Electric submersible pump vibration signal early warning device and method based on density DBSCAN and related products

Through the vibration signal early warning method of submersible oil electric pump based on density DBSCAN algorithm, abnormal data points are automatically identified and classified, which solves the problem of difficult to deal with complex fault modes and noise interference in the existing technology, and improves the accuracy and timeliness of fault diagnosis. , ensuring the stability and safety of oil field production.

CN119046713APending Publication Date: 2024-11-29HUAIWO (HANGZHOU) TECHNOLOGY CO LTD
View PDF 0 Cites 6 Cited by

Patent Information

Application Number
CN202411140604.6
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-20
Publication Date
2024-11-29

AI Technical Summary

Technical Problem

Existing vibration signal analysis technology of submersible oil electric pumps is difficult to cope with irregular changes and complex and changeable fault modes, and the mixing of noise and interference signals leads to increased difficulty in identifying outliers, which may cause false alarms or miss key fault signals, which will damage fault diagnosis. accuracy and timeliness.

Method used

The vibration signal early warning method of submersible oil electric pump based on density DBSCAN algorithm is adopted. By collecting vibration signal data points, setting the neighborhood radius and minimum number of points of spatial clustering, automatically identifying and classifying abnormal data points, generating abnormal data clusters, and performing Detect to trigger an early warning.

Benefits of technology

It improves the efficiency and accuracy of abnormal detection, reduces the impact of noise and interference signals, enhances the accuracy and timeliness of fault diagnosis, and ensures the safety and efficiency of oil field production.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119046713A_ABST
    Figure CN119046713A_ABST
Patent Text Reader

Abstract

The invention provides an electric submersible pump vibration signal early warning device and method based on density DBSCAN and related products, and the method comprises the steps: collecting vibration signal data points of an electric submersible pump, so as to obtain a vibration signal data set; setting a neighborhood radius xi and a minimum point number MinPts of spatial clustering based on a density DBSCAN algorithm; based on the vibration signal data set, selecting one vibration signal data point which is not accessed as a target vibration signal data point; comparing the number of the vibration signal data points in the neighborhood radius xi limited area of the target vibration signal data points with the minimum number MinPts to obtain a comparison result; judging whether the target vibration signal data points are candidate abnormal data points or not based on the comparison result; traversing the vibration signal data points which are not accessed for screening so as to generate an abnormal data cluster based on the candidate abnormal data points; and the abnormal data cluster is detected to obtain a detection result, and when the detection result meets a set condition, the early warning device gives an alarm to prompt that the electric submersible pump operates abnormally.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention belongs to the technical field of vibration detection of submersible electric pumps, and in particular relates to a density DBSCAN-based submersible electric pump vibration signal early warning device method and related products. Background Art

[0002] As the core equipment in the oil extraction industry, the stable operation of submersible electric pumps is crucial to the continuity and economic benefits of oilfield production. Given the extreme working environment and complex structure, they are prone to various faults during long-term operation. Therefore, it is particularly important to implement efficient and real-time fault diagnosis and monitoring of submersible electric pumps to ensure the safety and efficiency of oilfield production.

[0003] In this context, vibration signal analysis technology, with its advantages of non-invasiveness, instant feedback and high precision, occupies a core position in the health monitoring system of submersible electric pumps. Through in-depth analysis of the vibration signals generated during the operation of the equipment, it can not only preliminarily identify whether the equipment has potential failure risks, but also accurately determine the specific type and location of the failure, providing solid data support for the rapid formulation and implementation of maintenance strategies, thereby ensuring that the submersible electric pump can continue to serve the oilfield production reliably.

[0004] However, in the existing practice of outlier analysis of vibration signals of submersible pumps, there are several challenges that need to be overcome. Traditional analysis methods generally rely on fixed threshold settings or preset models, which limits their ability to flexibly respond to the irregular changes and complex and changeable failure modes that are prevalent in vibration signals of submersible pumps. To make matters more complicated, the mixing of noise and interference signals greatly increases the difficulty of outlier identification, which may not only cause unnecessary false alarms, but also miss key fault signals, thereby compromising the accuracy and timeliness of fault diagnosis, posing a potential threat to the stability and safety of oilfield production. Summary of the invention

[0005] Based on the above problems, the embodiments of the present application provide a submersible electric pump vibration signal warning device method based on density DBSCAN and related products to solve the problems existing in the above-mentioned prior art.

[0006] The present application discloses a vibration signal early warning method for a submersible electric pump based on density DBSCAN, comprising: S101: collecting vibration signal data points of a submersible electric pump to obtain a vibration signal data set; S201: setting a neighborhood radius ξ and a minimum number of points MinPts of spatial clustering based on a density DBSCAN algorithm; S301: based on the vibration signal data set, selecting one of the vibration signal data points that has not been visited as a target vibration signal data point; S401: comparing the vibration signal data points within the area limited by the neighborhood radius ξ of the target vibration signal data point; The comparison result is obtained by comparing the number of points of the data point with the minimum number of points MinPts; S501: judging whether the target vibration signal data point is a candidate abnormal data point based on the comparison result; S601: traversing the vibration signal data points that have not been visited, and executing steps S301 to S501 to generate an abnormal data cluster based on the candidate abnormal data points; S701: detecting the abnormal data cluster to obtain a detection result, and when the detection result meets the set condition, the early warning device alarms to prompt that the submersible electric pump is operating abnormally.

[0007] Optionally, setting the neighborhood radius ξ of spatial clustering includes: finding the number k of each vibration signal data point and its nearest neighbor points in the vibration signal data set based on a k-nearest neighbor graph method, so as to set the neighborhood radius ξ of spatial clustering based on the k value.

[0008] Optionally, based on the k-nearest neighbor graph method, the number k of each vibration signal data point and its nearest neighbor points in the vibration signal data set is found to set the neighborhood radius ξ of spatial clustering based on the k value, including: finding the k nearest neighbor points of the vibration signal data point based on the k-nearest neighbor graph method; calculating the distance of each vibration signal data point to the kth nearest neighbor point based on the silhouette coefficient method to obtain the silhouette coefficient ψ; traversing the possible range of k values, calculating the k value that maximizes the average value of the silhouette coefficient ψ to obtain the optimized k value; and setting the neighborhood radius ξ based on the optimized k value.

[0009] Optionally, the calculating the distance from each vibration signal data point to the kth nearest neighbor point based on the silhouette coefficient method to obtain the silhouette coefficient ψ includes: calculating the average distance a between the target vibration signal data point P in the vibration signal data set and the other vibration signal data points M in the same cluster except the target vibration signal data point P based on the silhouette coefficient method; calculating the average distance b between the target vibration signal data point P in the vibration signal data set and the other vibration signal data points N in the nearest cluster except the target vibration signal data point P based on the silhouette coefficient method; and calculating the silhouette coefficient ψ based on the average distance a and the average distance b.

[0010] Optionally, the minimum number of points MinPts is set based on the following formula: MinPts=2*ξ+1, where ξ is the neighborhood radius.

[0011] Optionally, the comparing the relationship between the number of the vibration signal data points contained in the neighborhood radius ξ of the target vibration signal data point P and the minimum number of points MinPts to obtain a comparison result includes: comparing the relationship between the number of all the vibration signal data points X within the area defined by the neighborhood radius ξ to which the target vibration signal data point P belongs, except the target vibration signal data point P, and the minimum number of points MinPts.

[0012] Optionally, the abnormal data cluster is detected to obtain a detection result, and when the detection result meets the set conditions, the early warning device sends an alarm to indicate that the submersible electric pump is operating abnormally, including: defining the normal operation setting value based on the set early warning parameter information; calculating the abnormal data cluster characteristics based on the abnormal data cluster; comparing the normal operation setting value with the abnormal data cluster characteristics to obtain a detection result, and when the detection result meets the set conditions, the early warning device sends an alarm to indicate that the submersible electric pump is operating abnormally.

[0013] The present application also provides a submersible electric pump vibration signal early warning device based on density DBSCAN, comprising: an information acquisition unit: collecting vibration signal data points of the submersible electric pump to obtain a vibration signal data set; a preset unit: setting a neighborhood radius ξ and a minimum number of points MinPts of spatial clustering based on a density DBSCAN algorithm; a comparison unit: based on the vibration signal data set, selecting one of the vibration signal data points that have not been visited as a target vibration signal data point; comparing the number of vibration signal data points within the area defined by the neighborhood radius ξ of the target vibration signal data point with the minimum number of points MinPts to obtain a comparison result; a screening unit: judging whether the target vibration signal data point is a candidate abnormal data point based on the comparison result; traversing the vibration signal data points that have not been visited, and executing the comparison unit to generate an abnormal data cluster based on the candidate abnormal data point; a judgment unit: detecting the abnormal data cluster to obtain a detection result, and when the detection result meets the set condition, the early warning device alarms to indicate that the submersible electric pump is operating abnormally.

[0014] The present application also provides an electronic device, comprising: a memory and a processor, wherein the memory stores a computer executable program, and the processor is used to execute the computer executable program to implement any one of the methods described in the present application.

[0015] A computer storage medium, characterized in that a computer executable program is stored on the computer storage medium, and the computer executable program can implement any method described in the present application when it is executed.

[0016] The present invention provides a method and related products of a vibration signal early warning device for a submersible electric pump based on density DBSCAN, wherein a method of a vibration signal early warning device for a submersible electric pump based on density DBSCAN comprises: S101: collecting vibration signal data points of a submersible electric pump to obtain a vibration signal data set; S201: setting a neighborhood radius ξ and a minimum number of points MinPts of spatial clustering based on a density DBSCAN algorithm; S301: based on the vibration signal data set, selecting one of the vibration signal data points that has not been visited as a target vibration signal data point; S401: comparing the target vibration signal data point The number of vibration signal data points within the area limited by the neighborhood radius ξ and the minimum number of points MinPts are compared to obtain a comparison result; S501: based on the comparison result, determine whether the target vibration signal data point is a candidate abnormal data point; S601: traverse the vibration signal data points that have not been visited, and execute steps S301 to S501 to generate abnormal data clusters based on the candidate abnormal data points; S701: detect the abnormal data clusters to obtain a detection result. When the detection result meets the set conditions, the early warning device alarms to indicate that the submersible electric pump is operating abnormally. In this application, clustering analysis of vibration signal data points is performed by the DBSCAN algorithm, which can automatically identify and classify potential abnormal data points without presetting a fixed threshold, thereby improving the efficiency and accuracy of abnormal detection. In addition, traditional analysis methods rely on fixed thresholds or preset models, which are difficult to deal with irregular changes and complex and changeable fault modes in the vibration signal of the submersible electric pump. The DBSCAN algorithm clusters based on the density of data points, which can better adapt to complex changes in vibration signals. When processing vibration signals, the mixing of noise and interference signals is a common problem. Moreover, the DBSCAN algorithm can effectively distinguish noise points and abnormal data points through density clustering, reduce false positives and false negatives, and improve the accuracy and timeliness of fault diagnosis. This method can collect and analyze the vibration signals of submersible electric pumps in real time, discover potential faults in a timely manner, provide data support for the rapid formulation and implementation of maintenance strategies, and ensure the safety and efficiency of oilfield production. In summary, through the method provided in this application, the flexibility of dealing with the irregular changes and complex and changeable fault modes that are prevalent in the vibration signals of submersible electric pumps is increased. Moreover, it can eliminate the interference of noise and interference signals, improve the accuracy and timeliness of fault diagnosis, and enhance the stability and safety of oilfield production. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0018] Figure 1 This is a flow chart of a submersible electric pump vibration signal early warning method based on density DBSCAN in Example 1 of the present application; DETAILED DESCRIPTION

[0019] The implementation of any technical solution of the embodiments of the present application does not necessarily require achieving all of the above advantages at the same time.

[0020] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only 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 ordinary technicians in this field without creative work are within the scope of protection of the present invention. First embodiment

[0021] Figure 1 Flow chart of the early warning method of vibration signal of submersible electric pump based on density DBSCAN in the first embodiment of the present application; Figure 1As shown, a vibration signal early warning method for a submersible electric pump based on density DBSCAN is characterized by comprising: S101: collecting vibration signal data points of a submersible electric pump to obtain a vibration signal data set; S201: setting a neighborhood radius ξ and a minimum number of points MinPts of spatial clustering based on a density DBSCAN algorithm; S301: based on the vibration signal data set, selecting one of the vibration signal data points that has not been visited as a target vibration signal data point; S401: comparing the vibration signal data points within the area limited by the neighborhood radius ξ of the target vibration signal data point. The number of points of the target vibration signal data point and the minimum number of points MinPts are compared to obtain the comparison result; S501: based on the comparison result, determine whether the target vibration signal data point is a candidate abnormal data point; S601: traverse the vibration signal data points that have not been visited, and execute steps S301 to S501 to generate abnormal data clusters based on the candidate abnormal data points; S701: detect the abnormal data cluster to obtain the detection result. When the detection result meets the set conditions, the early warning device alarms to indicate that the submersible electric pump is operating abnormally. In this embodiment, the method performs cluster analysis on the vibration signal data points through the DBSCAN algorithm, which can automatically identify and classify potential abnormal data points without presetting a fixed threshold, thereby improving the efficiency and accuracy of abnormal detection. In addition, the traditional analysis method relies on a fixed threshold or a preset model, which is difficult to deal with irregular changes and complex and changeable fault modes in the vibration signal of the submersible electric pump. The DBSCAN algorithm clusters based on the density of data points, which can better adapt to the complex changes in the vibration signal. When processing vibration signals, the mixing of noise and interference signals is a common problem. Moreover, the DBSCAN algorithm can effectively distinguish noise points and abnormal data points through density clustering, reduce false positives and false negatives, and improve the accuracy and timeliness of fault diagnosis. This method can collect and analyze the vibration signals of submersible electric pumps in real time, detect potential faults in a timely manner, provide data support for the rapid formulation and implementation of maintenance strategies, and ensure the safety and efficiency of oilfield production.

[0022] Optionally, the neighborhood radius ξ refers to the radius of a circular area around a data point, and all points in the area are regarded as "neighbors" or "neighborhood points" of the data point. In this embodiment, the setting of the neighborhood radius allows the algorithm to maintain high flexibility when processing data sets with different densities and distribution characteristics. By adjusting the size of the neighborhood radius, the spatial relationship between data points can be more effectively captured, thereby improving the accuracy and effectiveness of clustering. Moreover, the method can deal with noise and outliers in the data. In the DBSCAN algorithm, the setting of the neighborhood radius helps to distinguish core points, boundary points, and noise points. Noise points (i.e., data points with insufficient points in their neighborhood) can be identified and processed, thereby reducing interference with clustering results and improving the robustness of the clustering algorithm. In addition, through the definition of the neighborhood radius, the algorithm can accurately identify dense and sparse areas in the data. This is particularly important for the analysis of vibration signals of submersible electric pumps, because anomalies in vibration signals are usually closely related to changes in signal density. By capturing these changes, the algorithm can more accurately identify potential fault signals. In addition, by adjusting the size of the neighborhood radius, the scheme also has the ability of multi-scale analysis. For vibration signal features of different scales and densities, appropriate neighborhood radius can be selected for cluster analysis to obtain more comprehensive and in-depth information. Finally, the definition of circular neighborhood makes the algorithm more intuitive when processing and interpreting data. People can more easily understand the spatial relationship between data points and the formation process of clustering results, thereby enhancing their understanding and application of the algorithm.

[0023] Optionally, the minimum number of points MinPts is used to define the minimum number of points required for an area to become a dense area. Specifically, for a given data point p, if the number of other data points within its neighborhood radius ε is greater than or equal to MinPts, then the point is considered a core point. Here, "other data points" refer to points located in the ε-neighborhood of point p except point p itself.

[0024] Optionally, the unvisited data points are data points that have not been checked by the algorithm, have not been classified into any cluster, or have not been marked as noise during the DBSCAN clustering algorithm processing. In this embodiment, traditional analysis methods often rely on fixed thresholds or preset models, and it is difficult to flexibly respond to the irregular changes and complex and changeable fault modes that are prevalent in the vibration signals of submersible electric pumps. However, this solution can automatically adapt to the density changes of data points through a density-based clustering method (DBSCAN), and effectively identify and distinguish normal signals from abnormal signals. Moreover, by setting the minimum number of points MinPts and the neighborhood radius ε, the DBSCAN algorithm can distinguish between core points and noise points, thereby more accurately identifying abnormal data points. This method can reduce the impact of noise and interference signals on outlier identification and improve the accuracy of fault diagnosis. In addition, in the operating environment of submersible electric pumps, noise and interference signals are inevitable. Through density-based clustering analysis, this solution can effectively distinguish non-equipment vibration signals generated by environmental factors, sensor noise, electromagnetic interference, etc., thereby reducing false alarms and missed alarms and improving the reliability of the early warning system. In addition, the synchronization and storage processing ensures the consistency of data in time, which enables a comprehensive and multi-angle analysis of the operating status of the submersible pump, improving the accuracy and comprehensiveness of fault diagnosis.

[0025] Optionally, setting the neighborhood radius ξ of spatial clustering includes: finding the number k of each vibration signal data point and its nearest neighbor points in the vibration signal data set based on a k-nearest neighbor graph method, so as to set the neighborhood radius ξ of spatial clustering based on the k value.

[0026] Optionally, the method of finding the number k of each vibration signal data point and its nearest neighbor in the vibration signal data set based on the k-nearest neighbor graph method to set the neighborhood radius ξ of spatial clustering based on the k value includes: finding the nearest k neighbors of the vibration signal data point based on the k-nearest neighbor graph method; calculating the distance from each vibration signal data point to the kth nearest neighbor based on the silhouette coefficient method to obtain the silhouette coefficient ψ; traversing the possible k value range to calculate the k value that maximizes the average value of the silhouette coefficient ψ to obtain the optimized k value; and setting the neighborhood radius ξ based on the optimized k value. In this embodiment, the neighborhood radius can be determined based on the distribution characteristics of the data itself, rather than relying on a preset fixed value, by the k-nearest neighbor graph method. This enables the algorithm to adapt to the characteristics of different data sets more flexibly, improving the versatility and adaptability of the algorithm. Moreover, by optimizing the k value to select the best neighborhood radius, it is possible to ensure that a reasonable cluster structure is formed during the clustering process. This helps to reduce the problem of being too loose or too compact within the cluster due to improper setting of the neighborhood radius, thereby improving the quality and accuracy of clustering. In addition, since the neighborhood radius is determined based on the relative distance between data points, it is better able to identify and handle noise and outliers in the data set. In the clustering process, these noise and outliers are often regarded as outliers or, so as not to cause excessive interference to the formation of normal clusters. In addition, by adjusting the k value, clustering analysis of data at different scales can be achieved. This is particularly useful for situations where the data structure needs to be understood from multiple perspectives or levels. Finally, although the computational complexity of the k-nearest neighbor graph method is relatively high, once the appropriate k value and neighborhood radius are determined, the subsequent clustering process will be more efficient. In addition, the computational efficiency can be further improved by optimizing the algorithm implementation and selecting the appropriate data structure (such as kd tree, ball tree, etc.).

[0027] Optionally, the distance from each vibration signal data point to the kth nearest neighbor point is calculated based on the silhouette coefficient method to obtain the silhouette coefficient ψ, including: calculating the average distance a between the target vibration signal data point P in the vibration signal data set and other vibration signal data points M in the same cluster except the target vibration signal data point P based on the silhouette coefficient method; calculating the average distance b between the target vibration signal data point P in the vibration signal data set and other vibration signal data points N in the nearest cluster except the target vibration signal data point P based on the silhouette coefficient method; and calculating the silhouette coefficient ψ based on the average distance a and the average distance b. In this embodiment, the silhouette coefficient provides a quantitative indicator to measure the quality of clustering effect. By calculating the average distance (cohesion) of each data point to other data points in the cluster to which it belongs and the average distance (separation) to the data points in the nearest cluster, the compactness and separation of clustering can be objectively evaluated. In addition, when setting the neighborhood radius based on the k-nearest neighbor graph method, by traversing different k values ​​and calculating the k value that maximizes the average value of the silhouette coefficient (optimizing the k value), the optimal neighborhood radius setting can be found, thereby optimizing the parameters of DBSCAN clustering and improving the clustering effect. Moreover, since the silhouette coefficient can reflect the closeness between a data point and other data points in its cluster, when identifying abnormal data points, it can more accurately determine which data points may be signal points generated by noise or non-equipment vibration, thereby improving the accuracy of anomaly detection. In addition, by calculating and analyzing the silhouette coefficient of each data point, noise points and abnormal points can be more effectively identified and processed, reducing the impact of these points on clustering results and subsequent anomaly detection, thereby enhancing the robustness of the entire early warning method. Finally, the silhouette coefficient method is not only applicable to single-dimensional data, but also to multi-dimensional data. In the vibration signal early warning of submersible pumps, since the vibration signal data is often multi-dimensional, the silhouette coefficient method can effectively process these data and provide a more comprehensive clustering effect evaluation.

[0028] Optionally, calculate the silhouette coefficient ψ based on the following formula:

[0029] Wherein, ψ is the silhouette coefficient, and ψ∈[-1,1]; a is the average distance between the target vibration signal data point P in the vibration signal data set and the other vibration signal data points M in the same cluster except the target vibration signal data point P; b is the average distance between the target vibration signal data point P in the vibration signal data set and the other vibration signal data points N in the nearest cluster except the target vibration signal data point P. In this embodiment, the silhouette coefficient provides a quantitative way to evaluate the effect of clustering. It is not just a simple clustering completion mark, but can give the degree of clustering quality. This is particularly useful for optimizing the neighborhood radius ε and the minimum number of points MinPts parameters in the DBSCAN algorithm, which can ensure that the clustering results generated by the algorithm are more accurate and effective. In addition, by calculating the silhouette coefficient of each data point, those data points with poor clustering effects can be identified, which are likely to be abnormal data points or noise. This is particularly important for abnormal detection of vibration signals of submersible electric pumps, because abnormal signals often indicate potential failures of equipment.

[0030] Optionally, when the silhouette coefficient ψ=1, the clustering effect is reasonable and the points within the cluster are dense; when the silhouette coefficient ψ=0, the clustering effect is unreasonable and the vibration signal data points within the cluster are scattered; when the silhouette coefficient ψ=-1, the clustering result is unclear. In an embodiment of the present application, the clustering effect can be quantitatively evaluated by the silhouette coefficient. The closer the silhouette coefficient is to 1, the better the clustering effect, the denser the data points within the cluster, and the higher the separation between clusters. This provides a clear guiding direction for optimizing clustering parameters (such as neighborhood radius and minimum number of points MinPts). Moreover, when the silhouette coefficient is equal to 0, it indicates that the clustering effect is unreasonable and the data points within the cluster are scattered, which helps to timely discover and adjust the clustering strategy and avoid invalid clustering results. In addition, when the silhouette coefficient is -1, it means that the clustering result is unclear, which helps to identify those data points that are difficult to clearly classify, so that further analysis or processing measures can be taken.

[0031] Optionally, the minimum number of points MinPts is set based on the following formula: MinPts = 2*ξ+1, where ξ is the neighborhood radius. In this embodiment, by directly associating the minimum number of points MinPts with the neighborhood radius ε, the setting of MinPts is made more flexible and adaptive. As the neighborhood radius ε changes, MinPts will also be adjusted accordingly, so that it can better adapt to data sets with different densities and distribution characteristics. Moreover, the formula ensures that in the neighborhood of a data point, at least a certain number of other data points are required to be regarded as core points. This setting helps to reduce the impact of noise points on clustering results and improve the accuracy and stability of clustering. In addition, in the traditional DBSCAN algorithm, the neighborhood radius ε and the minimum number of points MinPts are two parameters that need to be set independently. This formula directly associates the setting of MinPts with ε, simplifies the process of parameter setting, and reduces the complexity of the user when using the algorithm. In addition, since the setting of MinPts is based on the neighborhood radius ε, the algorithm has a certain robustness to changes in the density of data points in the data set. Even if the density distribution of the data set is uneven, the algorithm can adapt to this change by adjusting MinPts, thereby obtaining a more reasonable clustering result. Finally, in the DBSCAN algorithm, the clustering process forms clusters by continuously expanding core points. Reasonable MinPts settings help accelerate the convergence process of the algorithm, reduce unnecessary iterations, and improve the execution efficiency of the algorithm.

[0032] Optionally, the comparing the number of vibration signal data points contained in the neighborhood radius ξ of the target vibration signal data point P with the minimum number of points MinPts to obtain a comparison result, and judging whether the target vibration signal data point P is a candidate abnormal data point based on the comparison result, includes: comparing the number of all vibration signal data points X except the target vibration signal data point P within the area defined by the neighborhood radius ξ to which the target vibration signal data point P belongs with the minimum number of points MinPts; if the vibration signal data point X is less than the minimum number of points MinPts, marking the target vibration signal data point P as a noise point, and removing or storing it from the vibration signal data set; if the vibration signal data point X is greater than the minimum number of points MinPts, marking the target vibration signal data point P as a core point to form a core cluster. In this embodiment, by comparing the number of vibration signal data points (X) contained in the neighborhood radius of the target vibration signal data point P with the minimum number of points MinPts, normal data points and potential abnormal data points can be accurately distinguished. When X is less than MinPts, the data point is identified as a noise point, which helps to reduce misjudgments caused by random noise or environmental interference. Moreover, in a data set, noise and outliers are often random and unpredictable. This embodiment reduces the interference of these noise points on subsequent data processing and analysis by treating points that do not meet the density requirements as noise points, and removing or storing them separately from the data set, thereby enhancing the robustness and stability of the entire algorithm. In addition, when X is greater than MinPts, the target vibration signal data point P is marked as a core point, and a core cluster is formed based on these core points. This classification method not only helps to distinguish normal data from abnormal data, but also provides a clearer and more organized data structure for subsequent data analysis and fault warning. In addition, by using the DBSCAN clustering algorithm, the vibration signal data points are divided into different clusters, which can more easily identify clusters that are significantly different from normal data behavior, that is, abnormal data clusters. Further analysis and detection of these abnormal data clusters can quickly locate the potential source of faults in submersible electric pumps and improve the efficiency and accuracy of fault diagnosis. Finally, traditional fault detection methods based on fixed thresholds or preset models are often difficult to adapt to the irregular changes and complex and changeable fault modes of vibration signals of submersible electric pumps. This embodiment, through a density-based clustering method, can more flexibly respond to these challenges, reduce unnecessary false alarms and missed alarms, and improve the reliability of fault warnings.

[0033] The background noise points are vibration signal data points representing that the equipment cannot operate normally. They are signal points generated by vibrations of non-equipment itself, such as environmental factors, sensor noise, and electromagnetic interference. These points may be randomly distributed throughout the signal domain, which may interfere with equipment fault diagnosis and early warning. In this embodiment, by clarifying the definition of background noise points, that is, signal points generated by vibrations of non-equipment itself, such as environmental factors, sensor noise, and electromagnetic interference, it is possible to effectively distinguish between true equipment fault signals and interference signals, thereby improving the accuracy of fault diagnosis. Moreover, background noise points may be randomly distributed throughout the signal domain. If they are not identified and excluded, they may be misjudged as fault signals, resulting in unnecessary false alarms. By identifying and marking these noise points, they can be excluded in subsequent data analysis and fault diagnosis, thereby significantly reducing the false alarm rate.

[0034] Optionally, if the vibration signal data point X is greater than or equal to the minimum number of points MinPts, the target vibration signal data point P is marked as a core point to form a core cluster, including: if the vibration signal data point is greater than or equal to the minimum number of points MinPts, the target vibration signal data point P is taken as the center, and all the vibration signal data points within the radius of ε are taken as the candidate vibration signal data point Q; the target vibration signal data point P is marked as the core point, and a candidate set N is created to classify the target vibration signal data point P and the candidate vibration signal data point Q; the target vibration signal data point P and the candidate vibration signal data point Q are added to the candidate set N, and the candidate vibration signal data point Q is marked as visited; the candidate vibration signal data point Q in the candidate set N is screened to form a core cluster. In this embodiment, by marking the target vibration signal data point P as a core point, and taking it as the center, all the vibration signal data points within the radius of ε are taken as candidate vibration signal data points Q, potential abnormal data points can be identified more accurately. This method reduces the misjudgment caused by the abnormality of a single data point and improves the accuracy of overall anomaly detection. Moreover, by creating a candidate set N and a new cluster C, the target vibration signal data point P and the candidate vibration signal data point Q are classified and screened to form a core cluster, which can more clearly divide the clustering structure of the data points. This structured clustering method helps to better understand the data distribution, so as to more accurately identify abnormal data clusters. In addition, marking the candidate vibration signal data point Q as visited avoids repeated calculations in subsequent processing and improves data processing efficiency. At the same time, by traversing the candidate vibration signal data points Q in the candidate set N for screening until the candidate set N is an empty set, it is ensured that all data points are properly processed and omissions are avoided. In addition, marking the vibration signal data points that are not classified into cluster C as background noise points helps to reduce the interference of noise signals on the anomaly detection results. This method can more accurately identify the signal points generated by the vibration of the equipment itself, thereby improving the reliability of anomaly detection. Finally, the method in this embodiment does not rely on fixed threshold settings or preset models, but performs cluster analysis based on the spatial distribution and density of data points, so it has strong adaptability and flexibility. It can better cope with the irregular changes and complex and changeable fault modes commonly found in the vibration signals of submersible electric pumps.

[0035] Optionally, if the vibration signal data point X is greater than or equal to the minimum number of points MinPts, the target vibration signal data point P is marked as a core point to form a core cluster, and further includes: traversing all the candidate vibration signal data points Q for screening until the candidate set N is an empty set; marking the vibration signal data points that are not included in the cluster C as background noise points. In this embodiment, by traversing all the candidate vibration signal data points Q for screening until the candidate set N is an empty set, this step ensures that the data points are processed comprehensively and systematically. The result of this is that the core points and noise points can be identified more accurately, thereby improving the accuracy of data clustering. Moreover, marking the vibration signal data points that are not included in cluster C as background noise points can effectively distinguish and eliminate those signal points that are generated by non-device vibrations such as environmental factors, sensor noise, and electromagnetic interference. These background noise points may be randomly distributed throughout the signal domain, interfering with equipment fault diagnosis and early warning. By identifying and marking them, the false alarm rate can be reduced and the accuracy of the early warning system can be improved. In addition, through the careful data screening and noise removal process, this embodiment enhances the robustness and reliability of the submersible pump vibration signal early warning method based on density DBSCAN. This enables the system to operate more stably and more accurately detect potential faults of the submersible pump when facing complex and changeable actual working environments. Finally, by removing background noise points and retaining only signal data directly related to equipment vibration, this method provides more accurate and reliable data support for subsequent fault diagnosis. This helps to locate the source of the fault more accurately, improve the efficiency and accuracy of fault diagnosis, and thus provide strong guarantees for the safety and stability of oilfield production.

[0036] Optionally, the screening of the candidate vibration signal data point Q in the candidate set N to form a core cluster includes: screening the candidate vibration signal data point Q in the candidate set N, if the candidate vibration signal data point Q in the candidate set N is the core point, then adding all the unvisited vibration signal data points within the neighborhood radius ξ- of the candidate vibration signal data point Q to the candidate set N to form a core cluster; if the candidate vibration signal data point Q is not the core point, then adding the candidate vibration signal data point Q to the cluster C to form; In this embodiment, by carefully screening the candidate vibration signal data point Q in the candidate set N, the core point and the noise point can be more accurately identified. The data points around the core point are included in the same cluster to form a core cluster, while the noise points are processed separately, avoiding the interference of noise on the clustering results, thereby improving the accuracy and reliability of clustering. Moreover, in the vibration signal analysis of submersible pumps, the presence of noise and interference signals often affects the accuracy of fault identification. Through the method of this embodiment, noise points can be effectively distinguished and eliminated, making the fault signal more prominent, thereby enhancing the ability to identify faults. In addition, since abnormal data clusters can be identified more accurately, the early warning system can respond to the abnormal state of the submersible electric pump more promptly and issue an alarm prompt, which helps to reduce the impact of faults on production and improve the continuity and economic benefits of oil field production. Finally, the method of this embodiment does not rely on fixed threshold settings or preset models, but performs cluster analysis based on the density and distribution characteristics of the data itself. This enables the method to more flexibly deal with the irregular changes and complex and changeable fault modes that are prevalent in the vibration signals of submersible electric pumps.

[0037] Optionally, if the vibration signal data point is greater than the minimum number of points MinPts, the vibration signal data point is marked as a core point to form a cluster, and further includes: assigning the boundary point to the cluster C where the adjacent core point is located. In this embodiment, the boundary points are usually located at the junction of different clusters, and their ownership is unclear. Assigning them to the cluster where the nearest core point is located can more reasonably divide the data points, making the clustering results more accurate. Moreover, the processing of boundary points is crucial to the stability of the clustering algorithm. By assigning boundary points to adjacent core points, the fluctuation of clustering results caused by the uncertainty of the ownership of boundary points can be reduced, and the stability of the clustering algorithm can be improved. In addition, the allocation of boundary points helps to define the boundaries of the cluster more clearly. In density-based clustering algorithms such as DBSCAN, the boundaries of clusters are often composed of boundary points. Reasonable allocation of boundary points to adjacent core points can make the boundaries of clusters clearer, which is conducive to subsequent data analysis and processing. Finally, in the application scenario of vibration signal early warning of submersible pumps, the reasonable processing of boundary points is crucial to the sensitivity of fault detection. Assigning boundary points to adjacent core points can make potential abnormal data points easier to identify, thereby improving the sensitivity and accuracy of fault detection.

[0038] Optionally, the abnormal data cluster is detected to obtain a detection result, and when the detection result meets the set condition, the early warning device alarms to indicate that the submersible electric pump is operating abnormally, including: defining the normal operation set value based on the set early warning parameter information; calculating the abnormal data cluster characteristics based on the abnormal data cluster; comparing the normal operation set value with the abnormal data cluster characteristics to obtain the detection result, and when the detection result meets the set condition, the early warning device alarms to indicate that the submersible electric pump is operating abnormally. In this embodiment, the normal operation set value is defined by setting the early warning parameter information, so that the method can adapt to the specific conditions and working environments of different submersible electric pumps. This means that for submersible electric pumps of different models and different working conditions, accurate abnormality detection can be achieved by adjusting the early warning parameters. Moreover, by comparing the characteristics calculated based on the abnormal data cluster (such as the size, density, center point position, and data point distribution of the cluster) with the normal operation set value, the abnormal state in the operation of the submersible electric pump can be more accurately identified. Compared with the traditional fixed threshold setting or preset model, this method can more accurately capture the irregular changes and complex and changeable fault modes in the vibration signal of the submersible electric pump. In addition, due to the density clustering characteristics of the DBSCAN algorithm itself, the method can effectively process noise and interference signals in the vibration signal. By excluding background noise points from the data, or by calculating the distribution and density characteristics of data points to reduce the impact of noise, the method can improve the accuracy and timeliness of fault diagnosis and reduce the risk of false alarms and missed reports. In addition, this embodiment detects the operating status of the submersible electric pump by comprehensively analyzing multiple parameters (such as cluster size, density, center point position, and data point distribution), and can provide comprehensive fault information. This comprehensive monitoring method helps to quickly locate the source of the fault and improve the efficiency and accuracy of maintenance. Moreover, the vibration signal analysis technology itself has the advantages of non-invasiveness and instant feedback. Combined with the data processing and detection method in this embodiment, real-time monitoring and early warning of the operating status of the submersible electric pump can be achieved. Finally, the submersible electric pump vibration signal early warning method based on density DBSCAN provided in this embodiment is not only applicable to submersible electric pumps, but can also be extended to the field of vibration signal analysis of other similar equipment. At the same time, with the development of technology and changes in demand, the early warning parameters and detection conditions can also be adjusted and optimized accordingly.

[0039] Optionally, the normal operation setting value includes: a normal operation range, a defined threshold value, and an early warning condition. In this embodiment, by setting the normal operation range, the early warning parameters can be flexibly adjusted according to the actual operation and working environment of the submersible pump, so that the early warning system is more in line with the actual situation and the accuracy of the early warning is improved. Moreover, the definition of the threshold value can ensure that when the vibration signal of the submersible pump exceeds the normal range, the early warning system can respond quickly and issue an alarm in time to avoid further deterioration of the fault. In addition, the setting of the early warning condition can be based on a comprehensive judgment of multiple parameters, reduce false alarms caused by abnormal single parameters, and improve the reliability of the early warning. In addition, different submersible pumps, different working environments and working conditions may require different early warning strategies. By setting different normal operation ranges, thresholds and early warning conditions, diverse early warning needs can be met and customized early warnings can be realized. When the early warning system is triggered, by comparing the abnormal data cluster with the normal operation setting value, the fault type and location of the submersible pump can be preliminarily determined, providing strong support for subsequent fault diagnosis and maintenance, and improving maintenance efficiency. Finally, through real-time monitoring and early warning of the vibration signal of the submersible electric pump, potential failure risks can be discovered in a timely manner, providing data support for the formulation of equipment maintenance plans and avoiding the impact of sudden equipment failures on production.

[0040] Optionally, the normal operating range includes: defining a normal size range size_normal; defining a normal density range density_normal; defining a normal center point position range center_normal; and defining a normal vibration signal data point distribution distribution_normal. In this embodiment, by defining the normal operating range in detail, including size, density, center point position, and data point distribution, the abnormality of the vibration signal of the submersible pump can be more accurately identified. These detailed definitions enable the early warning system to more accurately determine whether the vibration signal deviates from the normal range, thereby improving the accuracy and reliability of the early warning.

[0041] Optionally, the threshold value definition includes: defining a threshold value δ_size for size change; defining a threshold value δ_density for density change; defining a threshold value δ_center for center point position change; defining a threshold value δ_distribution for vibration signal data point distribution change. In this embodiment, by defining a threshold value for size change, the abnormal change of amplitude in the vibration signal can be accurately captured, so as to timely identify the increase or decrease of amplitude that may be caused by equipment failure. Defining a threshold value for density change helps to identify changes in the density of the distribution of vibration signals in space, and further analyze whether the operating state of the equipment is stable. Moreover, defining a threshold value for center point position change can keenly capture changes in the center position of the vibration signal, which is particularly important for identifying minor faults or offsets of the equipment, and improves the sensitivity of fault diagnosis. In addition, the threshold definition for changes in the distribution of vibration signal data points can fully reflect the distribution state of the signal in space, and any abnormal distribution may indicate potential problems with the equipment. Finally, by setting various types of thresholds (size, density, center point location, distribution), this embodiment can better adapt to the irregular changes and complex and changeable failure modes commonly found in the vibration signals of submersible electric pumps, thereby improving the flexibility and adaptability of the method.

[0042] Optionally, the characteristics of the abnormal data cluster include: the size of the cluster (C i ), the density of the cluster density (C i ), the center point position of the cluster center(Ci), the distribution of the vibration signal data points of the cluster distribution(C i ). In this embodiment, multiple key characteristics of abnormal data clusters are defined, including cluster size, cluster density, cluster center point location, and distribution of vibration signal data points of the cluster. These characteristics cover different dimensions of data clusters and help to fully understand the nature of abnormal data clusters from multiple perspectives.

[0043] The density of a cluster usually reflects the closeness of the data points within the cluster or the density of the local area. In vibration signal analysis, data points represent the vibration state or characteristics of the equipment at different time points. If the equipment is in normal operation, its vibration signal may form some relatively compact and uniformly dense clusters in the feature space. The density of these clusters reflects the stability and consistency of the equipment vibration to a certain extent. When the equipment fails or is abnormal, its vibration signal tends to change significantly, and these changes may appear as abnormal density of clusters in the feature space. For example, a fault may cause certain vibration features to become more prominent or frequent, thus forming a higher density cluster in the feature space; or a fault may cause the vibration signal to become chaotic, making the originally compact cluster loose or split into multiple small clusters, which is manifested as reduced density or uneven distribution. Specifically, the relationship between the density of clusters and the fault of equipment vibration may be manifested in the following aspects: Abnormal increase in density: If the density of a cluster increases abnormally, it may mean that the equipment has an abnormal concentration on a specific vibration feature, which may be caused by a certain fault mode. For example, bearing damage may cause a significant increase in the vibration signal of a specific frequency, thus forming a high-density cluster in the feature space. Unusually reduced density: Conversely, if the density of a cluster is abnormally reduced, it may indicate that the equipment vibration signal has become more scattered or inconsistent. This may be due to changes in the interaction between multiple parts of the equipment, or due to a fault that causes increased randomness in the vibration signal. Uneven density distribution: In some cases, the density of clusters may be unevenly distributed throughout space. This may mean that the equipment has significant differences in performance on different vibration characteristics, or that the equipment exhibits different vibration characteristics under different operating conditions. This uneven distribution may indicate some potential failure or instability of the equipment.

[0044] The center point of the cluster, center(Ci), is the average value of the coordinates of all data points in the cluster in each dimension. It represents the center or center of gravity of the cluster in multidimensional space. In vibration signal analysis, if the center point of a cluster shifts significantly, it may mean that the operating status of the equipment has changed, especially when the shift is beyond the normal range.

[0045] Optionally, calculate the size of the cluster based on the following formula: size(C i ); size(C i )=[N1,N2,...N d ], where N1, N2, ... N d Each subset C i In this embodiment, the size of the cluster is determined by calculating the total number of data points in each subset (i.e., different groups or regions in the cluster). This method can accurately reflect the amount of data contained in the cluster, providing a solid foundation for subsequent data analysis and fault diagnosis.

[0046] Optionally, the cluster density density (C i ):

[0047] density(C i )=[ρC1,ρC2,...ρC d ], where N i : subcluster; x i : subcluster; C i is the total number of vibration signal data points. In this embodiment, the density of the cluster is calculated by a specific formula, which can accurately quantify the density of the data points in each cluster. This quantification method is more objective and scientific than subjective judgment or fuzzy classification. Moreover, the density of the cluster is an important parameter in the analysis of the vibration signal of the submersible electric pump. The density may reflect the operating status or potential failure of the equipment. By accurately calculating the density, a more accurate basis can be provided for fault diagnosis. In addition, in the submersible electric pump vibration signal early warning method, by comparing the normal operating density range and the actual detected cluster density, it is possible to more accurately determine whether the equipment is in an abnormal state. This helps to reduce false alarms and missed alarms and improve the accuracy and reliability of early warnings.

[0048] Optionally, the center point position center(Ci) of the cluster is calculated based on the following formula:

[0049] Among them, cluster C is composed of d subsets C i Composition, N1, N2, ... N d is the total number of points in each subset. In this embodiment, by dividing the cluster into multiple subsets and considering the total number of points in each subset to calculate the center point position, this method can more accurately reflect the actual center of the cluster. This helps to more accurately locate the core area of ​​the cluster in complex data distribution. Moreover, the method does not rely on a single global center point calculation, but adjusts according to the distribution of the subsets within the cluster. This makes the method more flexible and effective when dealing with clusters with irregular shapes or multiple peaks.

[0050] Optionally, the distribution of the vibration signal data points of the cluster is calculated based on the following formula: i ):

[0051] in, It is C1, C2, ... C dMean of each subset. In this embodiment, the distribution of vibration signal data points within the cluster is evaluated by calculating the mean of each subset, so that the description of the characteristics of the cluster is more quantitative and accurate. This quantitative evaluation helps to understand the structure and characteristics of the cluster more accurately.

[0052] Optionally, the warning conditions include:

[0053]

[0054] The characteristics of the abnormal data cluster meet any of the early warning conditions, triggering the early warning device to act. In this embodiment, the early warning system can flexibly set and adjust the early warning conditions according to different operating states and monitoring requirements of the submersible electric pump. This flexibility enables the system to cope with a variety of complex and changeable fault modes, improving the accuracy and effectiveness of the early warning. Moreover, by setting multiple early warning conditions, the system can comprehensively detect abnormal conditions in the vibration signal of the submersible electric pump, including but not limited to the size, density, center point location and distribution of the cluster. This comprehensive detection method helps to discover potential faults in a timely manner and avoid the expansion of faults. In addition, the spatial clustering analysis based on the density DBSCAN algorithm can effectively distinguish between normal vibration signals and abnormal vibration signals. The DBSCAN algorithm has a certain anti-interference ability for noise and interference signals, which improves the accuracy and reliability of outlier identification. Finally, the system can collect the vibration signal data of the submersible electric pump in real time, and perform automatic processing and analysis. Once an abnormal data cluster that meets the early warning conditions is detected, the system immediately triggers the alarm device to indicate that the submersible electric pump is operating abnormally. This real-time and automated characteristics help to respond to faults quickly, reduce downtime and production losses.

[0055] By comparing the characteristics of a cluster with the normal operating range, we can set early warning conditions to determine whether the equipment is faulty. Specifically: if the size or density of a cluster deviates significantly from the normal range, or the center point position is greatly offset, or the distribution of data points in the cluster changes abnormally, these may indicate potential equipment failure. In order to achieve fault early warning, specific thresholds can be set. Once the characteristics of a cluster are detected to exceed these thresholds, the early warning mechanism is immediately triggered. The early warning mechanism prompts maintenance personnel to conduct further inspections and necessary repairs to prevent production interruptions and economic losses caused by potential equipment failures.

[0056] The vibration signal early warning method of a submersible electric pump based on density DBSCAN provided in the first embodiment of the present application has the following advantages over the traditional early warning method:

[0057] 1. Adaptive clustering capability without presetting the number of clusters

[0058] Breakthrough of traditional limitations: Traditional clustering algorithms such as K-means require users to preset the number of clusters, which becomes impractical in the analysis of submersible pump vibration signals due to signal complexity and uncertainty. Adaptive advantage: The DBSCAN algorithm does not require a preset number of clusters, and can automatically determine the number of clusters based on the density of data points, which is more adaptable to the complexity and variability of submersible pump vibration signals, significantly improving the flexibility and accuracy of clustering.

[0059] 2. Accurate recognition of complex shape clusters

[0060] Complex signal response: The vibration signal of a submersible pump may contain clusters of complex shapes such as concave and horseshoe shapes, which are difficult to effectively identify with traditional methods. Density clustering principle: The DBSCAN algorithm is based on the density clustering principle and can accurately identify these complex clusters, providing a more accurate information basis for fault diagnosis of submersible pumps.

[0061] 3. Intelligent filtering of noisy data

[0062] Noise interference challenge: Noise and interference signals in oilfield operating environments pose a major challenge to vibration signal analysis. Noise identification function: The DBSCAN algorithm can intelligently identify and exclude noise data during the clustering process, ensuring the purity and reliability of the clustering results, and improving the robustness and accuracy of fault diagnosis.

[0063] 4. Pioneering applications and enhanced early warning effectiveness

[0064] New path opening: Creatively applying the DBSCAN algorithm to the detection of outlier values ​​in the vibration signal of submersible electric pumps has opened up a new technical path for health monitoring and fault warning of submersible electric pumps. Improved warning efficiency: Through intelligent identification and processing of outliers in vibration signals, this method significantly improves the accuracy and timeliness of fault warnings, helps prevent potential equipment failures, and ensures safe and stable operation of oilfield production.

[0065] 5. Comprehensive cluster characteristics analysis and comprehensive fault diagnosis

[0066] Multi-dimensional analysis: This method not only focuses on the size and density of the cluster, but also introduces the center point position of the cluster and the distribution of data points within the cluster as analysis objects, realizing a comprehensive analysis of the vibration signal of the submersible pump. Fault mode capture: By comprehensively analyzing the characteristics of the cluster, the abnormal mode in the vibration signal of the submersible pump can be more comprehensively captured, providing richer information support for fault diagnosis.

[0067] The present embodiment also provides a submersible electric pump vibration signal early warning device based on density DBSCAN, comprising: an information acquisition unit: collecting vibration signal data points of the submersible electric pump to obtain a vibration signal data set; a preset unit: setting a neighborhood radius ξ and a minimum number of points MinPts of spatial clustering based on a density DBSCAN algorithm; a comparison unit: based on the vibration signal data set, selecting one of the vibration signal data points that have not been visited as a target vibration signal data point; comparing the number of vibration signal data points within the area defined by the neighborhood radius ξ of the target vibration signal data point with the minimum number of points MinPts to obtain a comparison result; a screening unit: judging whether the target vibration signal data point is a candidate abnormal data point based on the comparison result; traversing the vibration signal data points that have not been visited, executing the comparison unit to generate an abnormal data cluster based on the candidate abnormal data point; a judgment unit: detecting the abnormal data cluster to obtain a detection result, and when the detection result meets the set condition, the early warning device alarms to indicate that the submersible electric pump is operating abnormally. In this embodiment, by performing spatial clustering through the DBSCAN algorithm, abnormal data points in the vibration signal can be quickly and accurately identified, and abnormal data clusters can be generated. This helps to achieve real-time monitoring and early warning of the operating status of the submersible electric pump, and improves the efficiency and timeliness of fault diagnosis. Moreover, compared with the traditional anomaly detection method based on a fixed threshold or a preset model, the DBSCAN algorithm can more flexibly handle irregular changes and complex and changeable fault modes in the vibration signal, thereby improving the adaptability and accuracy of the system. In addition, by setting a reasonable neighborhood radius and a minimum number of points MinPts, the device can effectively distinguish between noise points and actual abnormal data points, reduce the influence of noise and interference signals on the identification of abnormal values, and reduce the false alarm rate and missed alarm rate. In addition, by collecting and analyzing the vibration signal data set of the submersible electric pump, the device can fully reflect the operating status of the equipment and provide more comprehensive data support for fault diagnosis and early warning. Finally, the entire early warning process is automated and intelligent, reducing manual intervention and improving the stability and reliability of the system. At the same time, the algorithm based on density clustering can automatically learn the distribution characteristics of the data, continuously optimize the early warning model, and improve the accuracy of the early warning.

[0068] This embodiment also provides a submersible electric pump vibration signal warning device based on density DBSCAN, including: a memory and a processor, the memory stores a computer executable program, and the processor is used to execute the computer executable program to implement any method described in Example 1.

[0069] This embodiment further provides a submersible electric pump vibration signal warning device based on density DBSCAN, wherein a computer executable program is stored on the computer storage medium, and when the computer executable program is executed, any method described in the first embodiment is implemented.

[0070] It should be noted that the same and similar parts between the various embodiments in this specification can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the device and system embodiments, since they are basically similar to the method embodiments, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiments. The device and system embodiments described above are merely schematic, wherein the modules described as separate components may or may not be physically separated, and the components indicated as modules may or may not be physical modules, that is, they may be located in one place, or they may be distributed on multiple network modules. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. A person of ordinary skill in the art can understand and implement it without paying creative labor.

[0071] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed in the present application should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A vibration signal early warning method for submersible electric pumps based on density DBSCAN, It is characterized in that include: S101: collecting vibration signal data points of the submersible electric pump to obtain a vibration signal data set; S2 01: Set the neighborhood radius ξ and the minimum number of points MinPts for spatial clustering based on the density DBSCAN algorithm; S301: Based on the vibration signal data set, selecting one of the vibration signal data points that has not been accessed as a target vibration signal data point; S401: comparing the number of vibration signal data points within the neighborhood radius ξ of the target vibration signal data point with the minimum number of points MinPts to obtain a comparison result; S501: Determine whether the target vibration signal data point is a candidate abnormal data point based on the comparison result; S601: traverse the vibration signal data points that have not been visited, and execute steps S301 to S501, generating an abnormal data cluster based on the candidate abnormal data points; S701: Detect the abnormal data cluster to obtain a detection result. When the detection result meets the set conditions, the early warning device alarms to indicate that the submersible electric pump is operating abnormally.

2. The submersible pump vibration signal early warning method based on density DBSCAN according to claim 1, wherein the neighborhood radius ξ of the set spatial clustering is include: Based on the k-nearest neighbor graph method, the number k of each vibration signal data point and its nearest neighbor points in the vibration signal data set is found, so as to set the neighborhood radius ξ of spatial clustering based on the k value.

3. The submersible pump vibration signal early warning method based on density DBSCAN according to claim 2, wherein the number k of each vibration signal data point and its nearest neighbor points in the vibration signal data set is found based on the k-nearest neighbor graph method, so as to set the neighborhood radius ξ of spatial clustering based on the k value, include: Find the k nearest neighbor points of the vibration signal data point based on the k-nearest neighbor graph method; Calculating the distance from each vibration signal data point to the kth nearest neighbor point based on the silhouette coefficient method to obtain a silhouette coefficient ψ; Traversing the possible range of k values, calculating the k value that maximizes the average value of the silhouette coefficient ψ, so as to obtain the optimized k value; The neighborhood radius ξ is set based on the optimized k value.

4. According to claim 3, the vibration signal early warning method of the submersible electric pump based on density DBSCAN, It is characterized in that The step of calculating the distance from each vibration signal data point to the kth nearest neighbor point based on the silhouette coefficient method to obtain the silhouette coefficient ψ comprises: Calculate the average distance a between the target vibration signal data point P in the vibration signal data set and the other vibration signal data points M in the same cluster except the target vibration signal data point P based on the silhouette coefficient method; Calculate the average distance b between the target vibration signal data point P in the vibration signal data set and the other vibration signal data points N in the nearest cluster except the target vibration signal data point P based on the silhouette coefficient method; Based on the average distance a and the average distance b, a silhouette coefficient ψ is calculated.

5. According to the density DBSCAN-based early warning method for vibration signals of submersible electric pumps according to claim 2, the minimum number of points MinPts is set based on the following formula: MinPts=2*ξ+1 in, ξ is the neighborhood radius.

6. According to the density DBSCAN-based early warning method for submersible pump vibration signals in claim 1, the comparison is performed by comparing the number of vibration signal data points contained in the neighborhood radius ξ of the target vibration signal data point P with the minimum number of points MinPts to obtain a comparison result. include: The relationship between the number of all the vibration signal data points X except the target vibration signal data point P within the area defined by the neighborhood radius ξ to which the target vibration signal data point P belongs is compared with the minimum number of points MinPts.

7. According to the density DBSCAN-based early warning method for vibration signals of submersible electric pumps as described in claim 1, the abnormal data cluster is detected to obtain a detection result, and when the detection result meets the set conditions, the early warning device alarms to indicate that the submersible electric pump is operating abnormally. include: Defining the normal operation setting value based on the set warning parameter information; Calculating the abnormal data cluster characteristics based on the abnormal data cluster; The normal operation setting value is compared with the abnormal data cluster characteristic to obtain a detection result. When the detection result meets the set condition, the early warning device alarms to indicate that the submersible electric pump is operating abnormally.

8. A vibration signal early warning device for submersible electric pump based on density DBSCAN, It is characterized in that include: Information acquisition unit: collects vibration signal data points of the submersible electric pump to obtain a vibration signal data set; Preset unit: Set the neighborhood radius ξ and the minimum number of points MinPts of spatial clustering based on the density DBSCAN algorithm; Comparison unit: based on the vibration signal data set, select one of the vibration signal data points that has not been visited as a target vibration signal data point; compare the number of vibration signal data points within the neighborhood radius ξ of the target vibration signal data point with the minimum number of points MinPts to obtain a comparison result; A screening unit: judging whether the target vibration signal data point is a candidate abnormal data point based on the comparison result; traversing the vibration signal data points that have not been visited, and executing the comparison unit to generate an abnormal data cluster based on the candidate abnormal data point; The judgment unit detects the abnormal data cluster to obtain a detection result. When the detection result meets the set conditions, the early warning device alarms to indicate that the submersible electric pump is operating abnormally.

9. An electronic device, It is characterized in that include: A memory and a processor, wherein the memory stores a computer executable program, and the processor is used to execute the computer executable program to implement the method according to any one of claims 1 to 7.

10. A computer storage medium, It is characterized in that The computer storage medium stores a computer executable program, and the computer executable program implements the method according to any one of claims 1 to 7 when executed.

Citation Information

Cited By

  • Intelligent gas meter data acquisition system based on machine learning

    CN120180165A

  • Cloud collaborative data exception analysis method and system based on job node information

    CN120832630A

  • Cloud collaboration data anomaly analysis method and system based on job node information

    CN120832630B

  • Method, device and equipment for diagnosing and monitoring vibration of power transmission system of unmanned AGV (Automatic Guided Vehicle)

    CN121256541A

  • Unmanned AGV power transmission system vibration diagnosis monitoring method, device and equipment

    CN121256541B