A building drainage network anomaly detection system

Abstract

This invention relates to the field of fault detection technology, specifically to an anomaly detection system for building drainage pipe networks. The system includes a water pressure chain construction module, a sudden drop identification module, a flow resistance classification module, a structure mapping module, and an attribution verification module. In this invention, a continuously changing chain is constructed between water pressure monitoring points in the drainage pipe network. Spatial correlation analysis is achieved based on the changes in water pressure direction and amplitude between adjacent nodes. Anomalies can be quickly identified using chain-like water pressure gradient fractures as clues. In anomaly sudden drop identification, the temporal matching of upstream rebound and downstream delay is used as a verification condition, providing a dynamic response basis for anomaly judgment. Flow resistance type classification is formed through segmented matching of water pressure curve trends, realizing the hierarchical and pipe segment attribution of anomaly locations. Overall, a closed-loop detection mechanism is achieved from signal chain change identification to spatial path confirmation, enabling continuous, traceable, and improved positioning accuracy in drainage system anomaly diagnosis.

Description

Technical Field

[0001] This invention relates to the field of fault detection technology, and in particular to an anomaly detection system for building drainage pipe networks. Background Technology

[0002] The field of fault detection technology involves the identification, analysis, and localization of abnormal states that occur in various systems, equipment, or components during operation. Core aspects include monitoring operating parameters, data acquisition and comparative analysis, anomaly pattern identification, fault source tracing, and state assessment. This technology relies on the systematic processing of sensor-acquired data, signal analysis methods, and diagnostic logic, and is widely applied in various scenarios such as industrial equipment, transportation facilities, and built environments to achieve real-time monitoring of system operating status and timely prediction of potential faults. Among these, traditional building drainage network anomaly detection systems refer to devices and methods used to detect blockages, leaks, and abnormal water flow in building drainage systems during operation. The technical issue addressed is the identification of structural or operational anomalies in building drainage networks. Real-time monitoring of drainage status is achieved by deploying level sensors, flow sensors, and water pressure sensors to collect internal operating parameters of the drainage network. Anomalies in the drainage system are identified through threshold comparison and trend analysis. Traditional methods employ a small-scale approach based on sensor signal acquisition and simple threshold judgment for anomaly detection.

[0003] Existing methods for detecting anomalies in building drainage pipe networks rely on single-point sensors to collect signals and compare thresholds. In complex pipe network structures, there is a lack of temporal and spatial correlation between different monitoring points, resulting in anomaly judgments that only reflect local conditions. It is difficult to form continuous feature identification in the early stages of pipe blockage or leakage. Furthermore, fixed thresholds are easily affected by environmental fluctuations, leading to misjudgments or missed detections. When different pipe sections experience synchronous fluctuations, it is impossible to distinguish the flow direction relationship, resulting in ambiguous anomaly location, delayed fault diagnosis, extended troubleshooting cycles, increased maintenance costs, and increased overall operational risks of the pipe network. This makes it difficult to meet the needs of refined monitoring and real-time diagnosis. Summary of the Invention

[0004] The purpose of this invention is to overcome the shortcomings of existing technologies and propose an anomaly detection system for building drainage pipe networks.

[0005] To achieve the above objectives, the present invention adopts the following technical solution: a building drainage pipe network anomaly detection system includes:

[0006] The water pressure chain construction module obtains real-time water pressure sequence data of water pressure monitoring points deployed in the riser, horizontal branch pipe and inspection well section of the building drainage pipe network, calls the water pressure change direction between three adjacent monitoring points and forms a continuous chain relationship according to the spatial arrangement order of the pipe section, and generates a list of broken chain segment initialization identifiers.

[0007] The sudden drop identification module extracts the nodes marked as broken according to the broken chain segment initialization identifier list, obtains the time series water pressure fluctuation path with adjacent upstream and downstream pipe segments, and generates a spatial sudden drop verification mark set by judging whether the upstream of the sudden drop point has water pressure reverse rebound in the time window.

[0008] The flow resistance classification module extracts the continuous water pressure change trend curve based on the spatial sudden drop verification mark set, divides the time span and change slope of the rising and falling segments, and obtains the flow resistance structure type identification table.

[0009] The structure mapping module retrieves the node coordinate information in the building structure plan and elevation layout model according to the flow resistance structure type identification table, locates the corresponding building hierarchical structure relationship, analyzes the position and connection sequence of the corresponding pipe segment on the vertical path in the building model, and generates a list of abnormal node spatial ownership information.

[0010] As a further embodiment of the present invention, the initialization identifier list of the fractured chain segment includes the direction reversal position, the water pressure gradient fracture node, the water pressure offset synchronization point, and the water pressure chain segment number; the spatial sudden drop verification mark set includes the sudden drop point node position, upstream and downstream response characteristics, time sequence verification identifier, and chain segment corresponding index; the flow resistance structure type identifier table includes the transient release type number, the stagnant accumulation type number, the change slope value, and the time span label; and the abnormal node spatial attribution information list includes the floor location information, the main and branch pipe type, the node spatial coordinates, and the connection sequence path.

[0011] As a further aspect of the present invention, the hydraulic chain construction module includes:

[0012] The water pressure sequence extraction submodule acquires real-time water pressure sequence data of water pressure monitoring points deployed in the riser, horizontal branch pipe and inspection well section of the building drainage pipe network, and extracts the water pressure change direction between three adjacent monitoring points according to the spatial arrangement of the monitoring points to obtain a water pressure direction sequence chain.

[0013] The chain segment continuity judgment submodule calls the water pressure direction sequence chain group, takes three consecutive water pressure monitoring points as chain segment units, analyzes the consistency of water pressure change direction, and judges whether there is a reverse state in the direction inside the chain segment. If there is, the chain segment is marked as a reverse direction chain segment, and a list of reverse direction chain segments is obtained.

[0014] The gradient fracture identification submodule extracts the difference in water pressure change amplitude of nodes within the chain segment based on the direction reversal chain segment list, performs synchronization offset judgment based on the spatial synchronization between the direction reversal position and the water pressure amplitude abrupt change position, calculates the water pressure offset synchronization index value, and determines that the chain segment is in a water pressure gradient fracture state when the water pressure offset synchronization index value is greater than the set water pressure offset synchronization threshold, and obtains the fractured chain segment initialization identifier list.

[0015] As a further aspect of the present invention, the sudden drop recognition module includes:

[0016] The fracture node extraction submodule extracts the chain segment nodes marked as fractured according to the fracture chain segment initialization identifier list, calls the upstream and downstream adjacent pipe segment numbers for the chain segment nodes, obtains the time series water pressure fluctuation path of the corresponding node, analyzes the amplitude and direction change trend of continuous fluctuation in water pressure data, and obtains the fracture node index structure.

[0017] Based on the fracture node index structure, the water pressure response analysis submodule selects the upstream and downstream adjacent pipe sections corresponding to the fracture node, collects water pressure data sequences within a fixed time window before and after the sudden drop point, calls the pressure amplitude, rebound rate and time lag interval values ​​of the node within the time window, calculates the joint feature value of spatial response, identifies the water pressure response behavior of the node, and generates a water pressure linkage marker vector.

[0018] The spatial verification submodule filters out fracture nodes with upstream pressure rebound and downstream response lag based on the water pressure linkage marker vector, counts joint behavior segments of nodes that meet the conditions, identifies the fracture node and the response relationship with adjacent upstream and downstream, integrates the node sudden drop type, water pressure response direction and time lag state, and obtains a spatial sudden drop verification marker set.

[0019] As a further aspect of the present invention, the joint eigenvalue of the spatial response is expressed by the formula:

[0020] ;

[0021] in, Represents the joint eigenvalues ​​of the spatial response. Indicates the first Changes in upstream water pressure at each node Indicates the first Downstream water pressure fluctuation at each node Indicates the first The reverse increase in water pressure caused by the response delay of each node Indicates the first The water pressure fluctuation amplitude at each node sudden drop point Indicates the total number of nodes. This is the index number for the fracture node.

[0022] As a further aspect of the present invention, the flow resistance classification module includes:

[0023] The trend curve extraction submodule extracts the corresponding continuous water pressure time series data based on the spatial sudden drop verification mark set, identifies the water pressure breakpoint position in the time series and limits the data segments before and after the mark, extracts the continuous water pressure sequence at adjacent measuring point positions, and reconstructs the time series through time index to generate a water pressure change trend data sequence.

[0024] The path feature calculation submodule calls the water pressure change trend data sequence, retrieves continuous data segments before and after the point of sudden drop, delineates the time index intervals of the water pressure rise and fall segments, calculates the duration of the corresponding segments based on the index difference, calculates the difference based on the water pressure values ​​within the segments, and generates a set of water pressure path change feature parameters.

[0025] The type label classification submodule selects the duration of the rising segment and the slope of the falling segment as classification indicators based on the set of water pressure path change characteristic parameters. It sets the time span threshold and slope judgment benchmark value for the classification of sudden drop chain segment types, compares the relationship between the classification indicators and the corresponding judgment benchmark, determines the type of the sudden drop chain segment and assigns the corresponding type label, and generates a flow resistance structure type identification table.

[0026] As a further aspect of the present invention, the structure mapping module includes:

[0027] The node coordinate retrieval submodule matches and extracts the corresponding node coordinate number in the building structure plan and elevation layout model based on the sudden drop chain segment index recorded in the flow resistance structure type identification table, calls the three-dimensional coordinate data associated with the node number, integrates the spatial position and associated number of the component to which the node belongs, and generates an abnormal node coordinate information set.

[0028] The path structure positioning submodule calls the abnormal node coordinate information set, parses the floor index number of the node in the building structure model and the connection relationship between the nodes on the upper and lower floors, combines the node grouping information set in the building hierarchy structure, determines whether adjacent nodes are located in the same drainage zone or cross-floor passage, and identifies the connection order on the vertical path according to the sequence relationship of the upstream and downstream nodes, and obtains the vertical pipe segment connection sequence data.

[0029] The spatial attribution labeling submodule reads the pipe segment identifier and floor number corresponding to the node based on the vertical pipe segment connection sequence data, retrieves the drainage direction field recorded in the flow direction information, compares it with the upstream and downstream positions of the node, determines the floor position of the abnormal node in the drainage path, and matches it with the structural classification fields of the main pipe and branch pipe in the building drainage, marks the corresponding pipe segment type code, and generates a list of spatial attribution information for abnormal nodes.

[0030] As a further aspect of the present invention, the system also includes an attribution verification module:

[0031] The attribution verification module extracts the response time records of the node and the two upstream and downstream liquid level monitoring points based on the list of spatial attribution information of the abnormal node. It determines whether the liquid level change of the node conforms to the response order from upstream to downstream when the abnormal event occurs. If the downstream response is delayed, the attribution location is verified to be reasonable. If the response time is advanced, the adjacent gradient chain segment is re-extracted for reverse tracing and judgment, and the set of abnormal verification nodes of the pipeline network is output.

[0032] The pipeline anomaly verification node set includes nodes with successful attribution verification, nodes with failed attribution verification, nodes with reverse tracing results, and response order consistency identifiers.

[0033] The attribution verification module includes:

[0034] The response order judgment submodule calls the list of spatial affiliation information of the abnormal node, extracts the abnormal node and the associated upstream and downstream liquid level monitoring point numbers, retrieves the liquid level monitoring data records with the corresponding numbers, determines the start time of the liquid level change of the three monitoring points within the time period corresponding to the abnormal event, compares the response order of the upstream and downstream monitoring points through the timestamp field, filters the node group whose downstream response time is after the upstream, and generates the liquid level response order verification result.

[0035] The source tracing and update submodule identifies downstream node numbers whose response times are earlier than those of upstream monitoring points based on the response time distribution in the liquid level response sequence verification results. It retrieves the neighboring gradient chain segment information corresponding to the node, extracts the continuous connection nodes of the upstream pipe segment through reverse retrieval, determines whether there are liquid level fluctuation records in the real-time upstream nodes that match the time period of the abnormal event, filters the node numbers that meet the conditions, and generates a set of pipeline network abnormal verification nodes.

[0036] Compared with the prior art, the advantages and positive effects of the present invention are as follows:

[0037] In this invention, a continuous change chain is constructed between water pressure monitoring points in the drainage pipe network. Spatial correlation analysis is achieved based on the changes in water pressure direction and amplitude of adjacent nodes. Abnormal areas can be quickly identified by using chain-like water pressure gradient fracture as a clue. In the identification of abnormal drop, the timing matching of upstream rebound and downstream delay is used as a verification condition, so that the abnormal judgment has a dynamic response basis. By segmenting and matching the trend of water pressure curves, the flow resistance type is classified, so that abnormal features such as blockage or stagnation have physical attribute orientation. In the structural mapping and attribution verification stage, the detection results are combined with the building space structure to realize the hierarchical and pipe segment attribution of abnormal location. The whole realizes a closed-loop detection mechanism from signal chain change identification to spatial path confirmation, so that the diagnosis of drainage system anomalies has continuity, traceability and improved positioning accuracy. Attached Figure Description

[0038] Figure 1 This is a system flowchart of the present invention;

[0039] Figure 2 This is a flowchart of the hydraulic chain construction module in this invention;

[0040] Figure 3 This is a flowchart of the sudden drop recognition module in this invention;

[0041] Figure 4 This is a flowchart of the flow resistance classification module in this invention;

[0042] Figure 5 This is a flowchart of the structure mapping module in this invention;

[0043] Figure 6 This is a flowchart of the attribution verification module in this invention. Detailed Implementation

[0044] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0045] In the description of this invention, it should be understood that the terms "length," "width," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," and "outer," etc., indicating orientation or positional relationships, are based on the orientation or positional relationships shown in the accompanying drawings and are only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the invention. Furthermore, in the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0046] Please see Figure 1 A building drainage network anomaly detection system includes:

[0047] The water pressure chain construction module acquires real-time water pressure sequence data from water pressure monitoring points deployed in the risers, horizontal branches, and inspection well sections of the building drainage network. It calls the water pressure change direction between three adjacent monitoring points within the same time window and forms a continuous chain relationship according to the spatial arrangement of the pipe sections. It selects three adjacent nodes as chain segment units, judges the chain continuity based on the consistency of the water pressure change direction, and marks the chain segments where the change reverses. It searches whether the position where the direction reverses within the same chain segment is synchronously offset with the water pressure change amplitude. If it is judged to be synchronously offset, the chain segment is identified as having a water pressure gradient breakage state, and a broken chain segment initialization identifier list is generated.

[0048] The sudden drop identification module initializes the list of broken chain segments, extracts the chain segment nodes marked as broken, obtains the time series water pressure fluctuation path with adjacent upstream and downstream pipe segments, extracts the water pressure response changes of adjacent nodes according to the time window before and after the sudden drop point, and generates a spatial sudden drop verification mark set by judging whether there is water pressure rebound upstream of the sudden drop point in the time window and judging whether there is response delay behavior of downstream nodes.

[0049] The flow resistance classification module extracts the continuous water pressure change trend curve based on the spatial sudden drop verification mark set, obtains the complete water pressure change path before and after the anomaly, divides the time span and change slope of the rising and falling segments, combines and matches the duration of the rising segment with the slope of the falling segment, performs the label classification operation, and classifies the sudden drop chain segments into transient release type or stagnant accumulation type respectively, and obtains the flow resistance structure type identification table.

[0050] The structural mapping module retrieves the node coordinate information in the building structure plan and elevation layout model based on the flow resistance structure type identification table, locates the building level structure relationship, analyzes the position and connection sequence of the corresponding pipe segment on the vertical path in the building model, and marks the abnormal position in the floor where the drainage is located and the type of the main branch pipe in combination with the flow direction information, and generates a list of abnormal node spatial ownership information.

[0051] The attribution verification module extracts the response time records of the node and the two upstream and downstream liquid level monitoring points based on the list of spatial attribution information of abnormal nodes. It determines whether the liquid level change of the node conforms to the response order from upstream to downstream when the abnormal event occurs. If the downstream response is delayed, the attribution location is verified to be reasonable. If the response time is advanced, the adjacent gradient chain segment is re-extracted for reverse tracing and judgment, and the set of abnormal verification nodes of the pipeline network is output.

[0052] The list of initialization identifiers for broken chain segments includes the direction reversal position, water pressure gradient break node, water pressure offset synchronization point, and water pressure chain segment number. The spatial drop verification mark set includes the drop point node location, upstream and downstream response characteristics, time-series verification identifier, and chain segment corresponding index. The flow resistance structure type identifier table includes the transient release type number, stagnant accumulation type number, change slope value, and time span label. The list of spatial attribution information for abnormal nodes includes floor location information, main and branch pipe types, node spatial coordinates, and connection sequence path. The set of abnormal verification nodes for the pipeline network includes nodes with successful attribution verification, nodes with failed attribution verification, nodes with reverse tracing results, and response sequence consistency identifier.

[0053] Please see Figure 2 The hydraulic chain building module includes:

[0054] The water pressure sequence extraction submodule acquires real-time water pressure sequence data of water pressure monitoring points deployed in the riser, horizontal branch pipe and inspection well section of the building drainage pipe network, and extracts the water pressure change direction between three adjacent monitoring points according to the spatial arrangement of the monitoring points to obtain a water pressure direction sequence chain.

[0055] Based on the spatial order of the monitoring points, the deployed water pressure monitoring points are numbered, and the numbering order is set according to the deployment order. For example, if there are 6 monitoring points in a building, they are numbered from P1 to P6. On this basis, the direction of water pressure change between three adjacent monitoring points (set as P1, P2, P3) is extracted. The water pressure change value ΔP1 between P1 and P2 per unit time and the water pressure change value ΔP2 between P2 and P3 are calculated. By comparing the positive and negative relationship between ΔP1 and ΔP2, if ΔP1>0 and ΔP2>0, the direction of water pressure change is recorded as a positive chain segment. If ΔP1>0 and ΔP2<0, the direction is determined to be reversed and recorded as a negative chain segment. The water pressure direction results of three adjacent points are recorded in sequence to obtain the water pressure direction sequence chain group.

[0056] The chain segment continuity judgment submodule calls the water pressure direction sequence chain group, takes three consecutive water pressure monitoring points as chain segment units, analyzes the consistency of the water pressure change direction, and judges whether there is a reverse state in the direction inside the chain segment. If there is, the chain segment is marked as a reverse direction chain segment, and a list of reverse direction chain segments is obtained.

[0057] In each chain segment consisting of three consecutive points, the consistency of the direction of water pressure change within each segment is determined sequentially. The judgment criterion is whether the angle between any two adjacent direction vectors within the chain segment is 180°. If they are the same, it means that the direction is continuous and consistent, and the chain segment is recorded as a positive chain segment. If they are different, it is a reverse turning point, and the index of the turning chain segment needs to be recorded. The direction sequence chain group is set as [1, 1, -1, -1, 1], where 1 represents positive and -1 represents negative. Then, the 2nd to 3rd point is the turning point, and the corresponding chain segment index 2 is recorded as the turning point index, thus obtaining the list of reverse direction chain segments.

[0058] The gradient fracture identification submodule extracts the difference in water pressure change amplitude between nodes within a segment based on the list of direction-reversed segments. It then determines the synchronization offset based on the spatial synchronicity between the direction-reversal location and the location of the sudden change in water pressure amplitude, using the following formula:

[0059] ;

[0060] Calculate the water pressure offset synchronization index value. When the water pressure offset synchronization index value is greater than the set water pressure offset synchronization threshold, determine that the chain segment is in a water pressure gradient breakage state and obtain the list of broken chain segment initialization identifiers.

[0061] in, For the first The water pressure offset synchronization index value of the chain segment Represents a chain segment Inner The water pressure difference between the node and the previous node, Indicates the first segment within the chain. The spatial distance between a node and its predecessor. This represents the peak value of the water pressure difference at the monitoring nodes within the same time window. This represents the normalized sum of squares index of the distance variation within a chain segment. This represents the number of chain segments;

[0062] Formula calculation logic: By accumulating the product of water pressure change values ​​between adjacent nodes within a chain segment and the spatial distance between nodes, the consistency trend of water pressure fluctuation in spatial distribution is reflected. Absolute value processing can avoid interference caused by directional changes. A dimensionless index value is constructed by sum-normalization. The denominator is set as the sum of the peak water pressure and the number of nodes within the time window. Water pressure fluctuation limit and chain segment complexity are jointly standardized to ensure comparability under different chain segments and different time scales. If the water pressure change within the chain segment is drastic and the node spacing is concentrated in a local range, the water pressure offset synchronization index value increases. Conversely, if the water pressure change amplitude is low or the directional change is frequent, causing the product results to cancel each other out, the water pressure offset synchronization index value tends to decrease.

[0063] The water pressure offset synchronization index is used to measure whether the water pressure changes and spatial positions of each monitoring node in the chain segment are synchronized. The index comprehensively considers the water pressure change amplitude, node spacing and extreme value normalization within the monitoring period to form a standardized result. The larger the value, the more concentrated the water pressure changes and the obvious offset trend. The smaller the value, the more dispersed the fluctuations or the inconsistent change trends.

[0064] Meaning of parameters and calculation process:

[0065] : chain segment The water pressure difference between the j-th node and the previous node, in kPa;

[0066] : chain segment The horizontal projected spatial distance between the j-th node and the previous node, in meters;

[0067] : chain segment The number of node pairs within;

[0068] The maximum value of the water pressure difference among the monitoring points within the current analysis time window T;

[0069] : No. The number of monitoring nodes contained in the chain segment, rounded to a positive integer. ≥3;

[0070] Practical example demonstration:

[0071] Let a certain chain segment be numbered as =2, including 4 monitoring points (P1, P2, P3, P4), that is =4, n=3, time window T is 60 seconds, water pressure difference between nodes in the chain segment With distance As shown in the table below:

[0072] Table 1: Calculation Data Table of Water Pressure Deviation Synchronization Index

[0073] ;

[0074] Let the maximum water pressure difference among all monitoring points in the network be within the current timeframe of T=60 seconds. =6.4kPa, substituting into the formula, the calculation is as follows:

[0075] ;

[0076] The results show that the water pressure offset synchronization index value is 1.035. Using a set threshold of 2.5 as the standard, G_2 = 1.035 < 2.5. Therefore, the segment does not meet the abrupt change criterion and belongs to a segment with a stable water pressure change trend, and thus does not enter the crack candidate list. If the value is ≥2.5, the chain segment is determined to be a high-offset chain segment.

[0077] Threshold 2.5 setting instructions and experimental basis:

[0078] An experiment was conducted based on 50 sets of monitoring chain segment data. 22 sets of cracked segments exhibited sudden water pressure shifts, and 28 sets of normal segments were recorded. Statistical analysis revealed that:

[0079] exist Of the segments with a length of ≥2.5, 20 groups were known crack segments, accounting for more than 90% of the crack groups;

[0080] exist Of the chain segments with a length <2.5, only two groups were cracked segments, and the rest were normal fluctuations.

[0081] Based on this, a recognition threshold of 2.5 is set, which can achieve a recognition strategy with a crack segment recognition rate of >90%, while also taking into account the control of false judgments;

[0082] The advantage of this formula lies in constructing a spatial-hydraulic synergy index by multiplying the water pressure difference by the physical distance, using absolute value processing to enhance robustness against directional reversal, and introducing the chain segment length. The correction improves the applicability under different point densities, and the overall construction of a unified and normalized criterion enables the extraction of crack identification features across space and time domains.

[0083] Please see Figure 3 The sudden drop detection module includes:

[0084] The fracture node extraction submodule extracts the chain segment nodes marked as fractured based on the fractured chain segment initialization identifier list, calls the upstream and downstream adjacent pipe segment numbers for the chain segment nodes, obtains the time series water pressure fluctuation path of the corresponding nodes, analyzes the amplitude and direction change trend of continuous fluctuations in the water pressure data, and obtains the fracture node index structure.

[0085] Nodes within a chain segment identified as potentially broken are numbered. Each chain segment consists of consecutively numbered monitoring nodes. The process iterates through each node in each chain segment, calling the numbers of adjacent upstream and downstream nodes, and recording them as follows: and After obtaining the node number, based on the corresponding water pressure fluctuation curve in the time series, water pressure data within a continuous time window T centered on the node is extracted. After aligning the data, the relative amplitude difference of water pressure fluctuations between nodes is compared. ,calculate The rate of change, i.e. And combined with the directional change sequence between nodes, such as the directional reversal chain segment table obtained in the previous section, the directional change interval of the node in the time series is marked, and the fluctuation amplitude of the node is analyzed to see if it is in an upward, downward or periodic oscillation state. By extracting the fluctuation frequency, it is compared with the frequency of adjacent nodes. If the node frequency differs from the upstream and downstream nodes, and If the amplitude exceeds the preset reference value of 2.0 kPa, the recorded node is the fracture feature index node, and the fracture node index structure is obtained.

[0086] The water pressure response analysis submodule, based on the fracture node index structure, selects the upstream and downstream adjacent pipe sections corresponding to the fracture node. It collects water pressure data sequences within a fixed time window before and after the sudden drop point, and retrieves the pressure amplitude, rebound rate, and time lag interval values ​​of the nodes within the time window, using the following formula:

[0087] ;

[0088] Calculate the joint eigenvalues ​​of the spatial response, identify the water pressure response behavior of the nodes, and generate a water pressure linkage label vector;

[0089] in, Represents the joint eigenvalues ​​of the spatial response. Indicates the first Changes in upstream water pressure at each node Indicates the first Downstream water pressure fluctuation at each node Indicates the first The reverse increase in water pressure caused by the response delay of each node Indicates the first The water pressure fluctuation amplitude at each node sudden drop point Indicates the total number of nodes. Assign index numbers to the fracture nodes;

[0090] Formula calculation logic: Taking the fracture node as the center, extract the characteristics of water pressure changes upstream and downstream, and use three parameters... , , These represent the upstream fluctuation amplitude, response delay, and downstream fluctuation amplitude, respectively. First, let's consider... and The purpose of summing and taking the arithmetic square root is to highlight the coupling characteristics between downstream fluctuations and response lags, and then multiplying the result by... The reciprocal of the result is used to complete the normalization adjustment, and... Sum, to obtain the first The response values ​​of each node are averaged over the results at the fault nodes to obtain the joint spatial response characteristic value RS, which reflects the degree of upstream and downstream hydraulic coupling of the entire set of fault points. The calculation structure of this formula is as follows: first, the composite fluctuation effect of the individual response of each node is processed, and then the whole is normalized and averaged to ensure... The values ​​are comparable under different node numbers, fluctuation amplitudes, and time delay characteristics, constituting a stable criterion for spatial hydraulic response characteristics;

[0091] The spatial response joint eigenvalue is used to measure the comprehensive response characteristics of the fault node and the changes in water pressure upstream and downstream. It integrates upstream water pressure fluctuations, response time delays and downstream disturbances, and unifies the dimensions through normalization. The larger the value, the more sensitive the fault node is to hydraulic disturbances and the stronger the linkage fluctuation characteristics.

[0092] Meaning of parameters and calculation process:

[0093] : No. The magnitude of water pressure change upstream of each node, in kPa;

[0094] : No. The magnitude of the reverse increase in water pressure caused by the response delay of each node, in kPa;

[0095] : No. Downstream fluctuation at each node, in kPa;

[0096] : No. Normalized baseline amplitude of water pressure fluctuation at each node, in kPa;

[0097] Total number of fracture nodes;

[0098] The following are sample data for three selected nodes:

[0099] Table 2: Calculation Parameters for Space Water Pressure Response

[0100] ;

[0101] Substituting into the formula, the calculations are as follows:

[0102] Node 1:

[0103] ;

[0104] Node 2:

[0105] ;

[0106] Node 3:

[0107] ;

[0108] Substitute into the formula to calculate:

[0109] ;

[0110] The results show that the joint eigenvalue of the spatial response is 4.9755, and the baseline value is set at 4.5. Based on the comparative test of 50 sets of fracture and non-fracture data, the following was found:

[0111] when When the value is ≥4.5, 95% of the corresponding nodes are true fracture responses;

[0112] when When the value is less than 4.5, most of the fluctuations are normal hydraulic fluctuations.

[0113] Therefore, 4.5 is used as the standard value for judging fracture response. Nodes with a value greater than 4.5 are included in the fracture hydraulic pressure response node set as spatial hydraulic linkage marker areas.

[0114] The spatial verification submodule filters out fracture nodes with upstream pressure rebound and downstream response lag based on the water pressure linkage marker vector, counts the joint behavior segments of nodes that meet the conditions, identifies the response relationship between fracture nodes and adjacent upstream and downstream nodes, integrates the node sudden drop type, water pressure response direction and time lag state, and obtains the spatial sudden drop verification marker set.

[0115] The set of identified fracture response nodes is traversed, and a chain-like backtracking is performed on the spatial upstream and downstream positions of the nodes. Forward and reverse verification chains are established separately. Water pressure sequences within two time windows before and after each node are extracted. Analysis is conducted to determine whether upstream nodes exhibit abrupt reversal behavior and whether downstream nodes exhibit water pressure attenuation behavior. The distribution of the rate of change is determined according to the following rules: if upstream Continuous increase and downstream If the value continuously decreases, the node is determined to be a node with instability caused by fracture. This is further analyzed by considering the difference between the node's fluctuation period Ti and the periods T{i±1} of neighboring nodes. If the time is greater than 2 seconds, it is considered that the node response characteristics are significantly related to the fracture process, and nodes that meet the above joint rules are included in the spatial drop verification label set.

[0116] Please see Figure 4 The flow resistance classification module includes:

[0117] The trend curve extraction submodule extracts the corresponding continuous water pressure time series data based on the spatial drop verification mark set, identifies the water pressure breakpoint location in the time series and limits the data segments before and after the mark, extracts the continuous water pressure sequence at adjacent measuring point locations, and reconstructs the time series through time index to generate a water pressure change trend data sequence.

[0118] Based on the marked geographical coordinates and corresponding timestamps, continuous water pressure time series data for the corresponding measuring point are retrieved from the water pressure monitoring database. The measuring point is designated S101, located at 105.21°E, 29.34°N, with a corresponding sudden drop marker time of 14:20 on October 10, 2025. The system automatically retrieves water pressure data from 14:00 to 14:20 and forward data from 14:20 to 14:40, constructing a complete trend series. Analysis of the extracted time series identifies the water pressure breakpoint, i.e., the point of sudden drop. This can be determined by comparing the water pressure difference between adjacent times and combining it with the set time. The sudden drop location is identified by span filtering. The data before and after the sudden drop point is divided into two segments, recorded as the pre-drop segment and the post-drop segment, respectively. At the same time, data from adjacent measuring points, such as S100 and S102, are retrieved for the same time period. The measuring points must meet certain spatial distance restrictions, such as a distance of no more than 50 meters. After extracting the water pressure sequence of the measuring points, the time index is unified according to the timestamp of the sudden drop point to ensure that the time series of the measuring points have the same time nodes. The water pressure data of the measuring points is re-aligned in minutes. Missing data can be filled using linear interpolation. The time series reconstruction of water pressure trend data across measuring points is completed, generating a water pressure change trend data sequence.

[0119] The path feature calculation submodule calls the water pressure change trend data sequence, retrieves the continuous data segments before and after the point of sudden drop, defines the time index intervals of the water pressure rise and fall segments, calculates the duration of the corresponding segments based on the index difference, calculates the difference based on the water pressure values ​​within the segments, and generates a set of water pressure path change feature parameters.

[0120] For each sudden drop event, the system automatically retrieves continuous water pressure data segments before and after the drop and divides them into two segments: a pressure drop segment and a pressure rise segment. The time intervals for each segment are defined based on the time of the drop. For example, for a drop at 14:20, the drop segment is defined as 14:00 to 14:20, and the rise segment as 14:20 to 14:40. The system records the start and end indices of these two time segments and calculates the duration of each segment (20 minutes for both). Simultaneously, the system reads the water pressure value at each time point and calculates the pressure change within each segment. For instance, in the drop segment, the initial water pressure drops from 2.1 MPa to a minimum of 1.4 MPa, and in the rise segment, it rises from 1.4 MPa to 2.0 MPa. The system records the pressure change values ​​and durations for these two segments and calculates the corresponding trend characteristics, generating a set of water pressure path change characteristic parameters.

[0121] The type label classification submodule selects the duration of the rising segment and the slope of the falling segment as classification indicators based on the water pressure path change characteristic parameter set. It sets the time span threshold and slope judgment benchmark value for the sudden drop chain segment type classification, compares the relationship between the classification indicators and the corresponding judgment benchmark, determines the type of the sudden drop chain segment and assigns the corresponding type label, and generates a flow resistance structure type identification table.

[0122] The system performs judgment and classification operations based on a pre-set time span threshold and a change slope benchmark. The set threshold for the rising segment time is 15 minutes. If the rising segment duration in an event is 20 minutes, it is identified as a "long-term recovery" event. The system judges the change slope of the rising segment, with a slope benchmark set at 0.025 MPa / min. If the rising slope of the event is 0.03 MPa / min, which is significantly higher than the judgment benchmark, it is classified into the "fast recovery" category. Based on the combination of these two judgment indicators, the system automatically matches the corresponding type label in the database. The system sets the category of sudden drop chain segments with a duration of more than 15 minutes and a rising slope greater than 0.025 as "medium-term fast recovery type". The classification criteria can be derived from industry experience or actual field observations and can be dynamically adjusted based on subsequent data feedback. The system assigns a unique type label to each sudden drop chain segment and generates a flow resistance structure type identification table.

[0123] Please see Figure 5 The structure mapping module includes:

[0124] The node coordinate retrieval submodule matches and extracts the corresponding node coordinate numbers in the building structure plan and elevation layout model based on the sudden drop chain segment index recorded in the flow resistance structure type identification table, calls the three-dimensional coordinate data associated with the node number, integrates the spatial position and associated number of the component to which the node belongs, and generates an abnormal node coordinate information set.

[0125] For each drop chain segment number, a coordinate matching operation is performed in the building structure plan and elevation layout model. This operation locates the component area in the model that includes the event corresponding to the drop chain segment using the event number. Then, it retrieves the node number information related to the area from the model structure database. If a certain event number corresponds to the pipe segment area numbered F3-17, the system searches for the included node numbers, such as N031 and N032, in the elevation model based on this number. The system then calls the 3D coordinate data corresponding to this set of node numbers and reads it from the 3D building model. The X, Y, and Z coordinate values ​​defined in the system, such as (15.2, 28.3, 3.6) for N031 and (15.2, 28.3, 6.6) for N032, indicate that the start and end points of the pipe segment are located on the same plane coordinates but the height difference of the Z axis is 3 meters. The system integrates the component number, component type, and relative position of the node in the component to which the node belongs, and sets the marker as the start and end node of the riser segment, the transition node of the horizontal branch pipe, or the valve connection node, etc. At the same time, it establishes a correspondence table between the node number and the component number and generates a set of abnormal node coordinate information.

[0126] The path structure positioning submodule calls the abnormal node coordinate information set, parses the floor index number of the node in the building structure model and the connectivity relationship between the nodes on the upper and lower floors, combines the node grouping information set in the building hierarchy structure, determines whether adjacent nodes are located in the same drainage zone or cross-floor passage, and identifies the connection order on the vertical path according to the sequence relationship of the upstream and downstream nodes, and obtains the vertical pipe segment connection sequence data.

[0127] The system parses the floor index number of each node from the building structure model. The Z-coordinate of node N031 is set to 3.6 meters. If each floor height is 3 meters, then the corresponding floor is floor 2. The system generates a floor index list for abnormal nodes in this way, searches the node connectivity information table for connections with nodes on the upper and lower floors. For example, if nodes N031 and N032 have a riser connection, the system records it as a vertical connection. It then checks the pre-set node grouping information in the building model to determine if N031 and N032 belong to the same drainage zone. If both are in the north drainage group, or if they cross a floor-to-floor riser, the system establishes an upstream and downstream connection sequence between each group of nodes, constructing a connection sequence record for the vertical drainage path, marked as "2F-N031→3F-N032→4F-N033". Simultaneously, it records the floor number, drainage zone number, and connection direction of each node in the connection sequence. If there is a cross-floor connection in the connection path but they belong to different zones, it is marked as "cross-floor connection in different zones". The system then obtains the vertical pipe segment connection sequence data.

[0128] The spatial attribution labeling submodule reads the pipe segment identifier and floor number corresponding to the node based on the vertical pipe segment connection sequence data, retrieves the drainage direction field recorded in the flow direction information, compares it with the upstream and downstream positions of the node, determines the floor position of the abnormal node in the drainage path, and matches it with the structural classification fields of the main pipe and branch pipe in the building drainage, marks the corresponding pipe segment type code, and generates a list of spatial attribution information for abnormal nodes.

[0129] The system reads the associated pipe segment identification information based on the node number in the connection sequence and extracts the floor number of the node from the floor structure database. For example, the pipe segment number corresponding to N032 is P56 and the floor number is 3. The system then retrieves the flow direction field information from the hydraulic model database to determine whether the water flow direction at the node is downward. If node N032 is the upstream end and the downstream node is N033, and the flow direction is from top to bottom, then the node is confirmed to be in the intermediate transmission section of the drainage system. If it is the terminal node and there is no downstream connection, then it is determined to be the terminal drainage point. The system reads the component structure classification field to distinguish between main pipe or branch pipe components. If the pipe segment where the node is located is a longitudinal riser connecting multiple branches, the system marks the pipe segment type as "main riser". If the component where the node is located is connected to only one horizontal drainage branch pipe, the system marks it as "branch pipe". The system integrates the five pieces of information, namely node number, pipe segment identification, floor number, flow direction status, and pipe segment type code, to form a list of abnormal node spatial ownership information.

[0130] Please see Figure 6 The attribution verification module includes:

[0131] The response order judgment submodule calls the list of abnormal node spatial ownership information, extracts the abnormal node and the associated upstream and downstream liquid level monitoring point numbers, retrieves the liquid level monitoring data records with the corresponding numbers, determines the start time of the liquid level change of the three monitoring points within the time period corresponding to the abnormal event, compares the response order of the upstream and downstream monitoring points through the timestamp field, filters the node group whose downstream response time is after the upstream, and generates the liquid level response order verification result.

[0132] The system retrieves the upstream and downstream liquid level monitoring point numbers corresponding to the abnormal node. These numbers are typically mapped to sensor numbers in building hydraulic monitoring equipment. For example, abnormal node N045 corresponds to upstream monitoring point L002, downstream monitoring point L003, and its own liquid level monitoring point L001. The system reads the corresponding liquid level monitoring data records based on the numbers, filters out the time period of the abnormal event (set as 14:15 to 14:45 on October 10, 2025), and extracts the liquid level record sequences from the three points within this time range. It performs initial change analysis on each sequence, determining the initial response time of the liquid level change by detecting the timestamp field of the liquid level abrupt change point, and marking it as T1 (upstream), T2 (abnormal), and T3 (downstream). The system compares the order of T1 and T3; if T3 > T1, the downstream response is determined to be after the upstream response, and the node is included in the acceptable response sequence. Conversely, if T3 < T1, the combination is excluded. The system then records the abnormal node groups and corresponding liquid level response times that conform to the upstream and downstream response logic, obtaining the liquid level response sequence verification results.

[0133] The source tracing and determination update submodule identifies downstream node numbers whose response times are earlier than those of upstream monitoring points based on the response time distribution in the liquid level response sequence verification results. It retrieves the neighboring gradient chain segment information corresponding to the node, extracts the continuous connection nodes of the upstream pipe segment through reverse retrieval, determines whether there are liquid level fluctuation records in the real-time upstream nodes that match the time period of the abnormal event, filters the node numbers that meet the conditions, and generates a set of pipeline network abnormal verification nodes.

[0134] The system analyzes the response time comparison between downstream and upstream nodes in each record. For downstream monitoring point numbers whose response time is earlier than that of upstream nodes, the system records them as suspicious node number groups and uses them as tracing targets. The system retrieves the information of the neighboring gradient chain segments corresponding to the node. The information includes the topological relationship between the node and multiple surrounding connected pipe segments. By retrieving the pipe segment connection table in reverse order, the system traces the continuous chain of connected nodes upwards and extracts the associated nodes in the upstream components in sequence. The system sets the tracing upwards from N053 to N052, N051, N050, etc., to form a complete reverse path. For each node on the path, the system calls the liquid level monitoring record to determine whether there is a similar or continuous liquid level fluctuation pattern during the abnormal event time period. If there is a sudden rise or fall response that matches the current event, the recorded node is a potential abnormal trigger source. The system continues to determine whether there are multiple nodes that meet this condition. If so, all of them are included in the verification set to generate a set of pipe network abnormal verification nodes.

[0135] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention in any other way. Any person skilled in the art may make changes or modifications to the above-disclosed technical content to create equivalent embodiments that can be applied to other fields. However, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the protection scope of the present invention.

Claims

1. A building drainage pipe network anomaly detection system, characterized in that, The system includes: The water pressure chain construction module obtains real-time water pressure sequence data of water pressure monitoring points deployed in the riser, horizontal branch pipe and inspection well section of the building drainage pipe network, calls the water pressure change direction between three adjacent monitoring points and forms a continuous chain relationship according to the spatial arrangement order of the pipe section, and generates a list of broken chain segment initialization identifiers. The hydraulic chain construction module includes: The water pressure sequence extraction submodule acquires real-time water pressure sequence data of water pressure monitoring points deployed in the riser, horizontal branch pipe and inspection well section of the building drainage pipe network, and extracts the water pressure change direction between three adjacent monitoring points according to the spatial arrangement of the monitoring points to obtain a water pressure direction sequence chain. The chain segment continuity judgment submodule calls the water pressure direction sequence chain group, takes three consecutive water pressure monitoring points as chain segment units, analyzes the consistency of water pressure change direction, and judges whether there is a reverse state in the direction inside the chain segment. If there is, the chain segment is marked as a reverse direction chain segment, and a list of reverse direction chain segments is obtained. The gradient fracture identification submodule extracts the difference in water pressure change amplitude of nodes within the chain segment based on the direction reversal chain segment list, performs synchronization offset judgment based on the spatial synchronization between the direction reversal position and the water pressure amplitude change position, calculates the water pressure offset synchronization index value, and determines that the chain segment is in a water pressure gradient fracture state when the water pressure offset synchronization index value is greater than the set water pressure offset synchronization threshold, and obtains the fractured chain segment initialization identifier list. The sudden drop identification module extracts the nodes marked as broken according to the broken chain segment initialization identifier list, obtains the time series water pressure fluctuation path with adjacent upstream and downstream pipe segments, and generates a spatial sudden drop verification mark set by judging whether the upstream of the sudden drop point has water pressure reverse rebound in the time window. The flow resistance classification module extracts the continuous water pressure change trend curve based on the spatial sudden drop verification mark set, divides the time span and change slope of the rising and falling segments, and obtains the flow resistance structure type identification table. The structure mapping module retrieves the node coordinate information in the building structure plan and elevation layout model according to the flow resistance structure type identification table, locates the corresponding building hierarchical structure relationship, analyzes the position and connection sequence of the corresponding pipe segment on the vertical path in the building model, and generates a list of abnormal node spatial ownership information.

2. The building drainage pipe network anomaly detection system according to claim 1, characterized in that, The list of initial identifiers for broken chain segments includes the direction reversal position, water pressure gradient break node, water pressure offset synchronization point, and water pressure chain segment number. The spatial drop verification mark set includes the drop point node location, upstream and downstream response characteristics, time-series verification identifier, and chain segment corresponding index. The flow resistance structure type identifier table includes the transient release type number, stagnant accumulation type number, change slope value, and time span label. The list of spatial attribution information for abnormal nodes includes floor location information, main and branch pipe type, node spatial coordinates, and connection sequence path.

3. The building drainage pipe network anomaly detection system according to claim 1, characterized in that, The sudden drop detection module includes: The fracture node extraction submodule extracts the chain segment nodes marked as fractured according to the fracture chain segment initialization identifier list, calls the upstream and downstream adjacent pipe segment numbers for the chain segment nodes, obtains the time series water pressure fluctuation path of the corresponding node, analyzes the amplitude and direction change trend of continuous fluctuation in water pressure data, and obtains the fracture node index structure. Based on the fracture node index structure, the water pressure response analysis submodule selects the upstream and downstream adjacent pipe sections corresponding to the fracture node, collects water pressure data sequences within a fixed time window before and after the sudden drop point, calls the pressure amplitude, rebound rate and time lag interval values ​​of the node within the time window, calculates the joint feature value of spatial response, identifies the water pressure response behavior of the node, and generates a water pressure linkage marker vector. The spatial verification submodule filters out fracture nodes with upstream pressure rebound and downstream response lag based on the water pressure linkage marker vector, counts joint behavior segments of nodes that meet the conditions, identifies the fracture node and the response relationship with adjacent upstream and downstream, integrates the node sudden drop type, water pressure response direction and time lag state, and obtains a spatial sudden drop verification marker set.

4. The building drainage pipe network anomaly detection system according to claim 3, characterized in that, The joint eigenvalues ​​of the spatial response are expressed by the formula: ； in, Represents the joint eigenvalues ​​of the spatial response. Indicates the first Changes in upstream water pressure at each node Indicates the first Downstream water pressure fluctuation at each node Indicates the first The reverse increase in water pressure caused by the response delay of each node Indicates the first The amplitude of water pressure fluctuation at each node sudden drop point Indicates the total number of nodes. This is the index number for the fracture node.

5. The building drainage pipe network anomaly detection system according to claim 3, characterized in that, The flow resistance classification module includes: The trend curve extraction submodule extracts the corresponding continuous water pressure time series data based on the spatial sudden drop verification mark set, identifies the water pressure breakpoint position in the time series and limits the data segments before and after the mark, extracts the continuous water pressure sequence at adjacent measuring point positions, and reconstructs the time series through time index to generate a water pressure change trend data sequence. The path feature calculation submodule calls the water pressure change trend data sequence, retrieves continuous data segments before and after the point of sudden drop, delineates the time index intervals of the water pressure rise and fall segments, calculates the duration of the corresponding segments based on the index difference, calculates the difference based on the water pressure values ​​within the segments, and generates a set of water pressure path change feature parameters. The type label classification submodule selects the duration of the rising segment and the slope of the falling segment as classification indicators based on the set of water pressure path change characteristic parameters. It sets the time span threshold and slope judgment benchmark value for the classification of sudden drop chain segment types, compares the relationship between the classification indicators and the corresponding judgment benchmark, determines the type of the sudden drop chain segment and assigns the corresponding type label, and generates a flow resistance structure type identification table.

6. The building drainage pipe network anomaly detection system according to claim 5, characterized in that, The structure mapping module includes: The node coordinate retrieval submodule matches and extracts the corresponding node coordinate number in the building structure plan and elevation layout model based on the sudden drop chain segment index recorded in the flow resistance structure type identification table, calls the three-dimensional coordinate data associated with the node number, integrates the spatial position and associated number of the component to which the node belongs, and generates an abnormal node coordinate information set. The path structure positioning submodule calls the abnormal node coordinate information set, parses the floor index number of the node in the building structure model and the connection relationship between the nodes on the upper and lower floors, combines the node grouping information set in the building hierarchy structure, determines whether adjacent nodes are located in the same drainage zone or cross-floor passage, and identifies the connection order on the vertical path according to the sequence relationship of the upstream and downstream nodes, and obtains the vertical pipe segment connection sequence data. The spatial attribution labeling submodule reads the pipe segment identifier and floor number corresponding to the node based on the vertical pipe segment connection sequence data, retrieves the drainage direction field recorded in the flow direction information, compares it with the upstream and downstream positions of the node, determines the floor position of the abnormal node in the drainage path, and matches it with the structural classification fields of the main pipe and branch pipe in the building drainage, marks the corresponding pipe segment type code, and generates a list of spatial attribution information for abnormal nodes.

7. The building drainage pipe network anomaly detection system according to claim 1, characterized in that, The system also includes an attribution verification module: The attribution verification module extracts the response time records of the node and the two upstream and downstream liquid level monitoring points based on the list of spatial attribution information of the abnormal node. It determines whether the liquid level change of the node conforms to the response order from upstream to downstream when the abnormal event occurs. If the downstream response is delayed, the attribution location is verified to be reasonable. If the response time is advanced, the adjacent gradient chain segment is re-extracted for reverse tracing and judgment, and the set of abnormal verification nodes of the pipeline network is output. The pipeline anomaly verification node set includes nodes with successful attribution verification, nodes with failed attribution verification, nodes with reverse tracing results, and response order consistency identifiers.

8. The building drainage pipe network anomaly detection system according to claim 7, characterized in that, The attribution verification module includes: The response order judgment submodule calls the list of spatial affiliation information of the abnormal node, extracts the abnormal node and the associated upstream and downstream liquid level monitoring point numbers, retrieves the liquid level monitoring data records with the corresponding numbers, determines the start time of the liquid level change of the three monitoring points within the time period corresponding to the abnormal event, compares the response order of the upstream and downstream monitoring points through the timestamp field, filters the node group whose downstream response time is after the upstream, and generates the liquid level response order verification result. The source tracing and update submodule identifies downstream node numbers whose response times are earlier than those of upstream monitoring points based on the response time distribution in the liquid level response sequence verification results. It retrieves the neighboring gradient chain segment information corresponding to the node, extracts the continuous connection nodes of the upstream pipe segment through reverse retrieval, determines whether there are liquid level fluctuation records in the real-time upstream nodes that match the time period of the abnormal event, filters the node numbers that meet the conditions, and generates a set of pipeline network abnormal verification nodes.

Images