Pipeline sealing intelligent monitoring method and system

Abstract

The invention relates to the technical field of data processing, in particular to an intelligent monitoring method and system for pipeline sealing performance, and the method comprises the steps: obtaining a pressure data sequence in a pipeline, carrying out the iterative clustering of the pressure data sequence through employing a DIANA algorithm, calculating the stability index of a clustering cluster after the clustering of any time is completed, and calculating the sealing performance of the pipeline. When the stability index is larger than a preset threshold value, iteration is stopped, clustering results are obtained, and the pipeline sealing performance is monitored according to a comparison result of the data size of the minimum clustering cluster in the clustering results and a preset number threshold value. According to the invention, the accuracy of the obtained clustering result can be improved, so that the sealing condition of the pipeline can be accurately monitored.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and more specifically, to a pipeline sealing intelligent monitoring method and system. Background Art

[0002] Pipelines are widely used in industries such as oil, gas, chemicals and water treatment to transport liquids and gases. Once a leak occurs, it may not only lead to a waste of resources, but also cause serious environmental pollution and even cause safety accidents. Through intelligent monitoring of pipeline sealing, potential leakage risks can be discovered and handled in a timely manner, thereby reducing the probability of accidents. Therefore, how to accurately monitor the sealing of pipelines is crucial.

[0003] In the related technology, for example, a patent application document with publication number CN111639711A discloses a method for monitoring oil pipeline leakage based on pressure monitoring time series data. The method includes: constructing multiple indicators to pre-process the data, and combining a type of support vector machine to model the sample data to identify pressure anomalies, using a hierarchical clustering algorithm to cluster all abnormal time series, and extracting leakage anomaly patterns, thereby determining the leakage of the pipeline based on the extracted abnormal patterns.

[0004] However, the above scheme uses multiple statistical indicators to preprocess the data, which has a huge amount of calculation and does not consider the influence of noise, resulting in invalid data in the sample set, making it impossible to accurately train the vector machine. The accuracy of the vector machine training results not only depends on the quality of the sample set, but also is affected by various parameters such as kernel function and penalty parameter. When using the hierarchical clustering algorithm to determine the abnormal clusters, it relies on the trained hierarchical clustering model, which may lead to deviations in the identified abnormal patterns, thereby affecting the monitoring of pipeline leakage. Summary of the invention

[0005] In order to solve the problem of inaccurate monitoring of pipeline leakage, the present invention provides a pipeline sealing intelligent monitoring method and system.

[0006] According to a first aspect of the present invention, there is provided a pipeline sealing intelligent monitoring method, comprising: Obtain the pressure data sequence in the pipeline; The DIANA algorithm is used to iteratively cluster the pressure data sequence, and the stability index of the cluster after any clustering is completed is calculated. When the stability index is greater than the preset threshold, the iteration is stopped to obtain the clustering result, so as to monitor the pipeline sealing according to the comparison result of the data volume of the smallest cluster in the clustering result and the preset quantity threshold; After any clustering is completed, the process of obtaining the stability index of the cluster is as follows: The modulus of the vector composed of the discrete degree of any cluster and the range of the data points in the cluster is used as the cluster feature of the cluster. The average difference of the cluster features of the same cluster after the current clustering and the previous clustering is calculated to obtain the intra-cluster change degree of the cluster. Calculate the separation between any two clusters. The separation is positively correlated with the difference in cluster characteristics between the two clusters and the distance between the two clusters. Obtain the difference in the average separation between the clusters after the current clustering and the previous clustering, and obtain the degree of inter-cluster change of the clusters. The stability index is calculated, which is positively correlated with the degree of change between clusters and negatively correlated with the degree of change within clusters.

[0007] When the present invention utilizes the DIANA algorithm to iteratively cluster the pressure data sequence, the determination of the iteration termination condition depends on the stability index of the obtained clustering clusters, and the stability index can characterize the change characteristics between the clustering clusters obtained after the clustering is completed, as well as the change characteristics within the clustering clusters, thereby ensuring the accuracy of the clustering results, so that abnormal data can be accurately identified and the sealing condition of the pipeline can be accurately monitored; and when the algorithm is used to identify abnormal data in the pressure data sequence, it relies on the iterative clustering method of the algorithm itself, can adapt to noise and abnormal data, and to a certain extent ensure the accuracy of the clustering results.

[0008] Preferably, the separation degree satisfies the following relationship: ; In the formula, For the After the clustering is completed The clusters and The separation between clusters; For the After the clustering is completed The cluster characteristics of the clusters; For the After the clustering is completed The cluster characteristics of the clusters; For the After the clustering is completed Clusters; For the After the clustering is completed Clusters; The symbol for distance.

[0009] The present invention combines the distance and the cluster characteristics to calculate the separation degree, thereby ensuring the accuracy of the calculated result.

[0010] Preferably, the method for obtaining the degree of variation among clusters includes: The average value of the separation between any two clusters after any clustering is completed is taken as the first average value, and the average value of the separation between any two clusters after the last clustering is completed is taken as the second average value; The absolute value of the difference between the first average value and the second average value is calculated to obtain the degree of inter-cluster change of the clusters after any clustering is completed.

[0011] The present invention can measure the dispersion of clusters after each clustering is completed, and judge whether there are overlapping or fuzzy areas between the clusters obtained after each clustering is completed, thereby providing a data basis for the iterative termination conditions of clustering.

[0012] Preferably, the method for obtaining the discrete degree includes: Get the cluster center of any cluster after any clustering is completed, calculate the average value of the distance between each data point in any cluster and the cluster center, and get the degree of dispersion of any cluster.

[0013] Preferably, the range is a comprehensive range, and a method for obtaining the comprehensive range includes: Get the maximum and minimum values ​​of the data points in any cluster after any clustering is completed, calculate the difference between the maximum and minimum values, and use the ratio of the difference to the number of data points in any cluster as the comprehensive range of the data points in any cluster.

[0014] Preferably, the stability index satisfies the following relationship: ; In the formula, For the The stability index of the cluster after the clustering is completed; For the The degree of change within the cluster after the clustering is completed; For the The degree of change among clusters after the clustering is completed; The natural constant An exponential function with base .

[0015] The present invention can reduce the amount of calculation and the difficulty of calculation by limiting the range of the stability index to 0-1.

[0016] Preferably, monitoring the pipeline sealing performance according to the comparison result between the data volume of the smallest cluster in the clustering result and a preset quantity threshold value includes: Get the preset quantity threshold; When the data volume of the smallest cluster in the clustering result is less than the quantity threshold, it is determined that there is abnormal data in the pressure data sequence, and an alarm is issued to remind the operator that there is a sealing defect in the pipeline.

[0017] The present invention can accurately detect the sealing defects of the pipeline and realize accurate monitoring of the sealing condition of the pipeline.

[0018] According to a second aspect of the present invention, a pipeline sealing intelligent monitoring system is provided. The system includes a memory and a processor. The memory stores a computer program. The processor executes the computer program to implement the steps of the first aspect of the present invention.

[0019] The present invention has the following effects: 1. The present invention improves the method for determining the iterative termination condition of the DIANA algorithm. By calculating the stability index of the clustering cluster after clustering is completed, it is determined whether to terminate the clustering process. It can adapt to the dynamic changes of the pressure data in the pipeline, obtain a clustering result with a higher clustering effect, and improve the accuracy of the clustering result.

[0020] 2. The present invention calculates the degree of intra-cluster variation and inter-cluster variation of clusters, which can measure the stability of the internal structure of each category of clusters and the dispersion between clusters. It evaluates the clustering effect of each clustering from multiple aspects and ensures the accuracy of the determined stability index.

[0021] 3. Based on the clustering results with high accuracy, the present invention can accurately identify abnormal data and realize accurate monitoring of the sealing condition of the pipeline. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] By reading the following detailed description with reference to the accompanying drawings, the above and other objects, features and advantages of the exemplary embodiments of the present invention will become readily understood. In the accompanying drawings, several embodiments of the present invention are shown in an exemplary and non-limiting manner, and the same or corresponding reference numerals represent the same or corresponding parts, wherein: Figure 1 It is a schematic diagram of the steps of a pipeline sealing intelligent monitoring method according to an embodiment of the present invention. DETAILED DESCRIPTION

[0023] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0024] The specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0025] Reference Figure 1 , a pipeline sealing intelligent monitoring method, including steps S1 to S3, specifically as follows: S1: Get the pressure data sequence in the pipeline.

[0026] Specifically, a pressure sensor can be installed at an appropriate position in the pipeline, such as before and after a valve or at a pipe elbow. During a sampling period, such as 15 minutes, the readings of the pressure sensor are collected at a certain collection frequency, such as 1 Hz, to obtain a pressure data sequence.

[0027] S2: Use the DIANA algorithm to iteratively cluster the pressure data sequence, calculate the stability index of the cluster after any clustering is completed, and stop the iteration when the stability index is greater than the preset threshold to obtain the clustering result.

[0028] It should be noted that when the DIANA (Divisive Analysis, hierarchical clustering) algorithm uses a bottom-up strategy to iteratively cluster sample data, there is no need to set the number of clusters in advance, and the algorithm can identify abnormal data when there is noise in the sample data. Therefore, the algorithm can be used to identify whether there is abnormal data in the pressure data sequence, so as to monitor the sealing condition of the pipeline based on the identification results.

[0029] It should be further explained that the algorithm iteration termination condition is generally that the number of clusters reaches the expected number, or the diameter of the cluster is less than the set threshold, or the distance between the clusters exceeds the set threshold, etc. However, since the pressure data in the pipeline is usually dynamically changing and susceptible to various factors, such as flow, temperature, and external environment, the traditional iteration termination condition may not be able to adapt to this change, thereby reducing the accuracy of the clustering results. Therefore, the present invention improves the determination of the algorithm iteration termination condition, by calculating the stability index of the cluster obtained after clustering is completed, judging whether to terminate the iteration, so that the iteration termination condition can adapt to the change of pressure data, thereby ensuring the accuracy of the determined clustering results.

[0030] Specifically, calculating the stability index of the cluster after any clustering is completed includes the following steps: Step 1: Take the modulus of the vector composed of the discrete degree of any cluster and the extreme value of the data points in the cluster as the cluster feature of the cluster, calculate the average difference of the cluster features of the same cluster after the current clustering and the previous clustering, and obtain the degree of intra-cluster change of the cluster; Among them, the degree of dispersion reflects the distance between data points in the cluster; the range reflects the distribution range of the data point values ​​in the cluster, which can characterize the diversity of samples in the cluster; the degree of variation within the cluster reflects the similarity between data points in the cluster of the same category. For example, after multiple iterations, if the changes in the cluster characteristics of the clusters of the same category are small, it means that the similarity of the data points in the clusters of this category is high, and the internal structure is relatively consistent, then the clustering results are more reliable.

[0031] In an exemplary embodiment of the present invention, the determination of the discreteness of any cluster can be achieved by the following steps: Get the cluster center of any cluster after any clustering is completed, calculate the average value of the distance between each data point in any cluster and the cluster center, and get the degree of dispersion of any cluster.

[0032] Optionally, when the overall distance between each data point in any cluster and the cluster center of the cluster is large, it means that the data points in the cluster are distributed more dispersedly, and the corresponding discrete degree of the cluster is large.

[0033] In an exemplary embodiment of the present invention, the range of the data points in the cluster is a comprehensive range, and the comprehensive range can be determined by the following steps: Get the maximum and minimum values ​​of the data points in any cluster after any clustering is completed, calculate the difference between the maximum and minimum values, and use the ratio of the difference to the number of data points in any cluster as the comprehensive range of the data points in any cluster.

[0034] Specifically, after any iteration, the comprehensive range of the data points in any cluster satisfies the following relationship: ; In the formula, For the After the clustering is completed The combined range of the data points in the clusters; The maximum value of the data point in the cluster; The minimum value of the data points in the cluster; is the number of data points in the cluster.

[0035] in, It reflects the distribution range of the data point values ​​in the cluster. The larger the value, the larger the distribution range of the data point values ​​in the cluster. When the number of data points in the cluster is small, it means that the differences between the data points in the cluster are large.

[0036] Furthermore, the discrete degree of each cluster after any clustering is completed and the modulus of the two-dimensional vector composed of the comprehensive range of the data points in the corresponding cluster can be used as the cluster characteristics of the corresponding cluster, so that the cluster characteristics of each cluster after each clustering is completed can be obtained.

[0037] Furthermore, the changes in the cluster characteristics of each cluster after each clustering is completed can be used to evaluate the similarity between data points in clusters of the same category, thereby obtaining the degree of intra-cluster change of the cluster after each clustering is completed.

[0038] Specifically, after any clustering is completed, the degree of change within the cluster satisfies the following relationship: ; In the formula, For the The degree of change within the cluster after the clustering is completed; For the After the clustering is completed The cluster characteristics of the clusters; For the After the clustering is completed The cluster characteristics of the clusters; is the absolute value symbol; For the The number of clusters after the clustering is completed; is the summation symbol.

[0039] in, Reflects the The cluster characteristics of each cluster after the first clustering are The overall difference between the cluster characteristics of the corresponding clusters after the first clustering is completed. The larger the value, the After the clustering is completed, the similarity of the data points in each cluster is low, and the corresponding intra-cluster variation degree of the cluster is large.

[0040] Step 2: Calculate the separation between any two clusters. The separation is positively correlated with the difference in cluster characteristics between the two clusters and the distance between the two clusters. Obtain the difference in the average separation between the clusters after the current clustering and the last clustering, and obtain the degree of inter-cluster change of the clusters. The separation degree reflects the separation between clusters and can reflect the clustering effect of clusters. For example, when the separation degree between clusters is high after any clustering is completed, it means that there is no cluster overlap between clusters, and the clustering effect is good.

[0041] The degree of change between clusters reflects the change in the separation between clusters obtained after each clustering is completed. The smaller the value, the smaller the change in the separation between clusters. There may be overlapping or fuzzy areas between clusters. If the iteration is stopped at this time, clusters of different categories cannot be effectively distinguished, resulting in poor clustering effect.

[0042] Specifically, after any clustering is completed, the separation degree between any two clusters satisfies the following relationship: ; In the formula, For the After the clustering is completed The clusters and The separation between clusters; For the After the clustering is completed The cluster characteristics of the clusters; For the After the clustering is completed The cluster characteristics of the clusters; For the After the clustering is completed Clusters; For the After the clustering is completed Clusters; The symbol for distance.

[0043] In another embodiment, the separation degree between any two clusters may also be calculated by other calculation formulas, such as by not using the arithmetic square root calculation formula but by summing up the difference in cluster characteristics of any two clusters and the distance between the two clusters and normalizing the calculation formula.

[0044] In an exemplary embodiment of the present invention, the following steps can be used to determine the degree of change between clusters after each clustering is completed: The average value of the separation between any two clusters after any clustering is completed is taken as the first average value, and the average value of the separation between any two clusters after the previous clustering is completed is taken as the second average value; the absolute value of the difference between the first average value and the second average value is calculated to obtain the degree of inter-cluster change of the clusters after any clustering is completed.

[0045] Optionally, when the difference between the first average value and the second average value is larger, it means that in the iterative clustering process, the clusters of different categories are more dispersed, and the clusters of different categories are less likely to overlap, and the corresponding clustering effect is better.

[0046] Step 3: Calculate the stability index, which is positively correlated with the degree of change between clusters and negatively correlated with the degree of change within clusters.

[0047] Specifically, after any clustering is completed, the stability index of the cluster satisfies the following relationship: ; In the formula, For the The stability index of the cluster after the clustering is completed; For the The degree of change within the cluster after the clustering is completed; For the The degree of change among clusters after the clustering is completed; The natural constant An exponential function with base .

[0048] Optionally, when the degree of variation within a cluster is small after any clustering is completed, it means that the internal structures of the clusters are relatively consistent, and the corresponding stability index is relatively large; when the degree of variation between clusters is large, it means that the clusters of different categories are more dispersed, the clustering effect is better, and the corresponding stability index is relatively large.

[0049] In another embodiment, the stability index may be calculated by other calculation formulas, such as not using a proportional relationship to characterize the correlation between the stability index and the degree of inter-cluster change and the degree of intra-cluster change, but using a difference calculation formula to characterize the correlation.

[0050] Further, when the stability index of the cluster after any clustering is greater than a preset threshold, such as 0.73, the iteration process can be stopped to obtain the clustering result. It should be noted that the process of iterative clustering using the DIANA algorithm is a prior art and is not described in detail in this embodiment.

[0051] S3: Monitoring the pipeline sealing according to the comparison result between the data amount of the smallest cluster in the clustering result and the preset quantity threshold.

[0052] In an exemplary embodiment of the present invention, the pipeline sealing performance can be monitored by the following steps: A preset quantity threshold is obtained; when the data amount of the smallest cluster in the clustering result is less than the quantity threshold, it is determined that there is abnormal data in the pressure data sequence, and an alarm is issued to remind the operator that there is a sealing defect in the pipeline.

[0053] It should be noted that DIANA can identify abnormal data and cluster the abnormal data into a separate category. In a stable pipeline system, the number of abnormal pressure data in the pipeline is relatively small. Therefore, the amount of data in the smallest cluster in the clustering result can be measured to evaluate whether there is abnormal data in the pressure data sequence.

[0054] Optionally, the quantity threshold can be set to 3. If the data volume of the smallest cluster in the clustering result is less than 3, it indicates that there are abnormal data in the pressure data sequence, and an alarm can be issued to remind the operator that there are defects in the sealing of the pipeline, so that corresponding measures can be taken in time to repair the sealing defects of the pipeline and ensure the safe use of the pipeline. The quantity threshold set in this embodiment can be set according to the data volume of the processed sample. When the data volume of the sample is large, a larger quantity threshold can be set. This embodiment does not specifically limit the setting of the quantity threshold.

[0055] The present invention also provides a pipeline sealing intelligent monitoring system, the system includes a memory and a processor, and a computer program is stored in the memory, the computer program integrates the function of a pipeline sealing intelligent monitoring method, when the computer program is executed, the accuracy of the obtained clustering results can be improved through a pipeline sealing intelligent monitoring method, so that the sealing condition of the pipeline can be accurately monitored.

[0056] In the description of this specification, "plurality" or "several" means at least two, such as two, three or more, etc., unless otherwise clearly and specifically defined.

[0057] Although this specification has shown and described a number of embodiments of the present invention, it will be apparent to those skilled in the art that such embodiments are provided by way of example only. Those skilled in the art will conceive of many modifications, changes and alternatives without departing from the ideas and spirit of the present invention. It should be understood that in the practice of the present invention, various alternatives to the embodiments of the present invention described herein may be employed.

Claims

1. A pipeline sealing intelligent monitoring method, characterized in that: include: Obtain the pressure data sequence in the pipeline; Iteratively clustering the pressure data sequence using the DIANA algorithm, calculating the stability index of the cluster after any clustering is completed, and stopping the iteration when the stability index is greater than a preset threshold to obtain a clustering result, so as to monitor the pipeline sealing according to the comparison result between the data amount of the smallest cluster in the clustering result and the preset quantity threshold; After any clustering is completed, the process of obtaining the stability index of the cluster is as follows: The modulus of the vector composed of the discrete degree of any cluster and the range of the data points in the cluster is used as the cluster feature of the cluster. The average difference of the cluster features of the same cluster after the current clustering and the previous clustering is calculated to obtain the intra-cluster change degree of the cluster. Calculate the separation between any two clusters. The separation is positively correlated with the difference in cluster characteristics between the two clusters and the distance between the two clusters. Obtain the difference in the average separation between the clusters after the current clustering and the last clustering, and obtain the degree of inter-cluster change of the clusters. A stability index is calculated, the stability index being positively correlated with the degree of variation between clusters and negatively correlated with the degree of variation within clusters.

2. A pipeline sealing intelligent monitoring method according to claim 1, characterized in that: The separation degree satisfies the following relationship: ； In the formula, For the After the clustering is completed, The clusters and The separation between clusters; For the After the clustering is completed, The cluster characteristics of the clusters; For the After the clustering is completed, The cluster characteristics of the clusters; For the After the clustering is completed, Clusters; For the After the clustering is completed, Clusters; The symbol for distance.

3. A pipeline sealing intelligent monitoring method according to claim 2, characterized in that: The method for obtaining the degree of inter-cluster variation includes: The average value of the separation between any two clusters after any clustering is completed is taken as the first average value, and the average value of the separation between any two clusters after the last clustering is completed is taken as the second average value; The absolute value of the difference between the first average value and the second average value is calculated to obtain the degree of inter-cluster change of the clusters after any clustering is completed.

4. The pipeline sealing intelligent monitoring method according to claim 1 is characterized in that: The method for obtaining the discrete degree includes: The cluster center of any cluster after any clustering is completed is obtained, and the average value of the distance between each data point in any cluster and the cluster center is calculated to obtain the degree of dispersion of any cluster.

5. The pipeline sealing intelligent monitoring method according to claim 1 is characterized in that: The range is a comprehensive range, and the method for obtaining the comprehensive range includes: After any clustering is completed, the maximum and minimum values ​​of the data points in any clustering cluster are obtained, the difference between the maximum and the minimum is calculated, and the ratio of the difference to the number of data points in any clustering cluster is used as the comprehensive range of the data points in any clustering cluster.

6. The pipeline sealing intelligent monitoring method according to claim 1 is characterized in that: The stability index satisfies the following relationship: ； In the formula, For the The stability index of the cluster after the clustering is completed; For the The degree of change within the cluster after the clustering is completed; For the The degree of change among clusters after the clustering is completed; The natural constant An exponential function with base .

7. The pipeline sealing intelligent monitoring method according to claim 1 is characterized in that: The monitoring of pipeline sealing according to the comparison result between the data amount of the smallest cluster in the clustering result and the preset quantity threshold comprises: Get the preset quantity threshold; When the data amount of the smallest cluster in the clustering result is less than the quantity threshold, it is determined that abnormal data exists in the pressure data sequence, and an alarm is issued to remind the operator that there is a sealing defect in the pipeline.

8. A pipeline sealing intelligent monitoring system, characterized in that: The pipeline sealing intelligent monitoring system comprises a memory and a processor, wherein a computer program is stored in the memory, and the processor executes the computer program to implement the steps of the method according to any one of claims 1 to 7.

Images