibration data corresponding to the i-th pressure data is sequentially recorded as the instability sequence of the vibration data corresponding to the i-th pressure data.

[0011] Preferably, obtaining the preference degree of the first clustering result according to the first clustering result and the characteristic value of each pressure data includes: In the first clustering result, the LOF algorithm is used to obtain the local outlier factor of each pressure data in its clustering cluster, which is recorded as the abnormality of each pressure data; according to the order of the sampling time of the pressure data, the abnormality of each pressure data is sorted to obtain an abnormality sequence, and the eigenvalue of each pressure data is sorted to obtain a eigenvalue sequence; the Pearson correlation coefficient between the abnormality sequence and the eigenvalue sequence is used as the preferred degree of the first clustering result.

[0012] This facilitates the subsequent acquisition of the best clustering result based on the degree of optimization of each clustering result.

[0013] Preferably, judging whether to continue iterating the number of initial midpoints according to the preference of the first clustering result and the second clustering result includes: The least squares method is used to fit the preferred degree of the first and second clustering results to obtain a fitting curve. If the derivative value of the fitting curve is negative, the iteration of the initial midpoint number is stopped. If the derivative value of the fitting curve is positive, the iteration of the initial midpoint number continues.

[0014] According to the decreasing trend of the preference degree of each clustering result, the iteration of the number of midpoints is stopped, and accurate clustering results can be obtained.

[0015] Preferably, the method of obtaining the final abnormality of each pressure data according to the best clustering result, and then judging the failure of the coating equipment, includes: In the best clustering result, the LOF algorithm is used to obtain the local outlier factor of each pressure data in its cluster to which it belongs, and recorded as the final abnormality of each pressure data. If the final abnormality of any pressure data is greater than the abnormal threshold T, the pressure data is abnormal data. At this time, the coating equipment fails, and the remote monitoring and alarm of the coating equipment is immediately triggered to notify the relevant personnel in time for maintenance.

[0016] The quality and accuracy of clustering are improved, which in turn improves the accuracy of anomaly detection.

[0017] Preferably, the collecting of pressure data and vibration data includes: The pressure sensor is placed between the coating roller and the support roller of the coating machine, and the vibration sensor is placed on the coating roller. The preset sampling time is 2 times / second. During the operation of the coating machine, the pressure data and vibration data are collected.

[0018] The present invention has the following beneficial effects: the purpose of the present invention is to obtain the characteristic value of pressure data according to the change characteristics of pressure data and its local data, so as to reflect the abnormal situation of pressure data, and then cluster the pressure data each time by iterating the number of midpoints of PAM algorithm, and obtain the preferred degree of each clustering result according to the characteristic value of each pressure data and the correlation of the abnormality of each pressure data in each clustering result, and obtain the best clustering result, and finally obtain the final abnormality of each pressure data according to the best clustering result, so as to judge the fault of coating equipment, and avoid the problem of instability and uncertainty of clustering results caused by artificially setting the number of midpoints in traditional algorithms, which often relies on experience and subjective judgment, thereby improving the quality and accuracy of clustering, and thus improving the efficiency of abnormality detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] 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: Figure 1 The present invention is a flowchart of the steps of a method for real-time early warning of coating equipment failure according to an embodiment of the present invention. DETAILED DESCRIPTION

[0020] 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.

[0021] The specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0022] See also Figure 1 , which shows a flowchart of a method for real-time early warning of coating equipment failure provided by an embodiment of the present invention, the method comprising the following steps: S001. Collect pressure data and vibration data.

[0023] In an embodiment of the present invention, a pressure sensor is placed between the coating roller and the support roller of the coating machine, and a vibration sensor is placed on the coating roller. The preset sampling time is 2 times / second. During the operation of the coating machine, pressure data and vibration data are collected. It should be noted that the pressure data and the vibration data are time-aligned.

[0024] S002. Obtain the instability of each pressure data according to the surrounding data of each pressure data, and obtain the characteristic value of each pressure data according to the instability of each pressure data.

[0025] It should be noted that the number of midpoints in the existing PAM algorithm needs to be set manually, but the pressure data between the coating roller and the support roller during the operation of the coater varies, and it is difficult to determine the optimal number of midpoints by manual setting. If the number of midpoints is too small, the clustering results may be too general and unable to accurately reflect the internal structure of the pressure data; if the number of midpoints is too large, the clustering results may be too fragmented, increasing the difficulty of subsequent anomaly detection. Therefore, the present invention needs to iteratively and adaptively determine the number of midpoints, so as to obtain the optimal clustering result of the pressure data between the coating roller and the support roller during the operation of the coater, thereby improving the anomaly detection results and realizing real-time early warning of coating equipment failures.

[0026] It needs to be further explained that, when the coating machine fails (for example, the bearing of the coating roller is not lubricated enough or other failures occur), the pressure data between the coating roller and the support roller will fluctuate unstably. Therefore, the instability of each pressure data can be obtained first, which is represented by the mean value of the fitting error of all the surrounding data of the pressure data by curve fitting. Because curve fitting is a process of finding a curve so that the pressure data is as close to the curve as possible, the more stable the pressure data change, the more conducive to the curve fitting. Therefore, the smaller the mean value of the fitting error of all the surrounding data of the pressure data, the more stable the pressure data is, and vice versa. In addition, since the pressure sensor will be affected by the surrounding noise when collecting pressure data, the pressure data will also experience unstable fluctuations. However, the instability of the pressure data caused by some faults of the coating machine is because the vibration of the coating roller causes the pressure data to fluctuate, that is, the unstable fluctuation of the vibration data will cause the unstable fluctuation of the pressure data. Therefore, the correlation between the instability of the pressure data and its corresponding vibration data can be used as the credibility to correct the instability of each pressure data. The greater the correlation, the greater the credibility of the instability of each pressure data. Therefore, the characteristic value of each pressure data is obtained according to the instability of each pressure data and the correlation between the instability of each pressure data and its corresponding vibration data.

[0027] In an embodiment of the present invention, the i-th pressure data, the M pressure data before the i-th pressure data, and the M pressure data after the i-th pressure data are used as the surrounding data of the i-th pressure data; the i-th vibration data, the M vibration data before the i-th vibration data, and the M vibration data after the i-th vibration data are used as the surrounding data of the i-th vibration data; the preset number of neighborhood data M=7, in other embodiments, the implementer may preset the value of M according to the specific implementation method.

[0028] Use the least square method to fit the surrounding data of the i-th pressure data to obtain the fitting value of each surrounding data of the i-th pressure data; Get the instability of each pressure data: ; In the formula, represents the instability of the i-th pressure data; Represents the number of surrounding data of the i-th pressure data; represents the value of the jth surrounding data of the i-th pressure data; represents the fitting value of the jth surrounding data of the ith pressure data; || represents the absolute value symbol; if the difference between the fitting value of each surrounding data of the ith pressure data and the actual value is larger, it means that the ith pressure data is more unstable.

[0029] The instability of each surrounding data of the i-th pressure data is recorded in sequence as the instability sequence of the i-th pressure data, and the instability of each surrounding data of the vibration data corresponding to the i-th pressure data is recorded in sequence as the instability sequence of the vibration data corresponding to the i-th pressure data; it should be noted that the vibration data corresponding to the sampling moment of the pressure data is the vibration data corresponding to the pressure data.

[0030] Get the characteristic value of each pressure data: ; In the formula, Represents the characteristic value of the i-th pressure data; represents the instability of the i-th pressure data; norm() represents the normalization function; The Pearson correlation coefficient representing the instability sequence of the i-th pressure data and the instability sequence of the vibration data corresponding to the i-th pressure data, the larger the value, the more credible the instability of the i-th pressure data. The larger the value of is, the more abnormal the i-th pressure data is.

[0031] S003. According to the characteristic value of each pressure data and the correlation of the abnormality of each pressure data in each clustering result, the preferred degree of each clustering result is obtained, and then the optimal number of midpoints is obtained, and then the optimal clustering result is obtained.

[0032] It should be noted that the number of midpoints in the known existing PAM algorithm needs to be set manually, and the pressure data between the coating roller and the support roller during the operation of the coating machine varies. It is difficult to determine the optimal number of midpoints by manual setting methods. Therefore, the present invention iterates the number of midpoints and obtains each clustering result. According to any clustering result, the abnormality of each pressure data in any clustering result is obtained. Since the characteristic value of each pressure data reflects the abnormal situation of each pressure data, the more the abnormality of each pressure data in the clustering result is correlated with the characteristic value of each pressure data, the more accurate the clustering result is. Therefore, according to the correlation between the abnormality of each pressure data in each clustering result and the characteristic value of each pressure data, the preferred degree of each clustering result is obtained. Finally, when the priority of the clustering result in the iterative process begins to decrease, the iteration is stopped to obtain the optimal number of midpoints.

[0033] In an embodiment of the present invention, the number of initial midpoints is preset to K. According to the number of initial midpoints, the pressure data is clustered using the PAM algorithm to obtain the first clustering result; in the first clustering result, the LOF algorithm is used to obtain the local outlier factor of each pressure data in the cluster to which it belongs, which is recorded as the abnormality of each pressure data; according to the order of the sampling time of the pressure data, the abnormality of each pressure data is sorted to obtain an abnormality sequence, and the eigenvalue of each pressure data is sorted to obtain a eigenvalue sequence; the Pearson correlation coefficient between the abnormality sequence and the eigenvalue sequence is used as the preferred degree of the first clustering result.

[0034] The iteration step of the number of midpoints is preset to 1, and K plus one is used as the number of midpoints for the second time. According to the number of midpoints for the second time, the second clustering result and the degree of preference of the second clustering result are obtained; the degree of preference from the first to the second clustering results is fitted using the least squares method to obtain the first fitting curve. If the derivative value of the first fitting curve is negative, the iteration of the initial number of midpoints is stopped, and the second clustering result is the best clustering result; If the derivative value of the first fitting curve is positive, the initial number of midpoints continues to iterate, and the second number of midpoints plus one is used as the third number of midpoints to obtain the third clustering result and the degree of preference of the third clustering result; the least squares method is used to fit the degree of preference of the first to third clustering results to obtain the second fitting curve, and the derivative value of the second fitting curve is judged, and so on, until the derivative value of the vth fitting curve is negative, it can be stopped, and the v+1th clustering result is taken as the best clustering result.

[0035] It should be noted that the method for obtaining the degree of preference of each clustering result is the same as the method for obtaining the degree of preference of the first clustering result.

[0036] S004. According to the best clustering result, the final abnormality of each pressure data is obtained, and then the failure of the coating equipment is judged.

[0037] It should be noted that the best clustering result is obtained, and then the final abnormality degree of each pressure data is obtained according to the best clustering result to perform abnormal data identification.

[0038] In an embodiment of the present invention, in the best clustering result, the LOF algorithm is used to obtain the local outlier factor of each pressure data in the cluster to which it belongs, and record it as the final abnormality of each pressure data. If the final abnormality of any pressure data is greater than the abnormal threshold T, the pressure data is abnormal data. At this time, the coating equipment fails, and the remote monitoring and alarm of the coating equipment is immediately triggered to promptly notify relevant personnel to perform maintenance.

[0039] In the embodiment of the present invention, the abnormal threshold T is preset to 0.8. In other embodiments, implementers may preset the value of the abnormal threshold T according to the specific implementation method.

[0040] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A real-time early warning method for coating equipment failure, characterized in that: include: Collect pressure data and vibration data; The average of the absolute values ​​of the differences between the fitted values ​​of all surrounding data of each pressure data and the actual values ​​is taken as the instability of each pressure data; The product of the instability of each pressure data and the Pearson correlation coefficient of the instability sequence of each pressure data and the instability sequence of the vibration data corresponding to each pressure data is taken as the characteristic value of each pressure data; The number of initial midpoints is preset as K, the iterative step length of the number of midpoints is preset as 1, and the pressure data is clustered according to the initial number of midpoints to obtain the first clustering result; according to the first clustering result and the characteristic value of each pressure data, the preference degree of the first clustering result is obtained; K is added by one as the second number of midpoints, and the second clustering result is obtained, and the preference degree of the second clustering result is obtained according to the method for obtaining the preference degree of the first clustering result; according to the preference degree from the first clustering result to the second clustering result, it is judged whether the initial number of midpoints should continue to be iterated; If the iteration is stopped, the second clustering result is the best clustering result; if the iteration is continued, it stops until the best clustering result is obtained; According to the best clustering result, the final abnormality of each pressure data is obtained, and then the failure of the coating equipment is judged.

2. A coating equipment failure real-time early warning method according to claim 1, characterized in that: The acquisition of all surrounding data of each pressure data includes: The number of neighborhood data is preset as M, and the i-th pressure data, the M pressure data before the i-th pressure data, and the M pressure data after the i-th pressure data are used as the surrounding data of the i-th pressure data.

3. A coating equipment failure real-time early warning method according to claim 1, characterized in that: The acquisition of the fitting values ​​of all surrounding data of each pressure data includes: The surrounding data of the ith pressure data are fitted using the least square method to obtain the fitting value of each surrounding data of the ith pressure data.

4. A coating equipment failure real-time early warning method according to claim 1, characterized in that: The acquisition of the instability sequence of each pressure data and the instability sequence of the vibration data corresponding to each pressure data includes: Sequentially record the instability of each surrounding data of the i-th pressure data as the instability sequence of the i-th pressure data; The number of neighborhood data is preset as M, and the i-th vibration data, the M vibration data before the i-th vibration data, and the M vibration data after the i-th vibration data are taken as the surrounding data of the i-th vibration data; the instability of each surrounding data of the vibration data corresponding to the i-th pressure data is sequentially recorded as the instability sequence of the vibration data corresponding to the i-th pressure data.

5. A coating equipment failure real-time early warning method according to claim 1, characterized in that: The obtaining the preference degree of the first clustering result according to the first clustering result and the characteristic value of each pressure data includes: In the first clustering result, the LOF algorithm is used to obtain the local outlier factor of each pressure data in its clustering cluster, which is recorded as the abnormality of each pressure data; according to the order of the sampling time of the pressure data, the abnormality of each pressure data is sorted to obtain an abnormality sequence, and the eigenvalue of each pressure data is sorted to obtain a eigenvalue sequence; the Pearson correlation coefficient between the abnormality sequence and the eigenvalue sequence is used as the preferred degree of the first clustering result.

6. A coating equipment failure real-time early warning method according to claim 1, characterized in that: The step of judging whether to continue iterating the number of initial midpoints according to the preferred degree of the first clustering result and the second clustering result includes: The least squares method is used to fit the preferred degree of the first and second clustering results to obtain a fitting curve. If the derivative value of the fitting curve is negative, the iteration of the initial midpoint number is stopped. If the derivative value of the fitting curve is positive, the iteration of the initial midpoint number continues.

7. A coating equipment failure real-time early warning method according to claim 1, characterized in that: The method of obtaining the final abnormality of each pressure data according to the best clustering result, and then judging the failure of the coating equipment, includes: In the best clustering result, the LOF algorithm is used to obtain the local outlier factor of each pressure data in its cluster to which it belongs, and recorded as the final abnormality of each pressure data. If the final abnormality of any pressure data is greater than the abnormal threshold T, the pressure data is abnormal data. At this time, the coating equipment fails, and the remote monitoring and alarm of the coating equipment is immediately triggered to notify the relevant personnel in time for maintenance.

8. A coating equipment failure real-time early warning method according to claim 1, characterized in that: The collecting of pressure data and vibration data includes: The pressure sensor is placed between the coating roller and the support roller of the coating machine, and the vibration sensor is placed on the coating roller. The preset sampling time is 2 times / second. During the operation of the coating machine, the pressure data and vibration data are collected.

Images