CN-122015022-A - Multi-stage clustering method and system for positioning leakage points

CN122015022ACN 122015022 ACN122015022 ACN 122015022ACN-122015022-A

Abstract

The invention discloses a pipeline leakage positioning result clustering method and system based on a distance threshold and mode-median, and belongs to the technical field of pipeline leakage detection and positioning. The method comprises the steps of obtaining suspected leakage point data sequences, constructing candidate subsets based on distance threshold values, selecting a maximum subset as initial clusters through greedy iteration, selecting mode priority and median alternative calculation center positions, merging adjacent clusters based on the same threshold values, intelligently screening according to pipeline length and data quantity threshold values, and reserving leakage points with sufficient evidence in an effective range and outside the boundary. The invention can accurately separate different leakage points, provide stable center position estimation, combine scattered subgroups, avoid the false deletion of edge leakage points, and remarkably improve positioning accuracy, integrity and practicability.

Inventors

LU LU
LU JIANGBO
ZHU RUI
LI SHOULIANG
SONG CHAO
HUANG XIAOLEI
DU HONGTAO
CHEN HAO
YUAN HONGWEI

Assignees

山东普赛通信科技股份有限公司

Dates

Publication Date: 20260512
Application Date: 20260228

Claims (10)

1. A multi-stage clustering method for leak location, comprising: Acquiring position data of a suspected leakage point of a pipeline section, and sequencing the position data according to the value to form an initial data sequence; Traversing each data point in the initial data sequence, and constructing a candidate subset based on a preset neighborhood distance threshold; calculating the central position of each candidate subset, and generating an initial clustering result set based on the central position; Based on the neighborhood distance threshold, merging and optimizing adjacent clusters in the initial clustering result set to obtain an optimized clustering result; and sequencing the optimized clustering results according to the numerical value of the central position to obtain positioning results.
2. The multi-stage clustering method for positioning leakage points according to claim 1 is characterized by comprising the steps of carrying out acoustic measurement on the same pipeline section for a plurality of times by utilizing sensors at two ends of the pipeline in a preset time period, collecting sound signals in the pipeline, and obtaining a series of position data of the leakage suspected points by calculating the time difference of the leakage sound signals reaching the sensors at two ends.
3. The method for multi-stage clustering of leak source localization according to claim 1, wherein the constructing a plurality of candidate subsets based on a preset neighborhood distance threshold includes centering each data point, and classifying all data points in the data sequence having a distance from the data point less than or equal to the neighborhood distance threshold into the candidate subset corresponding to the data point.
4. The multi-stage clustering method for leak location according to claim 1, wherein the calculating the center position of each candidate subset generates an initial set of clustering results based on the center positions, specifically comprising: selecting a candidate subset with the largest number of data points from all candidate subsets as an initial cluster; Counting the occurrence frequency of all data points in the initial cluster, if a unique mode exists, taking the mode as the central position of the cluster, otherwise, taking the median of all the data points in the cluster as the central position; removing all data points contained in the initial cluster from the current data sequence; repeating the steps until all candidate subsets are processed, and obtaining an initial clustering result set.
5. The multi-stage clustering processing method for positioning leakage points according to claim 1 is characterized in that the combining optimization processing is carried out on adjacent clusters, and specifically comprises traversing an initial cluster result set, judging whether the distance between the central positions of any two clusters adjacent to each other is smaller than the neighborhood distance threshold value or not, if so, combining the two clusters into a new cluster, and recalculating the central position of the new cluster.
6. The multi-stage clustering method for leak location in accordance with claim 1, further comprising a cluster screening step prior to said outputting of the location result: and judging whether the central position of each optimized cluster is positioned in the effective length range of the pipeline section, and if so, reserving the clustering result.
7. The multi-stage clustering method for leak location according to claim 6, wherein if the central position of the cluster exceeds the effective length range of the pipeline segment, further judging whether the number of data points contained in the cluster exceeds a preset number threshold, and if so, retaining the clustering result.
8. A multi-stage clustering process for leak-point localization as claimed in claim 6 or claim 7 wherein the effective length range of the pipe section refers to the actual physical length of the pipe.
9. A multi-stage cluster processing system for leak location, comprising: the data acquisition module is configured to acquire position data of suspected leakage points of the pipeline section, and sequence the position data according to the numerical value to form a data sequence; a candidate subset construction module configured to traverse each data point in the data sequence and construct a plurality of candidate subsets based on a preset neighborhood distance threshold; The initial clustering module is configured to calculate the central position of each candidate subset to obtain an initial clustering result set; The merging and optimizing module is configured to merge and optimize adjacent clusters in the initial clustering result set based on the neighborhood distance threshold value to obtain an optimized clustering result; and the result output module is configured to sort the optimized clustering results according to the numerical value of the central position and output positioning results.
10. An electronic device comprising one or more processors, memory to store one or more programs, which when executed by the one or more processors, cause the electronic device to implement a multi-stage clustering method for leak location as claimed in any one of claims 1-8.

Description

Multi-stage clustering method and system for positioning leakage points Technical Field The invention relates to the technical field of pipeline leakage positioning, in particular to a multi-stage clustering processing method and system for positioning leakage points. Background Water pipelines are an important component of modern industrial and urban infrastructure, and their safe operation is directly related to resource utilization efficiency and public safety. In the technology of positioning the leakage point of the pipeline based on the acoustic principle, the position of the leakage point is usually calculated by calculating the time difference of the water leakage sound signal reaching the sensors at two ends of the pipeline. However, due to the existence of various interference factors such as turbulence of water flow in the pipeline, vibration of the pipeline, external environmental noise and the like, random errors often exist in the positioning result obtained by single correlation calculation, and reliability is difficult to guarantee. In view of high cost of pipeline excavation repair, in order to avoid misjudgment, repeated positioning analysis is generally adopted in actual engineering detection for continuously carrying out dozens of times or even hundreds of times on the same pipeline, so that a series of discrete suspected position data of the leakage point are obtained. How to extract the final trusted leak location from this batch of data becomes critical. For the processing of the discrete positioning data set, a simplified processing method based on statistical average and range cut-off is generally adopted by the conventional method. And simultaneously, in order to improve the physical rationality of the result, boundary processing is carried out on the data according to the preset pipeline physical length, and all the data points positioned outside the interval are directly removed. However, the above conventional methods have significant drawbacks. Firstly, because the acoustic signal is interfered by complex noise, the measured data set often presents a form of multimodal distribution (corresponding to a plurality of potential leakage sources or interference sources) or smooth distribution (without significant leakage signals), in such a case, the arithmetic average value is extremely easy to be interfered by an outlier, and the output position lacks statistical representativeness and has larger deviation from the actual leakage point position. Secondly, a 'one-cut' rejection strategy is adopted for the data outside the boundary, and the real leakage signal slightly exceeding the end point of the pipeline due to signal attenuation or calculation deviation can be misjudged as invalid data to be deleted, so that key information is lost. Thirdly, the method lacks the analysis capability of the internal space aggregation structure of the data, can not identify independent data clusters corresponding to different leakage sources, can not combine data subgroups which are scattered due to noise disturbance and belong to the same leakage point, and is difficult to cope with engineering scenes with a plurality of leakage points or complex data distribution. Therefore, the traditional method only depends on basic mathematical operation and physical boundary hard constraint, and an intelligent analysis mechanism capable of adapting to complex data distribution characteristics cannot be introduced, so that a technical bottleneck exists in a key link of improving positioning accuracy. Disclosure of Invention In order to solve the defects of the prior art, the invention provides a multi-stage clustering processing method and system for positioning leakage points, which can intelligently group discrete noise-containing suspected position data of the leakage points, accurately separate out different potential leakage points, calculate a stable and representative center positioning coordinate for each identified leakage point cluster, overcome the defect that the traditional arithmetic average value fails under the condition of multimodal or gentle distribution, automatically combine spatially adjacent sub-clusters possibly belonging to the same leakage point, optimize clustering results, avoid the same leakage point from being falsely segmented due to noise disturbance, output a final leakage point list in a designated detection area, intelligently reserve important suspected leakage points which are positioned outside the boundary of the area and fully exist, and effectively solve the problem that information outside the boundary is cut off. In one aspect, a multi-stage clustering method for locating leakage points is provided, including: Acquiring position data of a suspected leakage point of a pipeline section, and sequencing the position data according to the value to form an initial data sequence; Traversing each data point in the initial data sequence, and constructing a candidate subset ba