CN-121977174-A - Water supply network leakage monitoring method

CN121977174ACN 121977174 ACN121977174 ACN 121977174ACN-121977174-A

Abstract

The invention discloses a water supply network leakage monitoring method, which is characterized in that a multisource monitoring system of inlet flow and partition pressure is built based on water supply network metering partition, and a partition reference operation model is built in a stage of judging stable operation. The method comprises the steps of preprocessing real-time operation data, extracting multidimensional features such as inlet flow trend, pressure change, flow-pressure association deviation, night low-load stability and the like, and fusing the features to form a comprehensive deviation index. And after triggering early warning, outputting a suspected leakage area by combining a pipe network topological relation based on the variation amplitude difference and the variation sequence relation of pressure monitoring points in a partition, and providing an auxiliary decision for on-site refined leakage detection.

Inventors

ZHANG WEIPING
YANG CAIYUAN
CAO XIWEN
SUN ZHANXIANG

Assignees

西北工业大学

Dates

Publication Date: 20260505
Application Date: 20260327

Claims (10)

1. The water supply network leakage monitoring method is characterized by comprising the following steps of: S1, constructing metering subareas, namely dividing the water supply network into at least one independent metering subarea based on a topological structure, a valve boundary and a pipe diameter level of the water supply network, setting at least one inlet flow monitoring point in each metering subarea, and setting at least two pressure monitoring points in each subarea; s2, acquiring operation data, namely synchronously acquiring inlet flow data, partition pressure data and corresponding timestamp information of each metering partition according to a preset sampling period in the operation process of the water supply network; S3, data preprocessing, namely performing time alignment, deletion complement, abnormal rejection and smoothing processing on the acquired inlet flow data and pressure data to form continuous partition operation time sequence data; s4, constructing a reference operation model, namely building the reference operation model of a corresponding metering partition based on the partition operation time sequence data in a stable operation stage which is judged to be free of obvious leakage, wherein the reference operation model at least comprises a night minimum flow reference, a pressure fluctuation reference and a correlation reference between inlet flow and partition pressure; s5, extracting multi-dimensional leakage characteristics, namely analyzing real-time partition operation time sequence data by adopting a sliding time window, and extracting multi-dimensional leakage characteristics for representing leakage states, wherein the multi-dimensional leakage characteristics at least comprise inlet flow trend characteristics, pressure change characteristics, flow-pressure deviation characteristics and night low-load stability characteristics; And S6, judging leakage abnormality, namely comparing the multidimensional leakage characteristic with the reference operation model, calculating the comprehensive deviation index of the metering partition, judging the comprehensive deviation index based on the self-adaptive threshold and the persistence judging rule, and outputting a leakage early warning result of the corresponding metering partition when the preset condition is met.
2. The water supply network leakage monitoring method according to claim 1, wherein the dividing of the metering partition satisfies at least one of the following conditions: The metering partition boundary is formed by a controllable valve and has a definite water inlet and outlet boundary; the metering partition inlet flow monitoring point is arranged at the partition water inlet main pipe and is used for metering the total water supply flow entering the partition; The pressure monitoring points are arranged at least two types of positions in the end nodes of the subareas, the high points of the topography, the pipe diameter abrupt change positions or the user concentrated areas.
3. The water supply network leakage monitoring method according to claim 1, wherein the step S2 further comprises synchronously collecting pump station operation state data and/or valve opening state data, and introducing the operation state data as working condition characteristics into the step S4 and the step S6, so as to distinguish flow and pressure changes caused by water supply scheduling or working condition adjustment from abnormal changes caused by network leakage.
4. The water supply network leakage monitoring method according to claim 1, wherein the anomaly rejection in S3 includes: Setting physical rationality constraint and change rate constraint on inlet flow data and pressure data, and marking the sampling data as abnormal data when the sampling data exceeds the constraint conditions; And supplementing the data marked as abnormal or missing by adopting an adjacent time point interpolation, historical same-working-condition data interpolation or model prediction mode.
5. The water supply network leakage monitoring method according to claim 1, wherein the steady operation phase is determined by: In a preset time range, the night inlet flow is in a historical low value interval and has no continuous lifting trend; no persistent decay or abnormal fluctuation of the pressure data in the subarea occurs; And the comprehensive deviation index corresponding to the multidimensional leakage characteristic is continuously lower than a preset threshold value for a set time period, and the data segment meeting the conditions is selected as modeling data of a reference operation model.
6. The method for monitoring leakage of a water supply network according to claim 1, wherein the correlation reference between the inlet flow and the partition pressure establishes a mapping relationship between the partition inlet flow and the pressure of each pressure monitoring point through historical operation data, and predicts the corresponding pressure according to the current inlet flow or the back-thrust predicted flow according to the current pressure in a real-time operation process so as to form a reference for calculating the flow-pressure deviation characteristic.
7. The water supply network leakage monitoring method according to claim 1, wherein the inlet flow trend feature of the multi-dimensional leakage features is characterized by an inlet flow change trend in a sliding time window, the night low-load stability feature is characterized by a discrete degree change of inlet flow in a night period, and a stability level corresponding to the night minimum flow reference is compared and analyzed.
8. The method for monitoring leakage of a water supply network according to claim 1, wherein the adaptive threshold is dynamically adjusted according to a historical same type period, a seasonal characteristic, a weekday or non-weekday attribute, and a current working condition state, and the persistence criterion includes determining that the leakage abnormality exists when the integrated deviation index continuously exceeds the corresponding threshold for a preset duration or when the accumulated number of times exceeding the threshold within a preset time window reaches a preset number of times.
9. The method for monitoring leakage of a water supply network according to claim 1, wherein the leakage pre-warning result includes a leakage risk level, the leakage risk level is classified according to magnitude and duration of a comprehensive deviation index and a condition that at least two kinds of characteristics of multi-dimensional leakage characteristics are abnormal at the same time, and the risk level includes at least two or three stages of a prompt stage, a focus stage and an alarm stage.
10. The water supply network leakage monitoring method according to claim 1, further comprising a leakage positioning auxiliary step of: after triggering leakage early warning in a certain metering partition, determining a suspected leakage influence range based on pressure change amplitude differences and change sequence relations of a plurality of pressure monitoring points in the partition in the same time window, and outputting a corresponding suspected leakage pipe section or node set for guiding follow-up fine leakage detection or on-site investigation.

Description

Water supply network leakage monitoring method Technical Field The invention relates to the technical field of pipeline monitoring, in particular to a water supply network leakage monitoring method. Background The water supply network is taken as an important component of urban infrastructure and bears the task of stably conveying factory water of water plants to various users. With the expansion of urban scale and the growth of service life of pipe network, the factors such as material aging, interface loosening, corrosion perforation, third party construction disturbance and foundation settlement of the pipe network are easy to cause hidden leakage, seepage or burst rupture of the pipeline. The leakage not only causes direct water resource waste and the rise of water production and transmission and distribution energy consumption, but also can cause local water pressure fluctuation and water supply reliability reduction, even induces external pollutant invasion under the condition of negative pressure or sudden pressure change, and brings water quality safety risks. Therefore, how to find and locate leakage hidden trouble in time on the premise of not influencing water supply is one of the core problems of fine management and leakage control loss reduction of water supply enterprises. Existing leakage management means can be broadly divided into two categories, namely "offline investigation" and "online monitoring". The method of off-line investigation usually adopts methods such as leakage detection, correlation instrument, ground microphone, trace gas or sectional valve closing test, etc., can accurately confirm leakage points in a certain range, but often depends on manual inspection and on-site conditions, has high implementation cost and long period, is difficult to realize continuous coverage of the whole network, and is easy to cause long-term but undiscovered leakage for slow leakage under the conditions of leakage, micro leakage and low flow at night. The on-line monitoring method mainly relies on partition metering (DMA), inlet flow, pressure and partial key node monitoring data, judges leakage risk by establishing a reference and identifying deviation, has the advantages of wide coverage and strong real-time performance, and still has various pain points in practical engineering application. Firstly, the operation of the water supply system has obvious characteristics of time period and variable working conditions, and the water load difference between day and night, the difference between working days and non-working days, seasonal change, start-stop and variable frequency adjustment of a pump station, valve scheduling and hydraulic disturbance caused by temporary construction can all cause remarkable fluctuation of inlet flow and pressure. Part of the existing methods adopt a fixed threshold value or a single index (such as the minimum flow at night) for judgment, and are easy to realize, but false alarm easily occurs when the working condition changes or the water consumption behavior changes, and false alarm is possible to be missed when the change amplitude is small but the duration time is long for slow leakage, seepage and the like. Secondly, the quality of the pipe network monitoring data is difficult to stabilize for a long time, the on-site sensor has the problems of drift, packet loss, disconnection, inconsistent noise and time stamp, and the like, and if an effective abnormal rejection, missing completion and time alignment mechanism is lacked, model input distortion is easily caused, so that an early warning result is unreliable. And thirdly, part of the existing methods focus on judging whether leakage exists, but the quantitative grading and executable positioning assistance on the leakage risk degree are lacking, so that even if an operation and maintenance person receives an alarm, the operation and maintenance person can not determine a preferential investigation region, and can only carry out large-scale inspection depending on experience, thereby reducing the service value of the alarm. Finally, in the aspect of a data driving method, if a large number of historical leakage labeling samples or complex black box models are excessively relied on, the samples are difficult to quickly land under the actual data conditions of most water supply enterprises, and meanwhile, the trust and the adoption of alarms by an operation and maintenance department are influenced due to the lack of interpretability. Therefore, a method for monitoring leakage of a water supply network is urgently needed to solve the above problems. Disclosure of Invention The invention aims to solve the technical problems in the background art, and provides a water supply network leakage monitoring method, which comprises the following steps: S1, constructing metering subareas, namely dividing the water supply network into at least one independent metering subarea based on a topological structure, a valve boundar