A multi-scene-based sewage pipe network management information intelligent classification method and system

Abstract

The application discloses a multi-scene-based sewage pipe network management information intelligent classification method and system, relates to the technical field of pipe network monitoring, and comprises the following steps: acquiring multi-source monitoring data and pipe network topology structure data of a sewage pipe network, and determining the scene categories corresponding to the multi-source monitoring data; constructing a node connection relationship graph based on the pipe network topology structure data, extracting scene association information from the multi-source monitoring data according to the node connection relationship graph and the scene categories; performing topological weighted fusion on the scene association information of different nodes according to the upstream and downstream topological relationships in the node connection relationship graph, and generating a pipe network global state representation; and determining the pipe network management information categories corresponding to the multi-source monitoring data according to the pipe network global state representation and preset classification boundary conditions. The application solves the problem that the classification result is unreasonable due to the neglect of the pipe network physical structure, and improves the accuracy and reliability of sewage pipe network management information classification.

Description

Technical Field

[0001] This application relates to the field of pipeline monitoring technology, and in particular to an intelligent classification method and system for sewage pipeline management information based on multiple scenarios. Background Technology

[0002] Currently, wastewater pipe network management systems employ various data analysis methods to classify monitoring information, including rule-based classification based on threshold judgment, pattern recognition based on statistical analysis, and intelligent classification based on machine learning. These methods can, to some extent, identify the operational status of the pipe network and classify management information. However, when processing wastewater pipe network monitoring data, these methods typically analyze data from each monitoring point as independent samples, neglecting the physical topology and hydraulic propagation characteristics of the pipe network, resulting in classification results lacking physical rationality. Furthermore, existing methods struggle to distinguish data characteristics under different operational scenarios, mixing data from routine operations, rainfall-affected scenarios, and abnormal events, thus reducing classification accuracy. Additionally, the lack of a verification mechanism based on pipe network connectivity may lead to logical contradictions in classification results between upstream and downstream nodes, affecting the reliability of management decisions.

[0003] Therefore, the relevant technologies have the problem that when classifying sewage pipe network monitoring information, they ignore the pipe network topology and hydraulic propagation laws, resulting in classification results that lack physical rationality and cannot be adapted to different operating scenarios. Summary of the Invention

[0004] In view of the aforementioned problems, this application is hereby filed.

[0005] Therefore, this application provides an intelligent classification method and system for sewage pipe network management information based on multiple scenarios, which can solve the problems mentioned in the background art.

[0006] To solve the above-mentioned technical problems, this application provides the following technical solution:

[0007] In a first aspect, this application provides an intelligent classification method for sewage pipe network management information based on multiple scenarios, comprising: acquiring multi-source monitoring data and pipe network topology data of a sewage pipe network; determining the scenario category corresponding to the multi-source monitoring data based on the spatiotemporal distribution parameters of the multi-source monitoring data and the pipe network topology data; constructing a node connectivity graph based on the pipe network topology data; extracting scenario association information from the multi-source monitoring data based on the node connectivity graph and the scenario category; wherein the scenario association information includes node hydraulic propagation information and node state evolution information; performing topological weighted fusion on the scenario association information of different nodes based on the upstream and downstream topological relationships in the node connectivity graph to generate a global state representation of the pipe network; and determining the pipe network management information category corresponding to the multi-source monitoring data based on the global state representation of the pipe network and preset classification boundary conditions.

[0008] Preferably, determining the scenario category corresponding to the multi-source monitoring data based on the spatiotemporal distribution parameters of the multi-source monitoring data and the pipeline network topology data includes: extracting monitoring time and monitoring location identifiers from the multi-source monitoring data; searching for the corresponding pipe segment attributes and node types in the pipeline network topology data based on the monitoring location identifiers; obtaining rainfall data for the corresponding time period based on the monitoring time; if the rainfall data is zero and the pipe segment attribute is a sewage pipe segment, then the scenario category is determined as a normal operation scenario; if the rainfall data is greater than a preset rainfall threshold and the node type is a merging node, then the scenario category is determined as a combined sewer overflow scenario; if the flow rate change rate in the multi-source monitoring data exceeds a preset change threshold, then the scenario category is determined as an abnormal event scenario.

[0009] Preferably, the step of extracting scene association information from the multi-source monitoring data based on the node connectivity graph and the scene category includes: if the scene category is a normal operation scene, then based on the flow path in the node connectivity graph, calculating the flow transmission delay between adjacent nodes, and extracting the first node hydraulic propagation information from the multi-source monitoring data based on the flow transmission delay; if the scene category is a combined sewer overflow scene, then based on the confluence node in the node connectivity graph, calculating the impact range of rainwater inflow on downstream nodes, and extracting the second node hydraulic propagation information from the multi-source monitoring data based on the impact range; if the scene category is an abnormal event scene, then based on the node connectivity graph, tracing the propagation path of the abnormal signal, and extracting the third node hydraulic propagation information from the multi-source monitoring data based on the propagation path; and using the first node hydraulic propagation information, the second node hydraulic propagation information, or the third node hydraulic propagation information as the node hydraulic propagation information in the scene association information.

[0010] Preferably, the step of performing topological weighted fusion of the scene association information of different nodes based on the upstream and downstream topological relationships in the node connectivity graph includes: determining the upstream contributing node set and the downstream influencing node set of each node based on the node connectivity graph; calculating the flow contribution ratio of each node in the upstream contributing node set to the current node based on the hydraulic balance principle; weighting and aggregating the scene association information of the upstream contributing node set according to the flow contribution ratio to obtain upstream aggregation information; calculating the influence attenuation factor of the current node on the downstream influencing node set based on the pipe segment length and pipe diameter parameters in the node connectivity graph; generating the fusion state vector of the current node according to the influence attenuation factor and the upstream aggregation information; and combining the fusion state vectors of all nodes to obtain the global state representation of the pipeline network.

[0011] Preferably, determining the pipeline management information category corresponding to the multi-source monitoring data based on the pipeline global state representation and preset classification boundary conditions includes: inputting the pipeline global state representation into a pre-constructed classification model and outputting an initial classification vector; identifying, based on the node connectivity graph, confluence control nodes with more than a preset connection threshold and terminal nodes with one connected pipe segment; extracting the classification components corresponding to the confluence control nodes and the terminal nodes from the initial classification vector; if the difference between the classification components corresponding to the confluence control nodes and the classification components corresponding to the terminal nodes exceeds a preset consistency threshold, then performing consistency correction on the initial classification vector based on the propagation path of the node connectivity graph to obtain a corrected classification vector; and determining the pipeline management information category based on the corrected classification vector and the preset classification boundary conditions.

[0012] Preferably, the consistency correction of the initial classification vector based on the propagation path of the node connectivity graph includes: calculating the path transfer function from the confluence control node to the terminal node along the propagation path of the node connectivity graph; calculating the theoretical classification value after the classification component of the confluence control node is transferred to the terminal node according to the path transfer function; calculating the deviation between the theoretical classification value and the classification component corresponding to the terminal node; if the deviation is greater than a preset deviation threshold, recursively correcting the classification components of each node on the propagation path according to the path transfer function; if the deviation is less than or equal to the preset deviation threshold, keeping the initial classification vector unchanged.

[0013] Preferably, after determining the pipeline management information category corresponding to the multi-source monitoring data based on the pipeline global state representation and preset classification boundary conditions, the method further includes: acquiring pipeline operation feedback data within a preset time period after determining the pipeline management information category; calculating the matching degree between the classification result and the actual operation state based on the pipeline operation feedback data and the pipeline management information category; if the matching degree is lower than a preset matching threshold, adjusting the topological weight of the corresponding node in the node connectivity graph based on the abnormal node distribution in the pipeline operation feedback data, and updating the classification boundary conditions.

[0014] Secondly, this application also provides an intelligent classification system for sewage pipe network management information based on multiple scenarios, including:

[0015] Thirdly, this application also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to perform the following steps:

[0016] The process involves acquiring multi-source monitoring data and network topology data of a sewage pipe network. Based on the spatiotemporal distribution parameters of the multi-source monitoring data and the network topology data, the scenario category corresponding to the multi-source monitoring data is determined. A node connectivity graph is constructed based on the network topology data. Based on the node connectivity graph and the scenario category, scenario association information is extracted from the multi-source monitoring data. This scenario association information includes node hydraulic propagation information and node state evolution information. Based on the upstream and downstream topological relationships in the node connectivity graph, topological weighted fusion is performed on the scenario association information of different nodes to generate a global network state representation. Finally, based on the global network state representation and preset classification boundary conditions, the network management information category corresponding to the multi-source monitoring data is determined.

[0017] Fourthly, this application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, performs the following steps:

[0018] The process involves acquiring multi-source monitoring data and network topology data of a sewage pipe network. Based on the spatiotemporal distribution parameters of the multi-source monitoring data and the network topology data, the scenario category corresponding to the multi-source monitoring data is determined. A node connectivity graph is constructed based on the network topology data. Based on the node connectivity graph and the scenario category, scenario association information is extracted from the multi-source monitoring data. This scenario association information includes node hydraulic propagation information and node state evolution information. Based on the upstream and downstream topological relationships in the node connectivity graph, topological weighted fusion is performed on the scenario association information of different nodes to generate a global network state representation. Finally, based on the global network state representation and preset classification boundary conditions, the network management information category corresponding to the multi-source monitoring data is determined.

[0019] Implementing this application has the following beneficial effects: This application provides an intelligent classification method and system for sewage pipe network management information based on multiple scenarios. By deeply integrating the pipe network topology into the information classification process, it establishes for the first time a scenario-related information extraction mechanism based on node connectivity graphs. This makes the classification process no longer a simple processing of isolated data points, but fully considers the hydraulic propagation relationships and influence ranges between nodes in the pipe network. This structured information processing method fundamentally changes the limitations of traditional classification methods. By introducing flow contribution ratio calculation and influence attenuation factors based on the hydraulic balance principle, topology weighted fusion that conforms to the physical characteristics of the pipe network is achieved, so that the generated global state representation of the pipe network can truly reflect the overall operating status of the pipe network. In addition, this application verifies the logical consistency of the classification results between upstream and downstream nodes by calculating the path transfer function. This classification correction under physical constraints not only improves the classification accuracy, but also ensures that the classification results conform to the hydraulic propagation laws of the pipe network, providing a reliable basis for subsequent management decisions. Compared with the simple data classification or rule judgment of existing technologies, this application achieves a deep integration of data-driven and physical mechanisms, significantly improving the rationality and interpretability of the classification results while maintaining classification efficiency. Attached Figure Description

[0020] To more clearly illustrate the technical solutions in the embodiments of this application or related technologies, the drawings used in the description of the embodiments of this application or related technologies will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0021] Figure 1 This is a schematic diagram of a multi-scenario intelligent classification method for sewage pipe network management information involved in this application;

[0022] Figure 2This is a flowchart illustrating the consistency correction of the initial classification vector in an intelligent classification method for sewage pipe network management information based on multiple scenarios, as described in this application.

[0023] Figure 3 This is a schematic diagram of the overall structure of an intelligent classification system for sewage pipe network management information based on multiple scenarios, as described in this application.

[0024] Figure 4 This is a computer device diagram of a smart classification method for sewage pipe network management information based on multiple scenarios, which is the subject of this application. Detailed Implementation

[0025] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0026] Wastewater pipe networks are a crucial component of urban infrastructure, and their operational status directly impacts urban drainage safety and environmental protection. Wastewater pipe network management information encompasses various types, including network operational status information, fault warning information, and maintenance decision-making information. Accurate information classification is fundamental to achieving refined management. Wastewater pipe network monitoring systems collect multi-source data such as flow rate, water level, and water quality through sensors deployed at various nodes within the network. This data reflects the operational characteristics of the pipe network under different temporal and spatial conditions.

[0027] Currently, wastewater pipe network information classification mainly employs three methods: threshold-based rule-based classification, statistical pattern recognition, and machine learning-based intelligent classification. Rule-based classification methods judge monitoring data by setting fixed thresholds, offering simplicity but lacking flexibility. Statistical methods build classification models by analyzing the distribution characteristics of historical data, handling data uncertainty but struggling to adapt to complex scenarios. Machine learning methods achieve automatic classification by training classifiers, possessing strong adaptability but typically treating each monitoring point as an independent sample, neglecting the physical structural constraints of the pipe network.

[0028] Wastewater pipe networks possess unique topological characteristics. The state of upstream nodes influences downstream nodes through hydraulic propagation; this spatial interconnectivity is an essential attribute of pipe network systems. Furthermore, pipe networks exhibit different operational patterns under different scenarios. During normal operation, flow rates show periodic variations; during rainfall, combined sewer overflows occur; and during abnormal events, abrupt signal changes propagate along the network. These physical characteristics dictate that pipe network information classification cannot simply apply general data classification methods; it requires a thorough consideration of the network's topological constraints and hydraulic mechanisms.

[0029] Furthermore, the reliability of pipeline monitoring data also depends on the verification of the physical rationality of the classification results. Due to factors such as sensor failure and communication interference, monitoring data may exhibit anomalies. If the classification results are physically unreasonable, such as the upstream being normal while the downstream suddenly experiences anomalies, the accuracy of the classification needs to be reassessed. Existing methods lack this verification mechanism based on physical constraints, which may lead to classification results that violate the basic laws of pipeline operation.

[0030] Based on the above, this application provides an intelligent classification method for sewage pipe network management information based on multiple scenarios. It integrates the pipe network topology into the classification process by constructing a node connectivity graph, and extracts corresponding scenario-related information, including node hydraulic propagation information and state evolution information, according to different scenario categories. The method calculates the flow contribution ratio between nodes based on the hydraulic balance principle, achieving topology-weighted fusion that conforms to physical laws. The consistency of the classification results is verified through path transfer functions to ensure the logical rationality of the classification results for upstream and downstream nodes. The classification boundary conditions and topology weights are iteratively optimized to continuously improve classification performance. This achieves the effect of providing accurate and physically reasonable information classification while fully utilizing the constraints of the pipe network's physical structure, thereby supporting refined management decisions for sewage pipe networks.

[0031] According to the embodiments of this application, an embodiment of an intelligent classification method for sewage pipe network management information based on multiple scenarios is provided. It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer device with data processing capabilities, such as a sewage pipe network monitoring center server, an edge computing device, etc. Furthermore, although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in a different order than that shown here.

[0032] This embodiment provides an intelligent classification method for sewage pipe network management information based on multiple scenarios, which can be used in the aforementioned computer equipment. Figure 1 This is a flowchart of an intelligent classification method for sewage pipe network management information based on multiple scenarios, according to an embodiment of this application. Figure 1 As shown, the process includes steps 202 to 208:

[0033] Step 202: Obtain multi-source monitoring data and network topology data of the sewage pipe network. Based on the spatiotemporal distribution parameters of the multi-source monitoring data and the network topology data, determine the scenario category corresponding to the multi-source monitoring data.

[0034] Specifically, the multi-source monitoring data for the sewage pipe network comes from sensor devices deployed at various monitoring points in the network, including real-time data collected by monitoring instruments such as flow meters, water level gauges, and water quality analyzers. The multi-source monitoring data includes various types of data, such as flow rate data, water level data, water quality parameter data, and pressure data. Each data point is accompanied by a timestamp and monitoring point number. Flow rate data reflects the velocity and flow rate of sewage in the pipes; water level data reflects the pipe's fullness; water quality parameter data includes indicators such as chemical oxygen demand (COD), ammonia nitrogen concentration, and suspended solids concentration; and pressure data reflects the internal pressure status of the pipes.

[0035] Pipeline topology data is static data describing the physical connections of a sewage pipeline network, including pipe segment information, node information, and connection relationship information. Pipe segment information includes attributes such as segment number, diameter, length, material, and slope; node information includes attributes such as node number, node type, ground elevation, and manhole bottom elevation; connection relationship information describes the upstream and downstream connections between each pipe segment and node. Node types include manhole nodes, confluence nodes, branch nodes, pump station nodes, and discharge outlet nodes. Confluence nodes are nodes where multiple pipes converge; in combined sewer systems, they may also receive rainwater inflow.

[0036] Spatiotemporal distribution parameters are parameters extracted from multi-source monitoring data that characterize the distribution of data in time and space. Time-dimensional parameters include monitoring time, data acquisition frequency, and time series length; spatial-dimensional parameters include monitoring location identifiers, spatial coordinates of monitoring points, and the pipe section to which the monitoring point belongs. By analyzing spatiotemporal distribution parameters, the data acquisition environment and operating conditions can be determined.

[0037] Determining the scenario category requires a comprehensive analysis of the numerical characteristics of the monitoring data and the structural attributes of the pipe network. Wastewater pipe networks exhibit different hydraulic characteristics and data patterns under different operating scenarios. The main scenario categories include routine operation scenarios, combined sewer and stormwater scenarios, and abnormal event scenarios. Routine operation scenarios refer to the state where the pipe network operates normally according to design conditions under sunny conditions; combined sewer and stormwater scenarios refer to the state where rainwater enters the combined sewer system under rainfall conditions, causing a surge in flow; abnormal event scenarios refer to the state where the pipe network experiences abnormal situations such as blockage, leakage, or overflow.

[0038] Step 204: Construct a node connectivity graph based on the pipeline network topology data. Based on the node connectivity graph and scenario categories, extract scenario association information from multi-source monitoring data. This scenario association information includes node hydraulic propagation information and node state evolution information.

[0039] Specifically, a node connectivity graph is a graph theory representation of a pipeline network topology. It constructs a directed graph structure by using nodes in the pipeline network as vertices and pipe segments as edges. The construction process of the node connectivity graph includes: first, traversing all nodes in the pipeline network topology data and creating a graph vertex object for each node; then, traversing all pipe segments and establishing directed edges between corresponding vertices based on the starting and ending nodes of each segment; finally, assigning attribute information to each edge, including pipe segment length, pipe diameter, and flow direction. The completed node connectivity graph allows for quick querying of connectivity between nodes, shortest paths, and upstream / downstream relationships.

[0040] Scene association information reflects the propagation and evolution of the pipeline network's operational status among nodes under the scene category determined in step 202 (normal operation scenario, combined sewer overflow scenario, or abnormal event scenario). Node hydraulic propagation information describes the propagation process of water flow, pollutants, or abnormal signals in the pipeline network, including parameters such as propagation speed, propagation path, and propagation delay. Node state evolution information describes the changes in the node's operational status over time, including time-series data such as flow rate change curves, water level change curves, and water quality change curves.

[0041] Step 204, which involves extracting scene association information from multi-source monitoring data based on the node connectivity graph and scene category, requires different extraction methods for different scenes:

[0042] If the scenario category is a regular operation scenario, the flow transmission delay between adjacent nodes is calculated based on the flow path in the node connectivity graph, and the hydraulic propagation information of the first node is extracted from the multi-source monitoring data according to the flow transmission delay.

[0043] Specifically, in normal operating scenarios, sewage flows through the pipe network according to gravity. Flow propagation delay refers to the time required for water to flow from an upstream node to a downstream node. By calculating the flow propagation delay of each pipe segment, the hydraulic propagation time between any two nodes can be determined. In normal operating scenarios, sewage flows through the pipe network according to gravity. Flow propagation delay is the ratio of pipe segment length to average flow velocity. The average flow velocity is calculated using Manning's formula in hydraulics. Manning's formula is an empirical formula describing steady uniform flow in open channels and pipes, comprehensively considering three key parameters: the hydraulic radius, hydraulic gradient, and roughness coefficient of the pipe. The hydraulic radius is determined by the ratio of the pipe's cross-sectional area to its wetted perimeter. For circular sewage pipes, the hydraulic radius depends on the pipe diameter and the degree of filling. The hydraulic gradient, or the longitudinal slope of the pipe, reflects the ability of gravity to drive water flow. The roughness coefficient characterizes the roughness of the pipe's inner wall; the roughness coefficient for concrete pipes is typically 0.013-0.015, while for plastic pipes it is typically 0.009-0.011. By substituting the pipe segment length, diameter, slope, and roughness coefficient into Manning's formula, the average flow velocity of the pipe segment can be calculated, thus yielding the flow propagation time delay. By summing the flow propagation time delays of each pipe segment along the path, the hydraulic propagation time between any two nodes can be determined.

[0044] The process of extracting hydraulic propagation information at the first node is as follows: First, the time correspondence is determined based on the flow transmission delay, that is, the state of the upstream node at time t corresponds to the state of the downstream node at time t+ delay; then, the flow rate, water level and other data of each node at the corresponding time are extracted from the multi-source monitoring data; finally, the node state propagation matrix is ​​constructed, and the matrix elements represent the strength of the hydraulic propagation relationship between nodes.

[0045] If the scenario category is a combined stormwater and sewage scenario, the impact range of rainwater inflow on downstream nodes is calculated based on the confluence nodes in the node connectivity graph, and the hydraulic propagation information of the second node is extracted from the multi-source monitoring data according to the impact range.

[0046] Specifically, in combined sewer overflow scenarios, rainwater enters the pipe network through storm drains and merging nodes, causing a surge in flow in downstream pipe sections. The impact range refers to the set of downstream nodes affected by the rainwater inflow. The impact range is calculated as follows: starting from the merging node, a breadth-first search is performed along the node connectivity graph until a predetermined search depth is reached or a pipe section with sufficient capacity is encountered. The search depth is determined based on the pipe network size, typically involving 3-5 downstream nodes.

[0047] The extraction process of hydraulic propagation information for the second node is as follows: First, identify all confluence nodes and nodes within their influence range. Then, calculate the rainwater inflow. The determination of rainwater inflow requires comprehensive consideration of three main factors: rainfall intensity, reflecting the amount of rainfall per unit time, usually measured in millimeters per hour; catchment area, referring to the surface area where rainwater collects at the node, determined based on the service area of ​​the pipeline network and topography; and runoff coefficient, characterizing the proportion of rainfall converted into surface runoff, which is related to the type of surface paving. The runoff coefficient for hardened surfaces is 0.8-0.9, and for green areas it is 0.1-0.3. The rainwater inflow is calculated by comprehensively considering these three factors; the greater the rainfall intensity, the larger the catchment area, and the higher the runoff coefficient, the greater the rainwater inflow. Next, based on the principle of hydraulic balance, calculate the flow increment at each node within the influence range. Finally, extract the hydraulic response characteristics of the affected nodes from multi-source monitoring data. The flow mutation characteristic is obtained by comparing flow data before and after rainfall. For example, the average flow rate 30 minutes before the start of rainfall is calculated as the baseline flow rate, and the peak flow rate during rainfall is calculated. The difference between the two is the flow mutation amplitude, and the time difference between the peak flow rate and the start of rainfall is the response delay. The water level rise characteristic is obtained by analyzing water level time series data. Specifically, the initial water level before rainfall is recorded, the water level rise process during rainfall is monitored, and the highest water level value, water level rise rate, and water level recovery time are extracted. The flow mutation amplitude and the water level rise amplitude together constitute the hydraulic propagation information of the second node, reflecting the degree of influence of rainwater inflow on the node's operating status.

[0048] If the scenario category is an abnormal event scenario, the propagation path of the abnormal signal is traced based on the node connectivity graph, and the hydraulic propagation information of the third node is extracted from the multi-source monitoring data according to the propagation path.

[0049] Specifically, in abnormal event scenarios, abnormal signals propagate from the point of occurrence to the surrounding area. Abnormal signals include sudden changes in flow rate, abnormal water level, and deteriorating water quality. The identification of abnormal signals is based on the degree to which monitoring data deviates from normal operating ranges. For example, abnormal flow rate refers to a flow rate value exceeding the historical average for the same period by more than or equal to three standard deviations; abnormal water level refers to a water level exceeding 80% of the pipeline's design fill level or falling below 20% of the normal operating level; and abnormal water quality refers to a pollutant concentration exceeding the upper limit of the normal range by 1.5 times.

[0050] The method for tracing the propagation path is as follows: First, perform time series analysis on the multi-source monitoring data, sort the monitoring data of each node by timestamp, and detect the timing of anomalies by using a sliding time window. The time window length can be set (e.g., 5 minutes). The node that detects the anomaly first is identified as the anomaly source. Then, starting from the anomaly source node, search for subsequent nodes with anomalies downstream according to the topology of the node connectivity graph. By comparing the time sequence of anomalies appearing in adjacent nodes, the direction of anomaly propagation is determined. Finally, arrange the nodes that the anomaly passes through in sequence according to time to form the anomaly propagation path.

[0051] The extraction process of hydraulic propagation information at the third node is as follows: Anomaly characteristics are analyzed node by node along the propagation path. The time of anomaly occurrence is determined by the timestamp of the first time the monitoring data exceeds the normal range; the amplitude of anomaly is obtained by calculating the difference between the peak value of the anomaly and the normal baseline value, which can be taken as the average of the data in the hour preceding the anomaly; the duration of anomaly is the time from the onset of the anomaly to the return to the normal range. The propagation speed of the anomaly signal is calculated by the ratio of the distance between adjacent nodes to the time difference of the anomaly occurrence; the attenuation degree is determined by comparing the rate of change of the anomaly amplitude at upstream and downstream nodes, and the attenuation degree is equal to the ratio of the anomaly amplitude at the downstream node to the anomaly amplitude at the upstream node. Based on the extracted anomaly feature parameters, a two-dimensional coordinate system is constructed, with the horizontal axis representing time and the vertical axis representing the node position along the propagation path. The time of anomaly occurrence and the amplitude of anomaly at each node are marked in the coordinate system, forming a spatiotemporal distribution map of anomaly propagation, which intuitively shows the propagation process of the anomaly signal along the pipeline network.

[0052] The hydraulic propagation information of the first node, the second node, or the third node is used as the node hydraulic propagation information in the scene association information. Based on the actual scene category, the corresponding node hydraulic propagation information is selected as the input for subsequent processing.

[0053] Step 206: Based on the upstream and downstream topological relationships in the node connectivity graph, perform topological weighted fusion on the scene association information of different nodes to generate a global state representation of the pipeline network.

[0054] Specifically, upstream and downstream topology refers to the relative positions of nodes in a pipeline network, with sewage flowing from upstream nodes to downstream nodes. Topology weighted fusion is a process of weighting and combining the scenario association information of each node based on the physical structure of the pipeline network, with the weights determined by the strength of the hydraulic connection between nodes.

[0055] Step 206 involves performing topological weighted fusion of scene association information for different nodes based on the upstream and downstream topological relationships in the node connectivity graph, including steps B1 to B6:

[0056] Step B1: Determine the set of upstream contributing nodes and the set of downstream influencing nodes for each node based on the node connectivity graph.

[0057] Specifically, for each node in the node connectivity graph, the upstream contributing node set refers to all upstream nodes that directly or indirectly supply wastewater to the current node. The upstream contributing node set is determined by traversing backwards along the incoming edges from the current node to find all reachable upstream nodes. The downstream influencing node set refers to all downstream nodes that directly or indirectly receive wastewater from the current node. The downstream influencing node set is determined by traversing forwards along the outgoing edges from the current node to find all reachable downstream nodes. The traversal depth is limited by the network size and typically does not exceed level 5.

[0058] Step B2: Based on the principle of hydraulic balance, calculate the flow contribution ratio of each node in the upstream contributing node set to the current node.

[0059] Specifically, the hydraulic balance principle states that the inflow to a node equals the outflow plus the change in storage capacity. Under steady-state conditions, the inflow to a node equals the outflow. The flow contribution ratio represents the degree to which an upstream node contributes to the total inflow to the current node. The calculation method is as follows: First, obtain all paths from each node in the upstream contributing node set to the current node; then, calculate the flow transmission efficiency of each path, which is determined by comprehensively evaluating the transmission capacity of each pipe segment along the path. The transmission efficiency of a single pipe segment is positively correlated with the pipe diameter, with a larger diameter resulting in stronger transmission capacity; it is negatively correlated with the pipe segment length, with longer segments resulting in greater transmission losses. For multiple pipe segments connected in series, the flow received by the downstream pipe segment depends on the transmission efficiency of the upstream pipe segment. By calculating the transmission loss segment by segment, the overall transmission efficiency of the entire path is obtained; next, calculate the total flow contribution of each upstream node through all paths; finally, normalize the flow contribution of each node by dividing the flow contribution of each node by the sum of the flow contributions of all nodes, so that the sum of the flow contribution ratios of all nodes equals 1, resulting in a standardized flow contribution ratio.

[0060] Step B3: Based on the traffic contribution ratio, perform weighted aggregation of the scenario association information of the upstream contributing node set to obtain upstream aggregated information.

[0061] Specifically, weighted aggregation combines the scenario-related information of multiple upstream nodes according to their traffic contribution ratio. The process is as follows: for each data dimension of the scenario-related information, the value of each upstream node in that dimension is multiplied by its corresponding traffic contribution ratio. Then, all the products are summed to obtain the aggregated value for that dimension. The aggregated values ​​of all dimensions together constitute the upstream aggregated information. This upstream aggregated information includes the comprehensive impact of all upstream nodes on the current node, reflecting the overall state of the upstream.

[0062] Step B4: Based on the pipe segment length and pipe diameter parameters in the node connectivity graph, calculate the attenuation factor of the current node's influence on the downstream set of nodes.

[0063] Specifically, the attenuation factor represents the degree of attenuation when a state change at the current node propagates to downstream nodes. Attenuation is mainly caused by two factors: distance attenuation and capacity buffering. Distance attenuation is directly proportional to the pipe length; the longer the pipe, the greater the attenuation during propagation. Capacity buffering is inversely proportional to the pipe diameter; the larger the pipe diameter, the stronger the buffering capacity, and the greater the attenuation during propagation. The formula for calculating the attenuation factor is: Attenuation factor = exp(-α × pipe length / pipe diameter). The larger the ratio of pipe length to pipe diameter, the more significant the attenuation. The attenuation rate is controlled by the attenuation coefficient, where α is the attenuation coefficient. The attenuation coefficient is determined based on the pipe material roughness and the characteristics of the wastewater: smooth pipes have a smaller attenuation coefficient, while rough pipes have a larger attenuation coefficient; clean water has a smaller attenuation coefficient, while high-concentration wastewater has a larger attenuation coefficient.

[0064] Step B5: Generate the fusion state vector of the current node based on the influence attenuation factor and upstream aggregation information.

[0065] Specifically, the fused state vector is a multi-dimensional vector describing the overall state of a node. Each dimension corresponds to a type of monitoring parameter, such as flow rate, water level, and water quality concentration. The generation process of the fused state vector is as follows: First, upstream aggregated information is acquired, which reflects the cumulative impact of all upstream nodes on the current node. Then, local measured data is extracted from the monitoring equipment of the current node. Next, the upstream aggregated information and local measured data are fused. The fusion method is as follows: for flow rate and water volume-related parameters, a superposition method is used, that is, the local flow rate is equal to the sum of the upstream input flow rate and the local new flow rate. For water quality concentration parameters, a mixed calculation is used, and the concentration after mixing is determined according to the flow ratio of upstream inflow and local sewage. Finally, the fused parameter values ​​are combined to form the fused state vector, and the influence attenuation factor calculated in step B4 is saved as an additional attribute of this vector for subsequent attenuation calculation when propagating to downstream nodes.

[0066] Step B6: Combine the fused state vectors of all nodes to obtain the global state representation of the pipeline network.

[0067] Specifically, the global state representation of the pipeline network is an overall state description obtained by fusing information from all nodes, reflecting the global operational characteristics of the network. The global state representation is a set of fused state vectors from all nodes, which can be represented as a matrix, where rows correspond to nodes and columns correspond to state dimensions. To preserve spatial structure information, the topological positions of nodes within the network also need to be recorded. The global state representation comprehensively reflects the operational state of the pipeline network at the current moment, including the mutual influences between nodes and the overall pattern.

[0068] Step 208: Determine the category of pipeline management information corresponding to the multi-source monitoring data based on the global status representation of the pipeline network and the preset classification boundary conditions.

[0069] Specifically, the corrected classification vector includes the classification results of all nodes. The preset classification boundary conditions are a set of rules for determining the overall pipeline network management information category based on the node classification distribution. In this embodiment, the preset classification boundary conditions include four judgment rules: First, when all node classification components are normal category codes, the pipeline network management information category is determined to be normal operation information; Second, when a small number of nodes have abnormal category codes and the number of abnormal nodes does not exceed the preset warning ratio, it is determined to be warning information; Third, when the number of abnormal nodes exceeds the preset warning ratio but does not exceed the preset alarm ratio, or when a busbar control node malfunctions, it is determined to be alarm information; Fourth, when the number of abnormal nodes exceeds the preset alarm ratio, or multiple consecutive nodes malfunction simultaneously, or when the classification components contain fault feature codes, it is determined to be fault information. It should be noted that the preset warning ratio is the threshold for the proportion of abnormal nodes that triggers the warning state, determined based on the redundancy design capability of the pipeline network, reflecting the range of minor anomalies that the pipeline network can withstand; at the same time, the preset alarm ratio is the threshold for the proportion of abnormal nodes that triggers the alarm state, determined based on the safe operation standards of the pipeline network, reflecting the degree of anomalies requiring manual intervention. The early warning ratio is lower than the alarm ratio, and the specific values ​​for both are determined comprehensively based on the pipeline network scale, design standards, and management requirements. The final output pipeline management information categories are used to guide pipeline network operation and maintenance management decisions. Pipeline management information categories include normal operation information, early warning information, alarm information, and fault information. Normal operation information indicates that the pipeline network is operating within its design range; early warning information indicates that certain parameters are approaching thresholds and require attention; alarm information indicates that parameters have exceeded thresholds and require timely handling; fault information indicates that equipment failure or pipeline accidents have occurred and require emergency response.

[0070] Step 208 includes steps C1 to C4:

[0071] Step C1: Input the global state representation of the pipeline network into the pre-built classification model and output the initial classification vector.

[0072] Specifically, the pre-built classification model is a classifier trained on historical data, constructed using machine learning methods. The appropriate model structure is selected based on data characteristics and classification requirements; it can be a support vector machine, random forest, or neural network, among others. The input to the classification model is the global state representation matrix of the pipeline network, and the output is an initial classification vector. Each element of the initial classification vector corresponds to the classification result of a node, and its value is the encoding of the pipeline network management information category. The classification model learns the correspondence between state representations and management information categories in historical data to achieve classification prediction for new data.

[0073] Step C2: Based on the node connectivity diagram, identify the bus control nodes with more than the preset connection threshold and the end nodes with only one connection segment.

[0074] Specifically, the number of connected pipe segments refers to the total number of pipe segments directly connected to a node, including inflow and outflow segments. A preset connection threshold is determined based on the average connectivity of the pipe network and is set as an integer value greater than the average connectivity of the pipe network. Nodes with a number of connected pipe segments greater than this threshold are identified as confluence control nodes. Confluence control nodes play a role in flow distribution and aggregation in the pipe network, and have a significant impact on downstream areas. Terminal nodes are the end nodes of the pipe network, having only one inflow segment and no outflow segment, or only one outflow segment connected to the discharge outlet. The identification process is achieved by traversing the node connectivity graph and counting the degree of each node.

[0075] Step C3: Extract the classification components corresponding to the confluence control nodes and the classification components corresponding to the terminal nodes from the initial classification vector.

[0076] Specifically, the categorical components are the element values ​​corresponding to specific nodes in the initial classification vector. The classification results of the confluence control nodes and terminal nodes are extracted from the initial classification vector using node indices. The dissimilarity between the categorical components is calculated, and this dissimilarity can be measured using Hamming distance or Euclidean distance.

[0077] The logic for determining classification consistency is as follows:

[0078] If the difference between the classification component corresponding to the confluence control node and the classification component corresponding to the terminal node exceeds a preset consistency threshold, the initial classification vector is corrected for consistency based on the propagation path of the node connectivity graph to obtain a corrected classification vector. The preset consistency threshold is determined based on the total number of classification categories and is set as the maximum allowed number of category differences; for example, it can be set to 1, meaning a maximum of one category difference is allowed. When the difference exceeds the preset consistency threshold, it indicates that the classification result is spatially discontinuous, violating the physical propagation laws of the pipeline network, and correction is required.

[0079] If the difference between the classification component corresponding to the confluence control node and the classification component corresponding to the terminal node is less than or equal to a preset consistency threshold, the initial classification vector is used as the correction classification vector. The classification results are basically consistent in space, conforming to physical laws, and require no correction.

[0080] like Figure 2 The diagram shows a flowchart of the consistency correction of the initial classification vector based on the propagation path of the node connectivity graph in step C3, specifically including steps C31 to C33:

[0081] Step C31: Calculate the path transfer function from the confluence control node to the terminal node along the propagation path of the node connectivity graph.

[0082] Specifically, the propagation path is a directed path from the confluence control node to the terminal node, and there may be multiple such paths. The path transfer function describes the transformation relationship of the state propagation along the path. For each path, the path transfer function is a composite of the transfer functions of each pipe segment along the path. The pipe segment transfer function is determined based on the pipe segment properties, taking into account factors such as flow attenuation, time delay, and mixing dilution. The path transfer function can be in the form of a linear transformation matrix or a nonlinear mapping function.

[0083] Step C32: Calculate the theoretical classification value of the classification component of the confluence control node after it is passed to the end node according to the path transfer function.

[0084] Specifically, the classification component of the confluence control node is used as input, and the path transfer function is applied to obtain the classification value that should theoretically be passed to the end node. When multiple paths exist, the transfer results of multiple paths are combined: if the results of each path are consistent, the result is directly adopted; if the results are inconsistent, a weighted sum is performed based on the traffic share of each path, with the path with the larger traffic share having a greater impact.

[0085] Step C33: Calculate the deviation between the theoretical classification value and the classification component corresponding to the terminal node.

[0086] Specifically, bias is calculated by comparing theoretical classification values ​​with actual classification components. For classification problems, bias can be the number of inconsistent categories or the difference in probability distributions.

[0087] The deviation judgment and processing logic is as follows:

[0088] If the deviation exceeds a preset deviation threshold, the classification components of each node along the propagation path are recursively corrected according to the path transfer function. The preset deviation threshold is determined based on the classification accuracy requirements, and its value is between 0 and 1; the smaller the value, the lower the tolerance. The recursive correction method is as follows: starting from the confluence control node, the transfer function is applied node by node along the propagation path to update the classification components of each node on the path, so that the classification result conforms to the propagation law.

[0089] If the deviation is less than or equal to the preset deviation threshold, the initial classification vector remains unchanged. If the deviation is within an acceptable range, no correction is needed.

[0090] Step C4: Determine the category of pipeline management information based on the corrected classification vector and the preset classification boundary conditions.

[0091] Specifically, the corrected classification vector contains the classification results of all nodes. Preset classification boundary conditions define the judgment rules for each category. For example: when most (e.g., over 80%) nodes are classified as normal, the overall classification is determined as normal operation information; when any node is classified as faulty, the overall classification is determined as fault information; other cases are determined as early warning or alarm information based on the number and distribution of abnormal nodes. The final output pipeline management information categories are used to guide pipeline operation and maintenance management decisions.

[0092] Furthermore, after step 208, this application also includes the following steps:

[0093] After determining the category of pipeline management information, obtain pipeline operation feedback data within a preset time period. The preset time period is determined based on the pipeline response time, typically 24–48 hours, to collect sufficient feedback information. Pipeline operation feedback data includes information such as actual operation and maintenance events, processing results, and changes in pipeline status.

[0094] Based on pipeline operation feedback data and pipeline management information categories, the matching degree between the classification results and the actual operating status is calculated. The matching degree calculation process is as follows: First, the actual operation and maintenance response types are extracted from the pipeline operation feedback data. These types include no action required, routine inspection, maintenance handling, and emergency repair, corresponding to four management information categories: normal operation, early warning, alarm, and fault, respectively. Then, a correspondence is established: if the classified pipeline management information category corresponds to the actual operation and maintenance response type, it is considered a match; otherwise, it is considered a mismatch. Next, the matching status of all classified events within a preset time period is statistically analyzed. Finally, the matching degree is calculated, which equals the number of matched events divided by the total number of events. The matching degree ranges from 0 to 1; a higher value indicates higher classification accuracy.

[0095] The feedback adjustment logic is as follows:

[0096] If the matching degree is lower than the preset matching threshold, the topological weights of the corresponding nodes in the node connectivity graph are adjusted based on the distribution of abnormal nodes in the pipeline operation feedback data, and the classification boundary conditions are updated. The preset matching threshold is determined according to management requirements, usually set to 0.85, corresponding to an 85% accuracy requirement. The method for adjusting the topological weights is to increase the weights of correctly classified nodes and decrease the weights of incorrectly classified nodes. The classification boundary conditions are updated by re-statistically analyzing to determine new threshold parameters.

[0097] If the matching degree is greater than or equal to the preset matching threshold, the topological weights and classification boundary conditions in the node connectivity graph remain unchanged. The classification effect meets the requirements and no adjustment is needed.

[0098] The wastewater pipe network management information classification method provided in this embodiment, by introducing pipe network topology constraints, achieves information classification that conforms to physical laws, solves the problem of unreasonable classification caused by neglecting spatial correlation in traditional methods, and improves the scientificity and accuracy of wastewater pipe network management.

[0099] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0100] Based on the same inventive concept, this application also provides an intelligent classification system for sewage pipe network management information based on multiple scenarios. The solution provided by this system is similar to the solution described in the above method. Therefore, the specific limitations of one or more embodiments of the intelligent classification system for sewage pipe network management information based on multiple scenarios provided below can be found in the limitations of the intelligent classification method for sewage pipe network management information based on multiple scenarios above, and will not be repeated here.

[0101] In one exemplary embodiment, such as Figure 3 As shown, a multi-scenario-based intelligent classification system for sewage pipe network management information is provided, including:

[0102] The scene recognition module is used to acquire multi-source monitoring data and network topology data of the sewage pipe network, and determine the scene category corresponding to the multi-source monitoring data based on the spatiotemporal distribution parameters of the multi-source monitoring data and the network topology data.

[0103] The information extraction module is used to construct a node connectivity graph based on pipeline network topology data, and extract scene association information from multi-source monitoring data according to the node connectivity graph and scene category; the scene association information includes node hydraulic propagation information and node state evolution information.

[0104] The topology fusion module is used to perform topology weighted fusion of scene association information of different nodes based on the upstream and downstream topology relationships in the node connectivity graph, and generate a global state representation of the pipeline network.

[0105] The classification decision module is used to determine the category of pipeline management information corresponding to multi-source monitoring data based on the global status representation of the pipeline network and the preset classification boundary conditions.

[0106] The modules in the aforementioned intelligent classification system for sewage pipe network management information based on multiple scenarios can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device, or stored in the memory of a computer device as software, so that the processor can call and execute the corresponding operations of each module.

[0107] In one exemplary embodiment, a computer device is provided, which may be a terminal, and its internal structure diagram may be as follows: Figure 4 As shown, the computer device includes a processor, memory, input / output interfaces, a communication interface, a display unit, and an input device. The processor, memory, and input / output interfaces are connected via a system bus, and the communication interface, display unit, and input device are also connected to the system bus via the input / output interfaces. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The input / output interfaces are used for exchanging information between the processor and external devices. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, mobile cellular networks, Near Field Communication (NFC), or other technologies. When the computer program is executed by the processor, it implements an intelligent classification method for sewage pipe network management information based on multiple scenarios. The display unit is used to form a visually visible image and can be a display screen, a projection device, or a virtual reality imaging device. The display screen can be an LCD screen or an e-ink screen. The input device of the computer device can be a touch layer covering the display screen, or buttons, trackballs, or touchpads set on the casing of the computer device, or external keyboards, touchpads, or mice, etc.

[0108] Those skilled in the art will understand that Figure 4 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0109] In one embodiment, a computer device is also provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above method embodiments.

[0110] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon that, when executed by a processor, implements the steps in the above method embodiments.

[0111] In one embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps in the above method embodiments.

[0112] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data must comply with relevant regulations.

[0113] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, artificial intelligence (AI) processors, etc., and are not limited to these.

[0114] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this application.

[0115] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A method for intelligent classification of sewage pipe network management information based on multiple scenarios, characterized in that, include: Acquire multi-source monitoring data and network topology data of the sewage pipe network, and determine the scene category corresponding to the multi-source monitoring data based on the spatiotemporal distribution parameters of the multi-source monitoring data and the network topology data; Based on the pipeline network topology data, a node connectivity graph is constructed. According to the node connectivity graph and the scenario category, scenario association information is extracted from the multi-source monitoring data. The scenario association information includes node hydraulic propagation information and node state evolution information. Based on the upstream and downstream topological relationships in the node connectivity graph, topological weighted fusion is performed on the scene association information of different nodes, including: Based on the node connectivity graph, determine the set of upstream contributing nodes and the set of downstream influencing nodes for each node; Based on the principle of hydraulic balance, the proportion of flow contribution of each node in the upstream contributing node set to the current node is calculated; The scenario association information of the upstream contributing node set is weighted and aggregated according to the traffic contribution ratio to obtain upstream aggregated information; Based on the pipe segment length and pipe diameter parameters in the node connectivity graph, calculate the influence attenuation factor of the current node on the downstream influencing node set. Based on the influence attenuation factor and the upstream aggregation information, generate the fusion state vector of the current node; The fused state vectors of all nodes are combined to obtain the global state representation of the pipeline network; Based on the global status representation of the pipeline network and the preset classification boundary conditions, the category of pipeline management information corresponding to the multi-source monitoring data is determined.

2. The intelligent classification method for sewage pipe network management information based on multiple scenarios as described in claim 1, characterized in that: The step of determining the scene category corresponding to the multi-source monitoring data based on the spatiotemporal distribution parameters of the multi-source monitoring data and the pipeline network topology data includes: Extract the monitoring time and monitoring location identifiers from the multi-source monitoring data; Based on the monitoring location identifier, the corresponding pipe segment attribute and node type are searched in the pipeline topology data; Rainfall data for the corresponding time period is obtained based on the monitoring time. If the rainfall data is zero and the pipe segment attribute is a sewage pipe segment, then the scenario category is determined as a normal operation scenario; If the rainfall data is greater than the preset rainfall threshold and the node type is a merging node, then the scenario category is determined to be a combined sewer overflow scenario. If the rate of change of traffic in the multi-source monitoring data exceeds a preset change threshold, the scenario category will be determined as an abnormal event scenario.

3. The intelligent classification method for sewage pipe network management information based on multiple scenarios as described in claim 2, characterized in that: The step of extracting scene association information from the multi-source monitoring data based on the node connectivity graph and the scene category includes: If the scenario category is a normal operation scenario, then based on the flow path in the node connectivity graph, the flow transmission delay between adjacent nodes is calculated, and the hydraulic propagation information of the first node is extracted from the multi-source monitoring data according to the flow transmission delay. If the scenario category is a combined stormwater and sewage scenario, then based on the confluence node in the node connectivity graph, the impact range of rainwater inflow on downstream nodes is calculated, and the hydraulic propagation information of the second node is extracted from the multi-source monitoring data according to the impact range. If the scenario category is an abnormal event scenario, the propagation path of the abnormal signal is traced based on the node connectivity graph, and the hydraulic propagation information of the third node is extracted from the multi-source monitoring data according to the propagation path. The hydraulic propagation information of the first node, the hydraulic propagation information of the second node, or the hydraulic propagation information of the third node are used as the node hydraulic propagation information in the scene association information.

4. The intelligent classification method for sewage pipe network management information based on multiple scenarios as described in claim 3, characterized in that: The step of determining the pipeline management information category corresponding to the multi-source monitoring data based on the global state representation of the pipeline network and preset classification boundary conditions includes: The global state representation of the pipeline network is input into a pre-built classification model, and an initial classification vector is output. Based on the node connectivity diagram, identify the bus control nodes with a number of connected pipe segments greater than a preset connection threshold and the end nodes with a number of connected pipe segments of 1; Extract the classification components corresponding to the confluence control node and the end node from the initial classification vector; If the difference between the classification component corresponding to the confluence control node and the classification component corresponding to the end node exceeds a preset consistency threshold, then the initial classification vector is corrected for consistency based on the propagation path of the node connectivity graph to obtain a corrected classification vector. The category of pipeline management information is determined based on the corrected classification vector and the preset classification boundary conditions.

5. The intelligent classification method for sewage pipe network management information based on multiple scenarios as described in claim 4, characterized in that: The consistency correction of the initial classification vector based on the propagation path of the node connectivity graph includes: Calculate the path transfer function from the confluence control node to the terminal node along the propagation path of the node connectivity graph; Based on the path transfer function, calculate the theoretical classification value of the classification component of the confluence control node after it is transferred to the end node; Calculate the deviation between the theoretical classification value and the classification component corresponding to the terminal node; If the deviation is greater than the preset deviation threshold, the classification components of each node on the propagation path are recursively corrected according to the path transfer function. If the deviation is less than or equal to the preset deviation threshold, the initial classification vector remains unchanged.

6. The intelligent classification method for sewage pipe network management information based on multiple scenarios as described in claim 5, characterized in that: After determining the pipeline management information category corresponding to the multi-source monitoring data based on the pipeline global state representation and preset classification boundary conditions, the method further includes: Obtain pipeline operation feedback data within a preset time period after determining the category of pipeline management information; Based on the pipeline network operation feedback data and the pipeline network management information categories, calculate the matching degree between the classification results and the actual operating status; If the matching degree is lower than the preset matching threshold, the topological weight of the corresponding node in the node connectivity graph is adjusted according to the distribution of abnormal nodes in the pipeline operation feedback data, and the classification boundary conditions are updated; if the matching degree is greater than or equal to the preset matching threshold, the topological weight in the node connectivity graph and the classification boundary conditions remain unchanged.

7. A multi-scenario intelligent classification system for sewage pipe network management information, employing the multi-scenario intelligent classification method for sewage pipe network management information as described in any one of claims 1 to 6, characterized in that, include: The scene recognition module is used to acquire multi-source monitoring data and network topology data of the sewage pipe network, and determine the scene category corresponding to the multi-source monitoring data based on the spatiotemporal distribution parameters of the multi-source monitoring data and the network topology data. The information extraction module is used to construct a node connectivity graph based on the pipeline network topology data, and extract scene association information from the multi-source monitoring data according to the node connectivity graph and the scene category; wherein, the scene association information includes node hydraulic propagation information and node state evolution information; The topology fusion module is used to perform topology weighted fusion on the scene association information of different nodes according to the upstream and downstream topology relationships in the node connectivity graph, and generate a global state representation of the pipeline network. The classification decision module is used to determine the category of pipeline management information corresponding to the multi-source monitoring data based on the global status representation of the pipeline network and preset classification boundary conditions.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the intelligent classification method for sewage pipe network management information based on multiple scenarios as described in any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the intelligent classification method for sewage pipe network management information based on multiple scenarios as described in any one of claims 1 to 6.

Images