Industrial equipment abnormity early warning method and device based on K-Means clustering and storage medium

Abstract

The invention discloses an industrial equipment abnormity early warning method and device based on K-Means clustering and a storage medium, and relates to the technical field of industrial equipment abnormity early warning, and the method comprises the steps: periodically collecting the total number of ferromagnetic abrasive particles in lubricating oil of industrial equipment; a dynamic analysis data set is constructed according to the total number of ferromagnetic abrasive particles in the industrial equipment lubricating oil liquid collected in a preset monitoring period; calculating a difference value between the acquisition values of every two adjacent acquisition time points in the dynamic analysis data set to obtain a difference value array; performing clustering analysis on the difference value array based on a K-Means clustering algorithm according to a plurality of preset wear state grades; sorting the plurality of clustering centers according to numerical values, calculating an average value of two adjacent clustering centers, and setting the average value as a dynamic threshold value; and comparing the real-time difference value with a dynamic threshold value to obtain an abnormal early warning result. According to the invention, the accuracy and sensitivity of abnormal prediction of the industrial equipment are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of industrial equipment abnormality early warning, and in particular to an industrial equipment abnormality early warning method, device and storage medium based on K-Means clustering. Background Art

[0002] Currently, industrial equipment often uses a fixed threshold based on the number of ferromagnetic wear particles to determine the health of the equipment. However, this method lacks sensitivity and is difficult to dynamically adapt to changes in equipment operating conditions, which can easily lead to false or missed alerts. Especially in the early stages of wear, small changes in the number of ferromagnetic wear particles are difficult to detect, resulting in delayed response to equipment maintenance.

[0003] Therefore, a K-Means clustering-based industrial equipment abnormality warning method, device and storage medium are developed to solve the above problems. Summary of the Invention

[0004] The present invention proposes an industrial equipment abnormality early warning method, device and storage medium based on K-Means clustering to solve the problems of insufficient sensitivity, easy false alarm or missed alarm and delayed equipment maintenance response in existing industrial equipment abnormality early warning methods.

[0005] The present invention achieves the above-mentioned purpose through the following technical solutions: In a first aspect, the present invention provides an industrial equipment abnormality early warning method based on K-Means clustering, comprising: Periodically collect the total number of ferromagnetic wear particles in the lubricating oil of industrial equipment, and construct a dynamic analysis data set based on the total number of ferromagnetic wear particles in the lubricating oil of industrial equipment collected within a preset monitoring period; Calculating the difference between the collected values ​​of every two adjacent collection time points in the dynamic analysis data set to obtain a difference array; Performing cluster analysis on the difference array based on a K-Means clustering algorithm according to a number of preset wear status levels to obtain a number of cluster centers, each cluster center corresponding to a wear status level; Sort the plurality of cluster centers by numerical value, calculate the average value of two adjacent cluster centers, and set the average value as a dynamic threshold; Obtaining the total number of ferromagnetic wear particles in the lubricating oil of the industrial equipment at the current acquisition time point and the total number of ferromagnetic wear particles in the lubricating oil of the industrial equipment at the previous acquisition time point adjacent to the current acquisition time point; Calculate the difference between the total number of ferromagnetic wear particles in the lubricating oil of the industrial equipment at the current collection time point and the total number of ferromagnetic wear particles in the lubricating oil of the industrial equipment at the previous collection time point as a real-time difference; The real-time difference is compared with the dynamic threshold to obtain an abnormal warning result.

[0006] Furthermore, cluster analysis is performed on the difference array based on several wear status levels based on the K-Means clustering algorithm, including: The number of clusters is specified to be three, and three differences are randomly selected from the difference array as the initial cluster centers of each cluster. Each cluster center corresponds to a preset wear state level, including normal, slightly abnormal, and severely abnormal. The three wear state levels are set based on the fact that the number of ferromagnetic abrasive particles in the equipment lubricating oil will show characteristics of slow growth, obvious growth, and rapid growth with different degrees of wear, so the state is divided into three levels; Iteratively optimize the cluster centers. In each iteration, first calculate the Euclidean distance from each difference value in the difference array to each cluster center, classify the difference value into the cluster with the closest distance, and then recalculate the mean of each cluster as the new cluster center. Continue iterating until the position change of the cluster center is less than the set threshold or the maximum number of iterations is reached, then stop iterating and obtain the final cluster center.

[0007] Furthermore, sorting the plurality of cluster centers according to numerical values, calculating the average value of two adjacent cluster centers, and setting the average value as a dynamic threshold comprises: Sort the three cluster centers from small to large; Calculate the average values ​​of two adjacent cluster centers in sequence to obtain the first average value and the second average value; The first average value is set as a first dynamic threshold, and the second average value is set as a second dynamic threshold. The first dynamic threshold is used to distinguish between normal and mild abnormality, and the second dynamic threshold is used to distinguish between mild abnormality and severe abnormality.

[0008] Furthermore, the real-time difference is compared with the dynamic threshold to obtain an abnormal warning result, including: comparing the real-time difference with the first dynamic threshold and the second dynamic threshold; When the real-time difference does not exceed the first dynamic threshold, it is determined that the industrial equipment is in a normal operating state; When the real-time difference is between the first dynamic threshold and the second dynamic threshold, it is determined that the industrial equipment is in a slightly abnormal state; When the real-time difference exceeds the second dynamic threshold, it is determined that the industrial equipment is in a serious abnormal state.

[0009] Furthermore, it also includes issuing an alarm based on the abnormal warning result, including: When industrial equipment is determined to be in a slightly abnormal state, a primary alarm is triggered to prompt maintenance personnel to pay attention to the equipment status; When industrial equipment is determined to be in a serious abnormal state, a high-level alarm will be issued to prompt shutdown and inspection.

[0010] In a second aspect, the present invention further provides an industrial equipment abnormality warning device based on K-Means clustering, comprising: A construction module, the construction module is used to periodically collect the total number of ferromagnetic wear particles in the lubricating oil of industrial equipment, and to construct a dynamic analysis data set based on the total number of ferromagnetic wear particles in the lubricating oil of industrial equipment collected during a preset monitoring period; a first calculation module, configured to calculate the difference between the collected values ​​of each two adjacent collection time points in the dynamic analysis data set to obtain a difference array; A clustering module, configured to perform cluster analysis on the difference array based on a K-Means clustering algorithm according to a number of preset wear status levels, to obtain a number of cluster centers, each cluster center corresponding to a wear status level; A threshold module is used to sort the cluster centers according to their numerical values, calculate the average value of two adjacent cluster centers, and set the average value as a dynamic threshold; an acquisition module, the acquisition module being configured to acquire the total number of ferromagnetic wear particles in the lubricating oil of the industrial equipment at a current acquisition time point and the total number of ferromagnetic wear particles in the lubricating oil of the industrial equipment at a previous acquisition time point adjacent to the current acquisition time point; a second calculation module, configured to calculate a difference between the total number of ferromagnetic abrasive particles in the lubricating oil of the industrial equipment at the current acquisition time point and the total number of ferromagnetic abrasive particles in the lubricating oil of the industrial equipment at the previous acquisition time point as a real-time difference; A comparison module is used to compare the real-time difference with the dynamic threshold to obtain an abnormal warning result.

[0011] In a third aspect, the present invention provides a storage medium, which is a readable storage medium. A computer program is stored on the readable storage medium. When the computer program is executed by a processor, the industrial equipment abnormality warning method based on K-Means clustering is implemented.

[0012] The beneficial effects of the present invention are: The industrial equipment anomaly warning method, device and storage medium based on K-Means clustering proposed in the present invention improve the prediction accuracy and sensitivity of industrial equipment anomalies, and have the advantages of low deployment cost, high analysis efficiency and fast anomaly response. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] Figure 1This is a flow chart of an industrial equipment abnormality warning method based on K-Means clustering in this application. DETAILED DESCRIPTION

[0014] To make the objectives, technical solutions, and advantages of the embodiments of the present invention more apparent, the technical solutions of the embodiments of the present invention will be described clearly and completely below in conjunction with the accompanying drawings of the embodiments of the present invention. It should be understood that the described embodiments are only a portion of the embodiments of the present invention, not all of them. Generally, the components of the embodiments of the present invention described and illustrated in the drawings herein may be arranged and designed in a variety of different configurations.

[0015] Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the invention as claimed, but rather merely represents selected embodiments of the present invention. All other embodiments derived by persons of ordinary skill in the art based on the embodiments of the present invention without creative effort shall fall within the scope of protection of the present invention.

[0016] It should be noted that similar reference numerals and letters denote similar items in the following drawings, and therefore, once an item is defined in one drawing, it does not need to be further defined or explained in subsequent drawings.

[0017] Furthermore, the terms “first”, “second”, etc. are merely used for distinguishing descriptions and should not be understood as indicating or implying relative importance.

[0018] The specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0019] like Figure 1 As shown, a K-Means clustering-based industrial equipment abnormality early warning method includes: The present invention provides an industrial equipment abnormality early warning method based on K-Means clustering, comprising: S1: Periodically collect the total number of ferromagnetic wear particles in the lubricating oil of industrial equipment, and construct a dynamic analysis data set based on the total number of ferromagnetic wear particles in the lubricating oil of industrial equipment collected within a preset monitoring period; S2: Calculate the difference between the collected values ​​of each two adjacent collection time points in the dynamic analysis data set to obtain a difference array; perform differential processing on the adjacent data in the set to generate a difference sequence to describe the change trend; S3: performing cluster analysis on the difference array based on a K-Means clustering algorithm according to a number of preset wear status levels to obtain a number of cluster centers, each cluster center corresponding to a wear status level; S4: sorting the plurality of cluster centers according to their numerical values, calculating the average value of two adjacent cluster centers, and setting the average value as a dynamic threshold; S5: Obtaining the total number of ferromagnetic wear particles in the lubricating oil of the industrial equipment at the current acquisition time point and the total number of ferromagnetic wear particles in the lubricating oil of the industrial equipment at the previous acquisition time point adjacent to the current acquisition time point; S6: Calculate the difference between the total number of ferromagnetic wear particles in the lubricating oil of the industrial equipment at the current collection time point and the total number of ferromagnetic wear particles in the lubricating oil of the industrial equipment at the previous collection time point as a real-time difference; S7: Compare the real-time difference with the dynamic threshold to obtain an abnormality warning result.

[0020] In one embodiment, cluster analysis is performed on the difference array based on a K-Means clustering algorithm according to several wear status levels, including: Specifying three clusters, and randomly selecting three difference values ​​from the difference value array as initial cluster centers of each cluster, each cluster center corresponding to a preset wear status level, the three wear status levels including normal, slightly abnormal, and severely abnormal; Iteratively optimize the cluster centers. In each iteration, first calculate the Euclidean distance from each difference value in the difference array to each cluster center, classify the difference value into the cluster with the closest distance, and then recalculate the mean of each cluster as the new cluster center. Continue iterating until the position change of the cluster center is less than the set threshold or the maximum number of iterations is reached, then stop iterating and obtain the final cluster center.

[0021] In one embodiment, sorting the plurality of cluster centers by numerical value, calculating the average value of two adjacent cluster centers, and setting the average value as a dynamic threshold comprises: Sort the three cluster centers from small to large; Calculate the average values ​​of two adjacent cluster centers in sequence to obtain the first average value and the second average value; The first average value is set as a first dynamic threshold, and the second average value is set as a second dynamic threshold. The first dynamic threshold is used to distinguish between normal and mild abnormality, and the second dynamic threshold is used to distinguish between mild abnormality and severe abnormality.

[0022] In one embodiment, comparing the real-time difference with the dynamic threshold to obtain an abnormality warning result includes: comparing the real-time difference with the first dynamic threshold and the second dynamic threshold; When the real-time difference does not exceed the first dynamic threshold, it is determined that the industrial equipment is in a normal operating state; When the real-time difference is between the first dynamic threshold and the second dynamic threshold, it is determined that the industrial equipment is in a slightly abnormal state; When the real-time difference exceeds the second dynamic threshold, it is determined that the industrial equipment is in a serious abnormal state.

[0023] In one embodiment, the method further includes issuing an alarm based on the abnormal warning result, including: When industrial equipment is determined to be in a slightly abnormal state, a primary alarm is triggered to prompt maintenance personnel to pay attention to the equipment status; When industrial equipment is determined to be in a serious abnormal state, a high-level alarm will be issued to prompt shutdown and inspection.

[0024] For example, within a set monitoring period (e.g., one month), data on the total number of ferromagnetic wear particles in the oil is continuously collected at a fixed sampling interval of 30 minutes to ensure the timeliness and continuity of the monitoring data. During the data preprocessing stage, the difference in the number of ferromagnetic wear particles between two adjacent sampling periods (i.e., 30-minute intervals) is calculated to construct a complete difference sequence. This difference sequence can intuitively reflect the changes in the rate of wear particle generation during equipment operation, providing a quantitative basis for subsequent wear status assessment.

[0025] During the data analysis phase, a K-Means clustering algorithm was used to perform pattern recognition on the difference sequence. The algorithm initially randomly selected three cluster centers, corresponding to three preset wear status levels. Data classification was achieved through an iterative optimization process: in each iteration, the Euclidean distance from each data point to each cluster center was first calculated, and the data point was assigned to the cluster with the closest distance. The mean of each cluster was then recalculated as the new cluster center. This process continued until the change in the position of the cluster center was less than the set threshold or the maximum number of iterations was reached. After the algorithm converged, the three final cluster centers were sorted by numerical value, and the arithmetic mean of the adjacent centers was taken as the dynamic threshold for status determination: T1 (distinguishing between normal and mild abnormality) and T2 (distinguishing between mild abnormality and severe abnormality).

[0026] During the status determination and alarm output stage, the real-time monitoring data is compared and analyzed with the dynamic threshold. When the difference does not exceed T1, it is determined to be a normal operating state, and the system maintains routine monitoring; when the difference is between T1 and T2, it is determined to be a mild abnormality, and the system triggers a primary alarm, prompting maintenance personnel to pay attention to the equipment status; when the difference exceeds T2, it is determined to be a serious abnormality, and the system immediately issues a high-level alarm, recommending shutdown inspection. This method achieves accurate assessment of equipment wear conditions and graded early warning through adaptive threshold setting and three-level status classification, effectively improving the timeliness and scientific nature of equipment maintenance. At the same time, the system will record the complete analysis process and judgment results to provide data support for subsequent equipment health management.

[0027] In one embodiment, a K-Means clustering-based industrial equipment abnormality warning device includes: A construction module, the construction module is used to periodically collect the total number of ferromagnetic wear particles in the lubricating oil of industrial equipment, and to construct a dynamic analysis data set based on the total number of ferromagnetic wear particles in the lubricating oil of industrial equipment collected during a preset monitoring period; a first calculation module, configured to calculate the difference between the collected values ​​of each two adjacent collection time points in the dynamic analysis data set to obtain a difference array; A clustering module, configured to perform cluster analysis on the difference array based on a K-Means clustering algorithm according to a number of preset wear status levels, to obtain a number of cluster centers, each cluster center corresponding to a wear status level; A threshold module is used to sort the cluster centers according to their numerical values, calculate the average value of two adjacent cluster centers, and set the average value as a dynamic threshold; an acquisition module, the acquisition module being configured to acquire the total number of ferromagnetic wear particles in the lubricating oil of the industrial equipment at a current acquisition time point and the total number of ferromagnetic wear particles in the lubricating oil of the industrial equipment at a previous acquisition time point adjacent to the current acquisition time point; a second calculation module, configured to calculate a difference between the total number of ferromagnetic abrasive particles in the lubricating oil of the industrial equipment at the current acquisition time point and the total number of ferromagnetic abrasive particles in the lubricating oil of the industrial equipment at the previous acquisition time point as a real-time difference; A comparison module is used to compare the real-time difference with the dynamic threshold to obtain an abnormal warning result.

[0028] In one embodiment, a storage medium is provided, wherein the storage medium is a readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the method for abnormal warning of industrial equipment based on K-Means clustering is implemented.

[0029] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the technical principles of the present invention. These improvements and modifications should also be regarded as within the scope of protection of the present invention.

Claims

1. A K-Means clustering-based industrial equipment abnormality early warning method, characterized in that: include: Periodically collect the total number of ferromagnetic wear particles in the lubricating oil of industrial equipment, and construct a dynamic analysis data set based on the total number of ferromagnetic wear particles in the lubricating oil of industrial equipment collected within a preset monitoring period; Calculating the difference between the collected values ​​of every two adjacent collection time points in the dynamic analysis data set to obtain a difference array; Performing cluster analysis on the difference array based on a K-Means clustering algorithm according to a number of preset wear status levels to obtain a number of cluster centers, each cluster center corresponding to a wear status level; Sort the plurality of cluster centers by numerical value, calculate the average value of two adjacent cluster centers, and set the average value as a dynamic threshold; Obtaining the total number of ferromagnetic wear particles in the lubricating oil of the industrial equipment at the current acquisition time point and the total number of ferromagnetic wear particles in the lubricating oil of the industrial equipment at the previous acquisition time point adjacent to the current acquisition time point; Calculate the difference between the total number of ferromagnetic wear particles in the lubricating oil of the industrial equipment at the current collection time point and the total number of ferromagnetic wear particles in the lubricating oil of the industrial equipment at the previous collection time point as a real-time difference; The real-time difference is compared with the dynamic threshold to obtain an abnormal warning result.

2. The industrial equipment abnormality early warning method based on K-Means clustering according to claim 1 is characterized in that: Cluster analysis is performed on the difference array based on several wear status levels using a K-Means clustering algorithm, including: Specifying three clusters, and randomly selecting three difference values ​​from the difference value array as initial cluster centers of each cluster, each cluster center corresponding to a preset wear status level, the three wear status levels including normal, slightly abnormal, and severely abnormal; Iteratively optimize the cluster centers. In each iteration, first calculate the Euclidean distance from each difference value in the difference array to each cluster center, classify the difference value into the cluster with the closest distance, and then recalculate the mean of each cluster as the new cluster center. Continue iterating until the position change of the cluster center is less than the set threshold or the maximum number of iterations is reached, then stop iterating and obtain the final cluster center.

3. The industrial equipment abnormality early warning method based on K-Means clustering according to claim 2 is characterized in that: Sorting the plurality of cluster centers according to numerical values, calculating the average value of two adjacent cluster centers, and setting the average value as a dynamic threshold, including: Sort the three cluster centers from small to large; Calculate the average values ​​of two adjacent cluster centers in sequence to obtain the first average value and the second average value; The first average value is set as a first dynamic threshold, and the second average value is set as a second dynamic threshold. The first dynamic threshold is used to distinguish between normal and mild abnormality, and the second dynamic threshold is used to distinguish between mild abnormality and severe abnormality.

4. The industrial equipment abnormality early warning method based on K-Means clustering according to claim 3 is characterized in that: Comparing the real-time difference with the dynamic threshold to obtain an abnormal warning result includes: comparing the real-time difference with the first dynamic threshold and the second dynamic threshold; When the real-time difference does not exceed the first dynamic threshold, it is determined that the industrial equipment is in a normal operating state; When the real-time difference is between the first dynamic threshold and the second dynamic threshold, it is determined that the industrial equipment is in a slightly abnormal state; When the real-time difference exceeds the second dynamic threshold, it is determined that the industrial equipment is in a serious abnormal state.

5. The industrial equipment abnormality early warning method based on K-Means clustering according to claim 4 is characterized in that: It also includes issuing alarms based on abnormal warning results, including: When industrial equipment is determined to be in a slightly abnormal state, a primary alarm is triggered to prompt maintenance personnel to pay attention to the equipment status; When industrial equipment is determined to be in a serious abnormal state, a high-level alarm will be issued to prompt shutdown and inspection.

6. An industrial equipment abnormality warning device based on K-Means clustering, characterized in that: include A construction module, the construction module is used to periodically collect the total number of ferromagnetic wear particles in the lubricating oil of industrial equipment, and to construct a dynamic analysis data set based on the total number of ferromagnetic wear particles in the lubricating oil of industrial equipment collected during a preset monitoring period; a first calculation module, configured to calculate the difference between the collected values ​​of each two adjacent collection time points in the dynamic analysis data set to obtain a difference array; A clustering module, configured to perform cluster analysis on the difference array based on a K-Means clustering algorithm according to a number of preset wear status levels, to obtain a number of cluster centers, each cluster center corresponding to a wear status level; A threshold module is used to sort the cluster centers according to their numerical values, calculate the average value of two adjacent cluster centers, and set the average value as a dynamic threshold; an acquisition module, the acquisition module being configured to acquire the total number of ferromagnetic wear particles in the lubricating oil of the industrial equipment at a current acquisition time point and the total number of ferromagnetic wear particles in the lubricating oil of the industrial equipment at a previous acquisition time point adjacent to the current acquisition time point; a second calculation module, configured to calculate a difference between the total number of ferromagnetic abrasive particles in the lubricating oil of the industrial equipment at the current acquisition time point and the total number of ferromagnetic abrasive particles in the lubricating oil of the industrial equipment at the previous acquisition time point as a real-time difference; A comparison module is used to compare the real-time difference with the dynamic threshold to obtain an abnormal warning result.

7. A storage medium, characterized in that: The storage medium is a readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the method for abnormal warning of industrial equipment based on K-Means clustering as claimed in any one of claims 1 to 5 is implemented.

Images