Drainage pipe network abnormity tracing method and electronic equipment

Abstract

The invention relates to a drainage pipe network abnormity tracing method and electronic equipment, and the method comprises the steps: determining a water quality analysis time sequence and a flow analysis time sequence of any monitoring point according to an initial water quality monitoring time sequence and an initial flow time sequence of the monitoring point; performing multi-dimensional statistical analysis according to the water quality analysis time sequence and the flow analysis time sequence of the monitoring point site, and determining a target abnormal event of the monitoring point site; according to the time range of the target abnormal event of the monitoring point position and the topological relation of the monitoring point positions of the target drainage pipe network, an expected abnormal time window of each monitoring point position on the upstream of the monitoring point position is determined; and determining the source monitoring point position of the target abnormal event of the monitoring point position according to the expected abnormal time window of each monitoring point position at the upstream of the monitoring point position. The method can give full play to the advantages of the drainage pipe network on-line monitoring in time frequency and space density, and improves the accuracy, reliability and efficiency of abnormal event recognition.

Description

Technical Field

[0001] This disclosure relates to the field of water pollution management technology, and in particular to a method and electronic equipment for tracing the source of anomalies in drainage pipe networks. Background Technology

[0002] In existing technologies, numerous monitoring devices are deployed in urban drainage pipe networks to monitor water quality and flow (or volume) data in real time and at high frequency. This monitoring data provides a rich and reliable foundation for real-time understanding of the water conditions within the drainage network, accurate diagnosis of the network's health status, and detection of abnormal events such as excessive discharge or unauthorized sewage collection. However, existing methods for analyzing this monitoring data often struggle to accurately capture abnormal characteristics, resulting in low accuracy in identifying abnormal events and low efficiency in tracing their source, thus failing to fully leverage the advantages of high-frequency monitoring. Summary of the Invention

[0003] In view of this, this disclosure proposes a method for tracing the source of anomalies in drainage pipe networks and a technical solution for electronic equipment.

[0004] According to one aspect of this disclosure, a method for tracing the source of anomalies in a drainage network is provided, comprising: for any monitoring point in the target drainage network, determining the water quality analysis time series and flow analysis time series of the monitoring point based on the initial water quality monitoring time series and the initial flow time series of the monitoring point, wherein the initial water quality monitoring time series of the monitoring point includes the initial monitoring value time series of at least one water quality indicator of the monitoring point; and performing multi-dimensional statistical analysis based on the water quality analysis time series and flow analysis time series of the monitoring point to determine the target anomaly event of the monitoring point, wherein the multi-dimensional statistical analysis includes... The method combines the instantaneous state dimension and the evolutionary state dimension to perform statistical analysis on the monitoring values ​​of at least one water quality indicator in the water quality analysis time series and the monitoring values ​​of flow rate in the flow analysis time series. A target abnormal event at any monitoring point indicates that both water quality abnormality and flow rate abnormality occur at that monitoring point. Based on the time range of the target abnormal event at the monitoring point and the topological relationship of the monitoring points in the target drainage network, the expected abnormal time window of each monitoring point upstream of the monitoring point is determined. Based on the expected abnormal time window of each monitoring point upstream of the monitoring point, the source monitoring point of the target abnormal event at the monitoring point is determined.

[0005] In one possible implementation, determining the water quality analysis time series and flow analysis time series for any monitoring point in the target drainage network, based on the initial water quality monitoring time series and initial flow time series of that monitoring point, includes: for any monitoring point, performing data preprocessing based on the initial water quality monitoring time series and initial flow time series of that monitoring point to determine the processed water quality monitoring time series and processed flow time series of that monitoring point; performing time series decomposition and periodic term removal based on the processed water quality monitoring time series of that monitoring point to determine the water quality analysis time series of that monitoring point; and performing time series decomposition and periodic term removal based on the processed flow monitoring time series of that monitoring point to determine the flow analysis time series of that monitoring point.

[0006] In one possible implementation, the step of performing multi-dimensional statistical analysis based on the water quality analysis time series and flow analysis time series of the monitoring point to determine the target abnormal event of the monitoring point includes: for any monitoring point, performing instantaneous state anomaly analysis and evolution state anomaly analysis based on the water quality analysis time series of the monitoring point to determine the water quality anomaly data in the water quality analysis time series of the monitoring point; performing instantaneous state anomaly analysis and evolution state anomaly analysis based on the flow analysis time series of the monitoring point to determine the flow anomaly data in the flow analysis time series of the monitoring point; and determining the target abnormal event of the monitoring point based on the water quality anomaly data and flow anomaly data of the monitoring point.

[0007] In one possible implementation, the step of performing instantaneous state anomaly analysis and evolutionary state anomaly analysis on the water quality analysis time series of any given monitoring point to determine the water quality anomaly data in the water quality analysis time series of that monitoring point includes: for any given monitoring point, if, in the water quality analysis time series of that monitoring point, there exist multiple water quality indicators whose monitoring values ​​at any given moment satisfy a first preset condition, then the monitoring values ​​of the first preset number of multiple water quality indicators at that moment are determined as the water quality anomaly data in the water quality analysis time series of that monitoring point. The first candidate anomaly data; based on the water quality analysis time series of the monitoring point, determine the rate of change evaluation result of each water quality indicator at any time at the monitoring point; if the rate of change evaluation results of multiple water quality indicators at any time at the monitoring point meet the second preset condition, determine the monitoring value of the multiple water quality indicators at that time as the second preset number of multiple water quality indicators as the second candidate anomaly data in the water quality analysis time series of the monitoring point; based on the first candidate anomaly data and the second candidate anomaly data of the monitoring point, determine the water quality anomaly data in the water quality analysis time series of the monitoring point.

[0008] In one possible implementation, the step of performing instantaneous state anomaly analysis and evolutionary state anomaly analysis based on the flow analysis time series of the monitoring point to determine the abnormal flow data in the flow analysis time series of the monitoring point includes: for any monitoring point, if the monitoring value at any moment in the flow analysis time series of the monitoring point meets a third preset condition, determining the monitoring value at that moment as the third candidate abnormal data in the flow analysis time series of the monitoring point; determining the flow change rate evaluation result at any moment of the monitoring point based on the flow analysis time series of the monitoring point; if the flow change rate evaluation result at any moment of the monitoring point meets a fourth preset condition, determining the monitoring value at that moment as the fourth candidate abnormal data in the flow analysis time series of the monitoring point; and determining the abnormal flow data in the flow analysis time series of the monitoring point based on the third and fourth candidate abnormal data.

[0009] In one possible implementation, determining the target abnormal event of the monitoring point based on the abnormal water quality data and abnormal flow data of the monitoring point includes: for any given monitoring point, determining the target abnormal data of the monitoring point based on the abnormal water quality data and abnormal flow data of the monitoring point; and, if there are multiple target abnormal data points at the monitoring point that are consecutive in time and exceed a preset number, determining the target abnormal event of the monitoring point based on the multiple target abnormal data points.

[0010] In one possible implementation, determining the source monitoring point of the target abnormal event of the monitoring point based on the expected abnormal time window of each monitoring point upstream of the monitoring point includes: determining the number of candidate abnormal events corresponding to the target abnormal event of the monitoring point based on the expected abnormal time window of each monitoring point upstream of the monitoring point; and determining the source monitoring point of the target abnormal event of the monitoring point based on the number of candidate abnormal events corresponding to the target abnormal event of the monitoring point.

[0011] In one possible implementation, determining the source monitoring point of the target abnormal event based on the number of candidate abnormal events corresponding to the target abnormal event at the monitoring point includes: for any given monitoring point, if the number of candidate abnormal events corresponding to the target abnormal event meets a fifth preset condition, determining the monitoring point as the source monitoring point of the target abnormal event; if the number of candidate abnormal events corresponding to the target abnormal event meets a sixth preset condition, performing a multi-dimensional similarity analysis on the target abnormal event and all its corresponding candidate abnormal events to determine the source monitoring point of the target abnormal event, wherein the multi-dimensional similarity analysis includes: combining data waveform dimension, data amplitude dimension, and time dimension to perform a similarity analysis on the target abnormal event at any given monitoring point and each of its corresponding candidate abnormal events.

[0012] In one possible implementation, the step of performing a multi-dimensional similarity analysis on the target anomaly and all its corresponding candidate anomalies to determine the source monitoring point of the target anomaly when the number of candidate anomalies corresponding to the target anomaly meets a sixth preset condition includes: for any one monitoring point, when the number of candidate anomalies corresponding to the target anomaly meets the sixth preset condition, determining the multi-dimensional similarity analysis results between the target anomaly and each of its corresponding candidate anomalies; determining the confidence evaluation result of each candidate anomaly based on the multi-dimensional similarity analysis results between the target anomaly and each of its corresponding candidate anomalies; determining the source monitoring point of the target anomaly when the confidence evaluation results of all candidate anomalies do not meet the confidence threshold; and determining the target candidate anomaly based on all candidate anomalies that meet the confidence threshold when the confidence evaluation results of at least one candidate anomaly meet the confidence threshold, and determining the monitoring point corresponding to the target candidate anomaly as the source monitoring point of the target anomaly.

[0013] In one possible implementation, the method further includes: determining the rainfall event analysis results of the target drainage network based on the initial rainfall time series of the target drainage network, wherein the rainfall event analysis results are used to indicate the rainfall time and rainfall type of the target drainage network; and for any target abnormal event at a monitoring point, performing a causal analysis on the target abnormal event based on the rainfall event analysis results to determine the causal analysis results of the target abnormal event.

[0014] In one possible implementation, the step of performing a causal analysis on a target anomaly event at any monitoring point, based on the rainfall event analysis results, and determining the causal analysis result of the target anomaly event, includes: for a target anomaly event at any monitoring point, determining a rainfall correlation analysis result of the target anomaly event based on the water quality analysis time series and flow analysis time series of the monitoring point and the rainfall event analysis results, or based on the water quality analysis time series and flow analysis time series of the source monitoring point of the target anomaly event and the rainfall event analysis results, wherein the rainfall correlation analysis result indicates the degree of correlation between the target anomaly event and rainfall; and determining the causal analysis result of the target anomaly event based on the rainfall correlation analysis result of the target anomaly event.

[0015] According to another aspect of this disclosure, an electronic device is provided, including a memory, a processor, and a computer program stored in the memory, wherein the processor executes the computer program to implement the steps of the above-described method.

[0016] In this embodiment of the disclosure, for any monitoring point in the target drainage network, the water quality analysis time series and flow analysis time series of the monitoring point can be determined based on the initial water quality monitoring time series and the initial flow time series, which include the initial monitoring value time series of at least one water quality indicator. Based on the water quality analysis time series and flow analysis time series of the monitoring point, multi-dimensional statistical analysis can be performed on the monitoring values ​​of each water quality indicator in the water quality analysis time series and the monitoring values ​​of the flow in the flow analysis time series, combining the instantaneous state dimension and the evolutionary state dimension. This fully utilizes the data characteristics of the water quality analysis time series and flow analysis time series of the monitoring point to comprehensively capture water quality anomalies and flow anomalies at the monitoring point, thereby identifying target abnormal events where water quality and flow anomalies occur simultaneously at the monitoring point, and improving the accuracy and reliability of abnormal event identification. Based on the time range of the target anomaly at the monitoring point and the topological relationship of the monitoring points in the target drainage network, the expected anomaly time window for each upstream monitoring point can be determined. This narrows the time range required for upstream source tracing of the target anomaly at any given monitoring point, enabling comprehensive analysis of anomaly source tracing in both time and spatial dimensions. It fully leverages the advantages of water quality and flow analysis time series at each monitoring point in terms of temporal frequency, as well as the spatial density of monitoring points in the target drainage network, thereby improving the efficiency and accuracy of anomaly source tracing. Based on the expected anomaly time window of each upstream monitoring point, the source monitoring point of the target anomaly at that monitoring point can be determined. On this basis, with the source monitoring point of the target anomaly at that monitoring point as the center, targeted and rapid pollution source tracing can be carried out within the monitoring range of that monitoring point, further improving the efficiency and accuracy of anomaly source tracing. Compared with the common analysis methods for drainage pipe network monitoring data in the prior art, the embodiments of this disclosure can give full play to the advantages of online monitoring of drainage pipe networks in terms of time frequency and spatial density. It is not only suitable for identifying and tracing abnormal events in the long-term operation of the target drainage pipe network, but also can provide a reliable basis for real-time pipeline health diagnosis of the target drainage pipe network.

[0017] Other features and aspects of this disclosure will become clear from the following detailed description of exemplary embodiments with reference to the accompanying drawings. Attached Figure Description

[0018] The accompanying drawings, which are included in and form part of this specification, illustrate exemplary embodiments, features, and aspects of this disclosure together with the specification and serve to explain the principles of this disclosure.

[0019] Figure 1 A flowchart illustrating a method for tracing the source of an anomaly in a drainage pipe network according to an embodiment of the present disclosure is shown.

[0020] Figure 2 A schematic diagram showing the topology of a drainage network according to an embodiment of the present disclosure is provided.

[0021] Figure 3 A block diagram of an electronic device according to an embodiment of the present disclosure is shown. Detailed Implementation

[0022] Various exemplary embodiments, features, and aspects of this disclosure will now be described in detail with reference to the accompanying drawings. The same reference numerals in the drawings denote elements that have the same or similar functions. Although various aspects of the embodiments are shown in the drawings, they are not necessarily drawn to scale unless specifically indicated otherwise.

[0023] As used herein, the terms “comprising,” “including,” “having,” or variations thereof are open-ended and include one or more of the stated features, integrals, elements, steps, components, or functions, but do not exclude the presence or addition of one or more other features, integrals, elements, steps, components, functions, or groups thereof.

[0024] When an element is referred to as “connected,” “coupled,” “responding,” or a variation thereof relative to another element, it may be directly connected, coupled, or responding to another element, or there may be an intermediate element present.

[0025] Although the terms first, second, third, etc., may be used herein to describe various elements / operations, these elements / operations should not be limited by these terms. These terms are only used to distinguish one element / operation from another. Therefore, without departing from the teachings of the inventive concept, a first element / operation in some embodiments may be referred to as a second element / operation in other embodiments.

[0026] The term “exemplary” as used herein means “serving as an example, embodiment, or illustration.” Any embodiment illustrated herein as “exemplary” is not necessarily to be construed as superior to or better than other embodiments.

[0027] Furthermore, to better illustrate this disclosure, numerous specific details are set forth in the following detailed description. Those skilled in the art will understand that this disclosure can be practiced without certain specific details. In some instances, methods, means, components, and circuits well known to those skilled in the art have not been described in detail in order to highlight the main points of this disclosure.

[0028] It should be noted that the information (including but not limited to user device information, user personal information, etc.), data (including but not limited to data used for analysis, data stored, data displayed, etc.) and signals involved in this application are all authorized by the user or fully authorized by all parties, and the collection, use and processing of related data must comply with the relevant laws, regulations and standards of the relevant regions.

[0029] In existing technologies, numerous monitoring devices are deployed in urban sewage drainage networks to monitor water quality and flow (or volume) data in real time and at high frequency. This monitoring data provides a rich and reliable foundation for real-time understanding of the water conditions within the drainage network, accurate diagnosis of the network's health status, and detection of abnormal events such as excessive or illegal sewage discharge. However, the time it takes for monitoring data to reflect water quality anomalies is somewhat delayed compared to the actual time of pollution occurrence. By the time a monitoring device detects a pollutant, it may have already diffused and mixed within the drainage network, causing the monitored data to lose the unique spatial distribution characteristics of the pollution source.

[0030] Existing methods for analyzing this monitoring data often struggle to accurately capture anomalous features, resulting in low accuracy in identifying abnormal events within drainage networks and low efficiency in tracing and locating sources. This fails to fully leverage the advantages of high-frequency, high-density water quality monitoring in drainage networks. For example, a common existing method for water quality anomaly monitoring and analysis is the inversion method based on diffusion models such as hydrodynamic-water quality models. This method relies on parameters such as pollution source intensity and hydrological boundary conditions. However, in actual drainage network environments, these parameters exhibit high dynamism and uncertainty, causing the model output to easily deviate from the actual pollution characteristics. This is especially true in application scenarios where drainage networks are located in complex terrain or where dynamic pollution sources exist, where the analytical capability of this method is weak. Another common existing method for water quality anomaly monitoring and analysis is machine learning methods such as anomaly detection algorithms. However, this method requires a large amount of sample data, making it difficult to guarantee the accuracy of the machine learning model. Furthermore, this method has poor adaptability to drainage networks in various complex environments.

[0031] On the other hand, common monitoring and analysis methods in existing technologies are prone to missing the "peak" characteristics of pollutants in cases of instantaneous or intermittent emissions, while these characteristic signals are usually very important for the analysis and source tracing of anomalies in drainage pipe networks.

[0032] In view of this, this disclosure provides a method for tracing the source of anomalies in drainage pipe networks. This method fully leverages the advantages of online monitoring of drainage pipe networks in terms of temporal frequency and spatial density, comprehensively capturing the abnormal data characteristics in the initial water quality monitoring time series and initial flow time series of each monitoring point in the target drainage pipe network, thereby improving the accuracy, reliability, and efficiency of anomaly event identification. The method for tracing the source of anomalies in drainage pipe networks provided in this disclosure is described in detail below.

[0033] Figure 1 A flowchart illustrating a method for tracing the source of drainage network anomalies according to an embodiment of this disclosure is shown. This method can be executed by an electronic device such as a terminal device or a server. The terminal device can be a user equipment (UE), mobile device, user terminal, terminal, cellular phone, cordless phone, personal digital assistant (PDA), handheld device, computing device, vehicle-mounted device, wearable device, etc. The method can be implemented by a processor calling computer-readable instructions stored in memory. Alternatively, the method can be executed by a server. Figure 1 As shown, the methods for tracing the source of anomalies in this drainage network include:

[0034] In step S101, for any monitoring point in the target drainage network, the water quality analysis time series and flow analysis time series of the monitoring point are determined based on the initial water quality monitoring time series and the initial flow time series of the monitoring point. The initial water quality monitoring time series of the monitoring point includes the initial monitoring value time series of at least one water quality indicator of the monitoring point.

[0035] The target drainage network mentioned here can refer to any type of drainage network, and its specific form can be flexibly set according to actual usage needs, such as rainwater drainage network, sewage drainage network, or combined sewer system, etc. This disclosure does not impose specific limitations on it. The specific number of monitoring points in the target drainage network can be flexibly set according to actual usage needs, and this disclosure does not impose specific limitations on it; any monitoring point can refer to the location in the target drainage network where water quality monitoring equipment and flow monitoring equipment are deployed. The specific forms of the water quality monitoring equipment and flow monitoring equipment can be flexibly set according to actual usage needs, and this disclosure does not impose specific limitations on it.

[0036] In one example, the water quality monitoring device can be a spectral sensor, and the flow monitoring device can be a flow sensor. The spectral sensor can acquire the monitoring values ​​corresponding to the water quality indicators at the designated monitoring points based on a preset acquisition frequency. The spectral sensor enables online, in-situ, high-frequency, and real-time measurement, increasing the acquisition frequency from once a day to 3-60 minutes per acquisition, resulting in high monitoring efficiency and accuracy. Therefore, it can obtain water quality indicator monitoring values ​​at a higher frequency, effectively capturing the water quality characteristics of the drainage network. The spectral sensor is preferably a quantum dot spectral probe. The acquisition frequency of the spectral sensor can be flexibly set according to actual usage requirements, for example, it can be set to 5-30 minutes per acquisition, 8-20 minutes per acquisition, 10-15 minutes per acquisition, etc., and this disclosure does not specifically limit it in this way.

[0037] The initial water quality monitoring time series for any monitoring point may include the initial monitoring value time series of at least one water quality indicator at that monitoring point; the initial monitoring value time series of any water quality indicator may reflect the change of the monitoring value of that water quality indicator at that monitoring point over time, and its specific form may be flexibly set according to actual usage needs, and this disclosure does not impose specific limitations on it. The specific number and type of water quality indicators may be flexibly set according to actual usage needs, and this disclosure does not impose specific limitations on it.

[0038] In one possible implementation, water quality indicators may include at least one of the following: Chemical Oxygen Demand (COD), conductivity, ammonia nitrogen, five-day biochemical oxygen demand, total phosphorus, total nitrogen, suspended solids, total dissolved solids, petroleum hydrocarbons, pH (acidity / alkalinity), anionic surfactants, cyanide, sulfide, fluoride, chloride, organophosphorus compounds, sulfate, mercury, chromium, cadmium, arsenic, lead, nickel, beryllium, silver, selenium, copper, zinc, manganese, iron, volatile phenols, benzene series compounds, aniline compounds, nitrobenzene, and water temperature. The monitored values ​​for each water quality indicator can constitute a corresponding initial monitoring value time series.

[0039] The initial flow time series of any monitoring point can reflect the change of flow in the pipeline at that monitoring point over time. Its specific form can be flexibly set according to actual usage requirements, and this disclosure does not impose specific limitations on it.

[0040] Typically, the initial water quality and flow time series data from any given monitoring point may contain various defects, such as missing data, single-point noise, or abnormal data due to equipment malfunctions. These data are unsuitable for direct use in identifying and tracing anomalies at that monitoring point. Therefore, preliminary data analysis, including data preprocessing and feature extraction, can be performed on the initial water quality and flow time series to determine standardized and saliently characteristic water quality and flow analysis time series. This improves the accuracy and reliability of subsequent anomaly identification and tracing.

[0041] The following section will describe, in conjunction with possible implementation methods of this disclosure, the specific process of determining the water quality analysis time series and flow analysis time series of any monitoring point based on the initial water quality monitoring time series and initial flow time series of any monitoring point, which will not be elaborated here.

[0042] In step S102, multi-dimensional statistical analysis is performed based on the water quality analysis time series and flow analysis time series of the monitoring point to determine the target abnormal event of the monitoring point. The multi-dimensional statistical analysis includes: combining the instantaneous state dimension and the evolution state dimension to perform statistical analysis on the monitoring values ​​of at least one water quality indicator in the water quality analysis time series and the monitoring values ​​of flow in the flow analysis time series. The target abnormal event of any monitoring point indicates that the monitoring point has both water quality abnormality and flow abnormality.

[0043] For any given monitoring point, the monitoring values ​​of each water quality indicator in the water quality analysis time series and the monitoring values ​​of the flow rate in the flow analysis time series can be analyzed by combining the instantaneous state dimension and the evolution state dimension. In this way, multi-dimensional statistical analysis can be used to identify abnormal events at the monitoring point and determine the target abnormal event in which both water quality abnormality and flow abnormality occur at the monitoring point.

[0044] The instantaneous state dimension can be used to reflect the monitoring values ​​of each water quality indicator in the water quality analysis time series, and the monitoring values ​​of the flow rate in the flow analysis time series, at each moment. The evolutionary state dimension can be used to reflect the changing trends of the monitoring values ​​of each water quality indicator in the water quality analysis time series and the monitoring values ​​of the flow rate in the flow analysis time series, such as the change process between any two adjacent moments.

[0045] By comprehensively analyzing the water quality and flow analysis time series of any monitoring point in both instantaneous and evolving state dimensions, this method is not limited to a single water quality or flow analysis, nor is it limited to anomaly analysis based solely on numerical values. It can fully utilize the data characteristics of the water quality and flow analysis time series of the monitoring point to comprehensively capture water quality and flow anomalies, thereby improving the accuracy and reliability of anomaly event identification.

[0046] The following text will describe the specific process of determining the target abnormal event at the monitoring point by conducting multi-dimensional statistical analysis based on the water quality analysis time series and flow analysis time series of the monitoring point, in conjunction with the possible implementation methods of this disclosure. It will not be elaborated here.

[0047] In step S103, based on the time range of the target abnormal event at the monitoring point and the topological relationship of the monitoring points in the target drainage network, the expected abnormal time window of each monitoring point upstream of the monitoring point is determined.

[0048] The time range of the target abnormal event at any monitoring point can include information such as the start time, end time and duration of the target abnormal event at that monitoring point. It can be flexibly set according to actual usage needs, and this disclosure does not make specific limitations on it.

[0049] The topological relationship of monitoring points in the target drainage network can be used to indicate the positional and distance relationships between each monitoring point in the target drainage network. Its specific form and content can be flexibly set according to actual usage needs, and this disclosure does not impose specific limitations on it.

[0050] Figure 2 A schematic diagram illustrating the topology of a drainage network according to an embodiment of the present disclosure is shown. Figure 2 As shown, the target drainage network includes 8 monitoring points, namely monitoring point 1 to monitoring point 8; monitoring points 1 to 3 and monitoring points 5 to 8 are all located upstream of monitoring point 4; monitoring points 1 and 2 are located upstream of monitoring point 3; monitoring points 6 to 8 are located upstream of monitoring point 5; and monitoring points 7 and 8 are located upstream of monitoring point 6.

[0051] For any given monitoring point, the expected anomaly time window for each upstream monitoring point can be determined based on the time range of the target anomaly event at that monitoring point and the topological relationship of the monitoring points in the target drainage network. The expected anomaly time window for any upstream monitoring point can be used to indicate the time range for investigating anomalies at that upstream monitoring point during the process of tracing the source of the target anomaly event.

[0052] Specifically, for any given monitoring point, the distance between that monitoring point and any upstream monitoring point can be determined based on the topological relationship of the monitoring points in the target drainage network. Combined with the preset flow velocity of the pipe segment where both monitoring points are located, the theoretical propagation time between them can be determined. Based on the time range of the target abnormal event at the monitoring point and the theoretical propagation time between the monitoring point and the upstream monitoring point, the expected abnormal time window for the upstream monitoring point can be determined. For example, the time difference between the start time of the expected abnormal time window of the upstream monitoring point and the start time of the time range of the target abnormal event at the downstream monitoring point can be equal to the theoretical propagation time, and the duration of the expected abnormal time window of the upstream monitoring point can be equal to the duration of the target abnormal event at the downstream monitoring point. The specific value of the preset flow velocity for any pipe segment can be flexibly set according to actual usage requirements; for example, it can be set to the average flow velocity of the pipe segment, etc., and this disclosure does not impose specific limitations on it. The method for determining the average flow velocity of any pipe segment can refer to the implementation methods in related technologies, and this disclosure does not specifically limit it. For example, flow velocity sensors can be set up at monitoring points to obtain flow velocity values ​​in real time, and the average flow velocity of the pipe segment can be determined based on the real-time flow velocities at multiple monitoring points.

[0053] Based on the above Figure 2 For example, the time range of the target abnormal event at monitoring point 4 is from 10:00 to 13:00. Based on the distance between monitoring point 4 and monitoring point 3, and the preset flow velocity of the pipe sections where monitoring point 4 and monitoring point 3 are located, the theoretical propagation time between monitoring point 4 and monitoring point 3 is determined to be 1 hour. Therefore, the expected abnormal time window for monitoring point 3 can be determined to be from 9:00 to 12:00, so as to check whether any abnormal event has occurred at monitoring point 3 within this time.

[0054] Through the above process, based on the topological relationship of the monitoring points in the target drainage network, the time range that needs to be investigated when tracing the upstream source of an abnormal event at any monitoring point can be narrowed down. This enables comprehensive analysis of abnormal event tracing in both time and space dimensions, giving full play to the advantages of the water quality analysis time series and flow analysis time series of each monitoring point in terms of time frequency, as well as the advantages of the monitoring points in the target drainage network in terms of spatial density, thereby improving the efficiency and accuracy of abnormal event tracing.

[0055] In step S104, the source monitoring point of the target abnormal event of the monitoring point is determined based on the expected abnormal time window of each monitoring point upstream of the monitoring point.

[0056] For any given monitoring point, targeted investigation of abnormal events can be conducted at each upstream monitoring point within the expected abnormal time window. This allows for the rapid identification of the source monitoring point of the target abnormal event at that monitoring point. Based on this, with the source monitoring point of the target abnormal event at that monitoring point as the center, targeted and rapid pollution source tracing can be carried out within the monitoring range of that monitoring point. This improves the efficiency and accuracy of abnormal event tracing and fully leverages the advantages of online monitoring of drainage pipe networks in terms of time frequency and spatial density. It is not only suitable for identifying and tracing abnormal events during the long-term operation of target drainage pipe networks, but also provides a reliable basis for real-time pipeline health diagnosis of target drainage pipe networks.

[0057] The following text will describe, in conjunction with the possible implementation methods of this disclosure, the specific process of determining the source monitoring point of the target abnormal event of the monitoring point based on the expected abnormal time window of each monitoring point upstream of the monitoring point, which will not be elaborated here.

[0058] In this embodiment of the disclosure, for any monitoring point in the target drainage network, the water quality analysis time series and flow analysis time series of the monitoring point can be determined based on the initial water quality monitoring time series and the initial flow time series, which include the initial monitoring value time series of at least one water quality indicator. Based on the water quality analysis time series and flow analysis time series of the monitoring point, multi-dimensional statistical analysis can be performed on the monitoring values ​​of each water quality indicator in the water quality analysis time series and the monitoring values ​​of the flow in the flow analysis time series, combining the instantaneous state dimension and the evolutionary state dimension. This fully utilizes the data characteristics of the water quality analysis time series and flow analysis time series of the monitoring point to comprehensively capture water quality anomalies and flow anomalies at the monitoring point, thereby identifying target abnormal events where water quality and flow anomalies occur simultaneously at the monitoring point, and improving the accuracy and reliability of abnormal event identification. Based on the time range of the target anomaly at the monitoring point and the topological relationship of the monitoring points in the target drainage network, the expected anomaly time window for each upstream monitoring point can be determined. This narrows the time range required for upstream source tracing of the target anomaly at any given monitoring point, enabling comprehensive analysis of anomaly source tracing in both time and spatial dimensions. It fully leverages the advantages of water quality and flow analysis time series at each monitoring point in terms of temporal frequency, as well as the spatial density of monitoring points in the target drainage network, thereby improving the efficiency and accuracy of anomaly source tracing. Based on the expected anomaly time window of each upstream monitoring point, the source monitoring point of the target anomaly at that monitoring point can be determined. On this basis, with the source monitoring point of the target anomaly at that monitoring point as the center, targeted and rapid pollution source tracing can be carried out within the monitoring range of that monitoring point, further improving the efficiency and accuracy of anomaly source tracing. Compared with the common analysis methods for drainage pipe network monitoring data in the prior art, the embodiments of this disclosure can give full play to the advantages of online monitoring of drainage pipe networks in terms of time frequency and spatial density. It is not only suitable for identifying and tracing abnormal events in the long-term operation of the target drainage pipe network, but also can provide a reliable basis for real-time pipeline health diagnosis of the target drainage pipe network.

[0059] In one possible implementation, for any monitoring point in the target drainage network, the water quality analysis time series and flow analysis time series of that monitoring point are determined based on the initial water quality monitoring time series and initial flow time series of that monitoring point. This includes: for any monitoring point, performing data preprocessing based on the initial water quality monitoring time series and initial flow time series of that monitoring point to determine the processed water quality monitoring time series and processed flow time series of that monitoring point; performing time series decomposition and periodic term removal based on the processed water quality monitoring time series of that monitoring point to determine the water quality analysis time series of that monitoring point; and performing time series decomposition and periodic term removal based on the processed flow monitoring time series of that monitoring point to determine the flow analysis time series of that monitoring point.

[0060] The specific methods for data preprocessing can be flexibly set according to actual usage needs, and this disclosure does not impose specific limitations on them.

[0061] In one possible implementation, data preprocessing may include at least one of the following: removal and interpolation of abnormal data from monitoring equipment, removal of abnormal operating data from monitoring equipment, processing of out-of-range data, smoothing and filtering, time resampling, and filling in missing data.

[0062] Specifically, for any given monitoring point, if the water quality monitoring equipment and / or flow monitoring equipment deployed at that point exhibits abnormal operation, the monitoring data falling within the time range of the abnormal operation in the initial water quality monitoring time series and initial flow time series for that monitoring point are removed, and interpolation methods are used to supplement the removed monitoring data. The specific content of the abnormal operation, and the specific methods for determining whether the water quality and flow monitoring equipment has experienced abnormal operation, depend on the actual form of the water quality and flow monitoring equipment and can be flexibly set according to actual usage requirements; this disclosure does not impose specific limitations on this. The specific form of the interpolation method can refer to implementation methods in related technologies; for example, Lagrange interpolation can be used, and this disclosure does not impose specific limitations on this.

[0063] For any given monitoring point, if the water quality monitoring equipment and / or flow monitoring equipment deployed at that point are operating normally but an abnormal condition occurs, the monitoring data within the time range of the abnormal condition in the initial water quality monitoring time series and initial flow time series of that monitoring point can be removed. The specific content of the abnormal condition and the specific methods for determining whether the water quality monitoring equipment and flow monitoring equipment have experienced an abnormal condition can be flexibly set according to actual usage requirements; this disclosure does not impose specific limitations on them.

[0064] In one example, abnormal operating conditions may include: equipment leaving the water. For any monitoring point, if, in the time series of the initial monitoring values ​​of a first preset water quality indicator at that monitoring point, there are multiple monitoring values ​​that are consecutively greater than a first quantity threshold and are all less than the in-water threshold, it can be determined that the water quality monitoring equipment deployed at that monitoring point has experienced equipment leaving the water. The specific content of the first preset water quality indicator can be flexibly set according to actual usage requirements; for example, it can be set to conductivity, etc., and this disclosure does not specifically limit it. The specific value of the first quantity threshold can also be flexibly set according to actual usage requirements, and this disclosure does not specifically limit it.

[0065] In one example, time series forecasting methods can be used to supplement the removed monitoring data. The specific form of the time series forecasting method can be found in implementation methods in related technologies; for example, an Autoregressive Integrated Moving Average (ARIMA) algorithm can be used, but this disclosure does not specifically limit it.

[0066] For any given monitoring point, considering the potential presence of significant single-point noise in the initial water quality and flow time series, a smoothing filter can be applied to these data to reduce the impact of noise on anomaly identification. The specific method for this smoothing filter can be found in relevant technical implementations, such as the Savitzky-Golay (SG) filtering algorithm; this disclosure does not impose any specific limitations on it.

[0067] For the same monitoring point, the sampling frequency for different water quality indicators and flow rates may vary; and for different monitoring points, the sampling frequency for the same water quality indicator or flow rate may also be incorrect due to the differences in the water quality monitoring equipment or flow monitoring equipment deployed.

[0068] Therefore, to improve the temporal consistency between the time series analysis of monitoring values ​​of all water quality indicators and the time series analysis of flow rate at the same monitoring point, as well as the temporal consistency between the time series analysis of monitoring values ​​of the same water quality indicators and the time series analysis of flow rate at different monitoring points, and to improve data standardization, thereby reducing the impact of noise data in the initial water quality monitoring time series and the initial flow rate time series on the identification of abnormal events, reducing the total amount of data processed, and improving the efficiency of abnormal event identification, for any monitoring point, the initial water quality monitoring time series and the initial flow rate time series of that monitoring point can be time-resampled according to a preset sampling frequency. The specific value of the preset sampling frequency can be flexibly set according to actual usage requirements, for example, it can be set to 1 hour, etc., and this disclosure does not specifically limit it. The specific method of time resampling processing can refer to the implementation methods in related technologies, such as the time series resampling algorithm based on Pandas, etc., and this disclosure does not specifically limit it.

[0069] Typically, the initial water quality monitoring time series and initial flow time series of any monitoring point may contain missing data due to various factors. Therefore, to improve data integrity and ensure the accuracy of anomaly identification, missing data filling processing can be performed on the initial water quality monitoring time series and initial flow time series of that monitoring point. The specific methods for missing data filling processing can refer to implementation methods in related technologies, such as forward filling methods, etc., and this disclosure does not specifically limit them.

[0070] Typically, water quality and flow rate changes in the target drainage network exhibit significant periodicity. For example, flow rate may peak between 7:00 AM and 7:00 PM daily, while water quality may peak between 7:00 PM and midnight. This periodicity affects the accuracy of identifying non-periodic anomalies at each monitoring point. For instance, an incident of illegal sewage discharge might be misidentified as a periodic phenomenon because the timing of the discharge coincides with a periodic pollution peak.

[0071] Therefore, to reduce the impact of the periodicity of water quality and flow changes in the target drainage network on the identification of abnormal events and improve the accuracy of abnormal event identification, for any monitoring point, time series decomposition and periodic term removal can be performed based on the processed water quality monitoring time series of that monitoring point to determine the water quality analysis time series. Similarly, time series decomposition and periodic term removal can be performed based on the processed flow monitoring time series of that monitoring point to determine the flow analysis time series. By removing the periodic terms obtained from the time series decomposition, the non-periodic data characteristics in the water quality analysis time series and flow analysis time series become more apparent. The specific method for time series decomposition can be found in related technical implementations, and this disclosure does not impose specific limitations on it.

[0072] In one example, a time series decomposition method can be used to decompose the processed water quality monitoring time series of any monitoring point into a trend analysis series, a periodic analysis series, and a residual analysis series. The time series decomposition method can be expressed as formula (1).

[0073]

[0074] in, This represents the monitoring value of the i-th water quality indicator at the k-th monitoring point in the target drainage network at time j. This represents the trend term of the i-th water quality indicator at the k-th monitoring point in the target drainage network at time j. This represents the periodicity term of the monitoring value of the i-th water quality indicator at the k-th monitoring point in the target drainage network at collection time j. This represents the residual term of the i-th water quality index at the k-th monitoring point in the target drainage network at time j.

[0075] Based on this, by removing periodic terms Only retain the trend item and residuals This allows us to obtain the time series analysis of the monitoring values ​​of the i-th water quality indicator at the k-th monitoring point.

[0076] In one possible implementation, multi-dimensional statistical analysis is performed based on the water quality analysis time series and flow analysis time series of the monitoring point to determine the target abnormal event of the monitoring point. This includes: for any monitoring point, performing instantaneous state anomaly analysis and evolution state anomaly analysis based on the water quality analysis time series of the monitoring point to determine the water quality anomaly data in the water quality analysis time series of the monitoring point; performing instantaneous state anomaly analysis and evolution state anomaly analysis based on the flow analysis time series of the monitoring point to determine the flow anomaly data in the flow analysis time series of the monitoring point; and determining the target abnormal event of the monitoring point based on the water quality anomaly data and flow anomaly data of the monitoring point.

[0077] Specifically, for any given monitoring point, instantaneous state anomaly analysis and evolutionary state anomaly analysis can be performed based on the water quality analysis time series of that point. This involves statistical analysis of two aspects: whether there are anomalies in the monitored values ​​of at least one water quality indicator at any given time, and whether there are anomalies in the trend of change. This identifies the water quality anomaly data in the water quality analysis time series of that monitoring point. When there are multiple water quality indicators, each indicator can be analyzed separately to obtain the corresponding water quality anomaly data.

[0078] In one possible implementation, for any given monitoring point, based on the water quality analysis time series of that monitoring point, instantaneous state anomaly analysis and evolutionary state anomaly analysis are performed to determine the water quality anomaly data in the water quality analysis time series of that monitoring point. This includes: for any given monitoring point, if, in the water quality analysis time series of that monitoring point, there exist multiple water quality indicators whose monitoring values ​​at any given moment in the analysis time series satisfy a first preset condition, then the monitoring values ​​of the first preset number of multiple water quality indicators at that moment are determined as the water quality anomaly data in the water quality analysis time series of that monitoring point. First candidate anomaly data; based on the water quality analysis time series of the monitoring point, determine the rate of change evaluation result of each water quality indicator at any time at the monitoring point; if the rate of change evaluation results of multiple water quality indicators at any time at the monitoring point meet the second preset condition, determine the monitoring value of the multiple water quality indicators at that time as the second preset number of monitoring points as the second candidate anomaly data in the water quality analysis time series of the monitoring point; based on the first candidate anomaly data and the second candidate anomaly data of the monitoring point, determine the water quality anomaly data in the water quality analysis time series of the monitoring point.

[0079] For any given monitoring point, the system can analyze the monitoring time series of all water quality indicators at that point, along with a first preset condition, to filter out the first candidate abnormal data with numerical anomalies in the water quality analysis time series of that monitoring point, thereby achieving instantaneous anomaly analysis. The specific content of the first preset condition can be flexibly set according to actual usage requirements, and this disclosure does not impose specific limitations on it.

[0080] In one example, for any monitoring point, the first preset condition may include a first preset threshold and a second preset threshold for all water quality indicators at the monitoring point, and the first preset threshold is greater than the second threshold. If, in the time series analysis of the monitoring values ​​of all water quality indicators at the monitoring point, the monitoring value at any time is greater than the corresponding first preset threshold or less than the corresponding second preset threshold, the monitoring value of all water quality indicators at that time can be determined as the first candidate abnormal data in the water quality analysis time series of the monitoring point.

[0081] In one example, for any given monitoring point, the first preset condition may include a first preset threshold and a second preset threshold for all water quality indicators at that monitoring point, wherein the first preset threshold is greater than the second preset threshold. For any given monitoring point, in a time series analysis of multiple water quality indicators satisfying a first preset number, if any monitoring value at any given moment is greater than its corresponding first preset threshold or less than its corresponding second preset threshold, then the monitoring values ​​of these water quality indicators at that moment can be identified as the first candidate abnormal data in the water quality analysis time series of that monitoring point. The specific value of the first preset number can be flexibly set according to actual usage requirements; for example, it can be set to 1 or 3, etc., and this disclosure does not impose a specific limitation on it. Furthermore, the specific type of each water quality indicator among the multiple water quality indicators in the first preset number can be further set. For example, for any monitoring point, the first preset number can be set to be 3, including COD, ammonia nitrogen, and conductivity. In the time series analysis of the monitoring values ​​of COD, ammonia nitrogen, and conductivity at the monitoring point, if the monitoring value at any time is greater than the corresponding first preset threshold or less than the corresponding second preset threshold, the monitoring value of COD, ammonia nitrogen, and conductivity at that time can be determined as the first candidate abnormal data in the water quality analysis time series of the monitoring point. This disclosure does not make specific limitations in this regard.

[0082] For any given monitoring point, the rate of change evaluation result for any water quality indicator at that monitoring point can be determined based on the water quality analysis time series, reflecting the change of that water quality indicator at each time point. The specific content of the rate of change evaluation result can be flexibly set according to actual usage needs, and this disclosure does not impose specific limitations on it.

[0083] In one example, the evaluation result of the rate of change of any water quality indicator at any monitoring point can include the comprehensive standard score (Z-score) of the rate of change of the monitored value of the water quality indicator at any time, which can be expressed as formula (2):

[0084]

[0085] in, The comprehensive standard score represents the rate of change of the monitored value of the water quality indicator at any given time. This indicates the change in the absolute monitoring value of the water quality indicator at this moment compared to the previous moment; This represents the average absolute change in the water quality index across all points in the time series of its monitoring values. It represents the standard deviation of the absolute monitoring value change of the water quality indicator at all times in its monitoring value analysis time series.

[0086] For any given monitoring point, based on the evaluation results of the change rates of all water quality indicators at that monitoring point and the second preset conditions, an evolution state anomaly analysis can be performed to identify second candidate anomaly data in the water quality analysis time series of that monitoring point that exhibits anomalies in the change process. The specific content of the second preset conditions can be flexibly set according to actual usage requirements and is related to the actual content of the change rate evaluation results; this disclosure does not impose specific limitations on them.

[0087] In one example, the evaluation result of the rate of change of at least one water quality indicator at any monitoring point may include the comprehensive standard score of the rate of change of the monitored value of that water quality indicator at any given time. The second preset condition may include the third preset threshold corresponding to all water quality indicators at that monitoring point. If there are water quality indicators at that monitoring point whose comprehensive standard scores of the rate of change of the monitored value at any given time are greater than or equal to their respective third preset thresholds, it can be determined that the comprehensive standard scores of the rate of change of the monitored value of these water quality indicators at that monitoring point all meet the second preset condition, and the monitored values ​​of these water quality indicators at that time are identified as second candidate abnormal data. The specific value of the second preset number can be flexibly set according to actual usage needs, and may be equal to the first preset number; this disclosure does not specifically limit this. Furthermore, the specific type of each water quality indicator in the second preset number of water quality indicators can be further set. For example, for any monitoring point, the second preset number can be set to 3, including COD, ammonia nitrogen, and conductivity. If the comprehensive standard score of the rate of change of the monitoring values ​​of COD, ammonia nitrogen, and conductivity at the same time at the monitoring point is greater than or equal to their respective third preset thresholds, the monitoring values ​​of COD, ammonia nitrogen, and conductivity at that time can be determined as the second candidate abnormal data in the water quality analysis time series of the monitoring point. This disclosure does not make specific limitations in this regard.

[0088] In one example, the evaluation result of the rate of change of at least one water quality indicator at any monitoring point may include the comprehensive standard score of the rate of change of the monitored value of the water quality indicator at any given time. The second preset condition may include a third preset threshold corresponding to all water quality indicators at the monitoring point. If the comprehensive standard score of the rate of change of the monitored value of all water quality indicators at the monitoring point at any given time is greater than or equal to their respective third preset thresholds, it can be determined that the comprehensive standard score of the rate of change of the monitored value of all water quality indicators at the monitoring point at that time satisfies the second preset condition, and the monitored values ​​of all water quality indicators at that time are identified as second candidate abnormal data. If the comprehensive standard score of the rate of change of the monitored value of the water quality indicator at any given time is less than the third preset threshold corresponding to the water quality indicator, it can be determined that the comprehensive standard score of the rate of change of the monitored value of the water quality indicator at that time does not satisfy the second preset condition.

[0089] For any given monitoring point, based on the first and second candidate anomaly data for that monitoring point, the monitoring values ​​with numerical anomalies and those with abnormal change processes in the water quality analysis time series of that monitoring point can be identified as water quality anomaly data in the water quality analysis time series of that monitoring point, thereby improving the comprehensiveness and accuracy of water quality anomaly identification.

[0090] Similarly, for any monitoring point, instantaneous state anomaly analysis and evolution state anomaly analysis can be performed based on the flow analysis time series of that monitoring point. That is, statistical analysis is performed from two aspects: whether there are anomalies in the flow value at any time in the flow analysis time series, and whether there are anomalies in the trend of change, to determine the abnormal flow data in the flow analysis time series of that monitoring point.

[0091] In one possible implementation, based on the flow analysis time series of the monitoring point, instantaneous state anomaly analysis and evolution state anomaly analysis are performed to determine the abnormal flow data in the flow analysis time series of the monitoring point. This includes: for any monitoring point, if the monitoring value at any moment in the flow analysis time series of the monitoring point meets a third preset condition, the monitoring value at that moment is determined as the third candidate abnormal data in the flow analysis time series of the monitoring point; based on the flow analysis time series of the monitoring point, the evaluation result of the flow change rate at any moment of the monitoring point is determined; if the evaluation result of the flow change rate at any moment of the monitoring point meets a fourth preset condition, the monitoring value at that moment is determined as the fourth candidate abnormal data in the flow analysis time series of the monitoring point; based on the third and fourth candidate abnormal data of the monitoring point, the abnormal flow data in the flow analysis time series of the monitoring point is determined.

[0092] For any given monitoring point, the monitoring values ​​in the flow analysis time series of that point that meet the third preset condition can be identified as the third candidate abnormal data indicating numerical anomalies in the flow analysis time series of that monitoring point, thereby achieving instantaneous state anomaly analysis. The specific content of the third preset condition can be flexibly set according to actual usage requirements, and this disclosure does not impose specific limitations on it.

[0093] In one example, the third preset condition may include a fourth preset threshold and a fifth preset threshold for traffic, wherein the fourth preset threshold is greater than the fifth preset threshold. For any monitoring point, if any monitored value in the traffic analysis time series of that monitoring point is greater than the fourth preset threshold, or if any monitored value in the traffic analysis time series of that monitoring point is less than the fifth preset threshold, that monitored value can be identified as the third candidate abnormal data in the traffic analysis time series of that monitoring point.

[0094] For any given monitoring point, the rate of change of flow at that monitoring point can be determined based on the flow analysis time series to reflect the change in flow at each moment. The specific content of the rate of change of flow evaluation result can be flexibly set according to actual usage requirements, and this disclosure does not impose specific limitations on it.

[0095] In one example, the evaluation result of the flow rate change at any monitoring point can include the comprehensive standard score of the flow rate change at any monitoring point at any time, which can be expressed as formula (3):

[0096]

[0097] in, A composite standard score representing the rate of change of flow at any given moment; This indicates the absolute change in flow rate between this moment and the previous moment; This represents the average absolute flow change over all moments in the flow analysis time series at that monitoring point. It represents the standard deviation of the absolute flow rate change at all times in the time series analysis of the monitoring values ​​at this monitoring point.

[0098] For any given monitoring point, based on the evaluation results of the flow change rate at that monitoring point and the fourth preset condition, an evolution state anomaly analysis can be performed to identify the fourth candidate anomaly data in the flow analysis time series of that monitoring point that shows an abnormal change process. The specific content of the fourth preset condition can be flexibly set according to actual usage requirements and is related to the actual content of the flow change rate evaluation results; this disclosure does not impose specific limitations on it.

[0099] In one example, the evaluation result of the flow rate change rate at any monitoring point can include the comprehensive standard score of the flow rate change rate at any given time. The fourth preset condition can include the sixth preset threshold corresponding to the flow rate at that monitoring point. If the comprehensive standard score of the flow rate change rate at any given time is greater than or equal to the sixth preset threshold corresponding to the flow rate, it can be determined that the comprehensive standard score of the flow rate change rate at that time satisfies the fourth preset condition, and the monitoring value at that time is identified as the fourth candidate abnormal data in the flow analysis time series of that monitoring point. If the comprehensive standard score of the flow rate change rate at any given time is less than the sixth preset threshold corresponding to the flow rate, it can be determined whether the comprehensive standard score of the flow rate change rate at that time satisfies the fourth preset condition.

[0100] For any given monitoring point, based on the third and fourth candidate anomaly data for that monitoring point, the flow with numerical anomalies and the monitoring values ​​with abnormal change processes in the flow analysis time series of that monitoring point can be identified as flow anomaly data in the flow analysis time series of that monitoring point, thereby improving the comprehensiveness and accuracy of flow anomaly identification.

[0101] For any given monitoring point, based on its water quality and flow anomaly data, all moments when both anomalies occur simultaneously can be determined, thus identifying the target anomaly event at that monitoring point and achieving complete anomaly event identification. Furthermore, by combining statistical analysis with both instantaneous and evolving state dimensions, the accuracy and reliability of anomaly event identification are improved. The specific method for determining the target anomaly event at any given monitoring point based on its water quality and flow anomaly data can be flexibly configured according to actual usage requirements; this disclosure does not impose specific limitations on it.

[0102] In one example, for any given monitoring point, based on the water quality anomaly data and flow anomaly data of that monitoring point, all moments when water quality anomalies and flow anomalies occur simultaneously at that monitoring point can be determined. The monitoring values ​​of at least one water quality indicator and the monitoring values ​​of flow corresponding to these moments are then identified as the target anomaly data for that monitoring point. Based on all the target anomaly data for that monitoring point, the target anomaly event for that monitoring point can be identified, thereby reducing the possibility of missing short-term anomaly events during anomaly event identification and improving the comprehensiveness of anomaly event identification.

[0103] In one possible implementation, the target abnormal event of the monitoring point is determined based on the abnormal water quality data and abnormal flow data of the monitoring point, including: for any monitoring point, determining the target abnormal data of the monitoring point based on the abnormal water quality data and abnormal flow data of the monitoring point; and when there are multiple target abnormal data of the monitoring point that are continuous in time and greater than a preset number, determining the target abnormal event of the monitoring point based on the multiple target abnormal data.

[0104] Specifically, for any given monitoring point, based on the abnormal water quality data and abnormal flow data at that monitoring point, all moments when both water quality and flow abnormalities occur simultaneously at that monitoring point can be determined. Thus, the monitoring values ​​of all water quality indicators and flow rates corresponding to these moments can be identified as the target abnormal data for that monitoring point.

[0105] Furthermore, in addition to the aforementioned method of separately performing water quality anomaly analysis and flow anomaly analysis, and then determining the target anomaly data where water quality and flow anomalies occur simultaneously, one can also first simultaneously perform statistical analysis on the water quality analysis time series and flow analysis time series of any monitoring point in the instantaneous state dimension to determine the instantaneous state anomaly data of the simultaneous occurrence of numerical anomalies in water quality and flow at that monitoring point; then, based on the water quality analysis time series and flow analysis time series of that monitoring point, determine the evaluation results of the change rate of at least one water quality indicator and the evaluation results of the change rate of flow at that monitoring point; based on the evaluation results of the change rate of at least one water quality indicator and the evaluation results of the change rate of flow at that monitoring point, perform weighted processing to determine the evaluation results of the evolution state of that monitoring point, and perform statistical analysis in the evolution state dimension to determine the evolution state anomaly data of the simultaneous occurrence of numerical anomalies in water quality and flow at that monitoring point; finally, based on the instantaneous state anomaly data and the evolution state anomaly data of that monitoring point, determine the target anomaly data of that monitoring point. The specific methods for statistical analysis of the instantaneous state dimension and the evolutionary state dimension can be found in the aforementioned records and will not be elaborated here.

[0106] For any given monitoring point, if there are multiple target anomaly data points that are continuous over time and exceed a preset number, a target anomaly event can be determined to have occurred at that monitoring point based on these multiple target anomaly data points. This allows for the use of both preset quantity and time continuity as a dual filter to filter target anomaly data points with shorter durations, thereby improving the accuracy and reliability of anomaly event identification.

[0107] The specific value of the preset quantity is affected by the time series frequency of the water quality analysis time series and flow analysis time series of each monitoring point, and can be flexibly set according to actual usage needs. This disclosure does not impose specific limitations on this.

[0108] In one example, with the time series frequency of water quality analysis and flow analysis at each monitoring point being 1 hour, the preset quantity can be set to 3. In this case, if any monitoring point has four consecutive target anomaly data points, it can be determined that the monitoring point experienced both water quality and flow anomalies simultaneously for three consecutive hours. Therefore, based on these four target anomaly data points, it is determined that a target anomaly event occurred at that monitoring point.

[0109] In one possible implementation, determining the source monitoring point of the target abnormal event at the monitoring point based on the expected abnormal time window of each monitoring point upstream of the monitoring point includes: determining the number of candidate abnormal events corresponding to the target abnormal event at the monitoring point based on the expected abnormal time window of each monitoring point upstream of the monitoring point; and determining the source monitoring point of the target abnormal event at the monitoring point based on the number of candidate abnormal events corresponding to the target abnormal event at the monitoring point.

[0110] Specifically, for any given monitoring point, based on the expected abnormal time window of each monitoring point upstream of that point, anomaly event investigation can be conducted at each upstream monitoring point. This involves analyzing whether each upstream monitoring point has candidate anomalies that could potentially cause the target anomaly at that monitoring point, and determining the number of candidate anomalies corresponding to the target anomaly at that monitoring point. Here, any candidate anomaly event can represent an anomaly event occurring within the expected abnormal time window of any upstream monitoring point. The criteria for determining candidate anomalies can be similar to those described above; for example, if abnormal water quality data and abnormal flow data occur simultaneously, and the time is consecutive and exceeds a preset number, then a candidate anomaly event is considered to have occurred.

[0111] For any given monitoring point, based on the number of candidate abnormal events corresponding to the target abnormal event at that monitoring point, the source monitoring point that caused the target abnormal event at that monitoring point can be further traced.

[0112] In one possible implementation, the source monitoring point of the target anomaly at a monitoring point is determined based on the number of candidate anomalies corresponding to the target anomaly at that monitoring point. This includes: for any target anomaly at a monitoring point, if the number of candidate anomalies corresponding to the target anomaly meets a fifth preset condition, the monitoring point is determined as the source monitoring point of the target anomaly; if the number of candidate anomalies corresponding to the target anomaly meets a sixth preset condition, a multi-dimensional similarity analysis is performed on the target anomaly and all its corresponding candidate anomalies to determine the source monitoring point of the target anomaly. The multi-dimensional similarity analysis may include: combining the data waveform dimension, data amplitude dimension, and time dimension to perform a similarity analysis on the target anomaly at any monitoring point and each of its corresponding candidate anomalies.

[0113] The fifth preset condition may include that the number of candidate abnormal events corresponding to the target abnormal event at any monitoring point is 0; the sixth preset condition may include that the number of candidate abnormal events corresponding to the target abnormal event at any monitoring point is greater than or equal to 1.

[0114] For any target abnormal event at a monitoring point, if the number of candidate abnormal events corresponding to the target abnormal event meets the fifth preset condition, it can be directly determined that there is no monitoring point upstream of the monitoring point that caused the target abnormal event. Therefore, the monitoring point can be determined as the source monitoring point of the target abnormal event.

[0115] For any given monitoring point, if the number of candidate anomalies corresponding to the target anomaly meets the sixth preset condition, it is possible to determine the upstream monitoring point that may have triggered the target anomaly. Therefore, a multi-dimensional similarity analysis can be performed on the target anomaly and all its corresponding candidate anomalies. By filtering the candidate anomalies, the upstream monitoring points can be indirectly investigated, thereby determining the source monitoring point of the target anomaly. The specific method of multi-dimensional similarity analysis can be flexibly set according to actual usage requirements; this disclosure does not impose specific limitations on it.

[0116] In one possible implementation, if the number of candidate anomalies corresponding to the target anomaly meets a sixth preset condition, a multi-dimensional similarity analysis is performed on the target anomaly and all its corresponding candidate anomalies to determine the source monitoring point of the target anomaly. This includes: for any monitoring point, if the number of candidate anomalies corresponding to the target anomaly meets the sixth preset condition, determining the multi-dimensional similarity analysis results between the target anomaly and each of its corresponding candidate anomalies; determining the confidence evaluation result of each candidate anomaly based on the multi-dimensional similarity analysis results between the target anomaly and each of its corresponding candidate anomalies; if the confidence evaluation results of all candidate anomalies do not meet the confidence threshold, determining the monitoring point as the source monitoring point of the target anomaly; if the confidence evaluation result of at least one candidate anomaly meets the confidence threshold, determining the target candidate anomaly based on all candidate anomalies that meet the confidence threshold, and determining the monitoring point corresponding to the target candidate anomaly as the source monitoring point of the target anomaly.

[0117] For any target anomaly event at a monitoring point, if the number of candidate anomalies corresponding to the target anomaly event meets the sixth preset condition, the multi-dimensional similarity analysis results between the target anomaly event and each of its corresponding candidate anomalies can be determined as the basis for screening candidate anomalies. The specific content of the multi-dimensional similarity analysis results can be flexibly set according to actual usage needs, and this disclosure does not impose specific limitations on it.

[0118] In one possible implementation, the results of multi-dimensional similarity analysis may include: for any target anomaly event at a monitoring point, determining the waveform similarity, amplitude similarity, and temporal similarity between the target anomaly event and any corresponding candidate anomaly event.

[0119] Specifically, for any target anomaly event at a monitoring point, the waveform similarity between the target anomaly event and any corresponding candidate anomaly event can represent the similarity between the changing trend of the time series segment corresponding to the target anomaly event in the time series analysis of the monitoring values ​​of various water quality indicators at the monitoring point and the changing trend of the time series segment corresponding to the candidate anomaly event in the time series analysis of the water quality indicator at the monitoring point corresponding to the candidate anomaly event. Its specific form can be flexibly set according to actual usage requirements. For example, it can be represented by the Pearson correlation coefficient, etc. This disclosure does not make specific limitations on it.

[0120] For any target anomaly event at a monitoring point, the amplitude similarity between the target anomaly event and any corresponding candidate anomaly event can represent the similarity between the change amplitude of the time series segment corresponding to the target anomaly event in the time series analysis of the monitoring values ​​of various water quality indicators at the monitoring point and the change amplitude of the time series segment corresponding to the candidate anomaly event in the time series analysis of the monitoring values ​​of the water quality indicator at the monitoring point corresponding to the candidate anomaly event. Its specific form can be flexibly set according to actual usage needs, and this disclosure does not impose specific limitations on it.

[0121] In one example, for a target anomaly event at any monitoring point, the peak value of the time series segment corresponding to the target anomaly event in the time series analysis of at least one water quality indicator at that monitoring point can be determined as the ratio of the peak value of the time series segment corresponding to the target anomaly event in the time series analysis of the water quality indicator at any monitoring point corresponding to any candidate anomaly event; and the peak value of the time series segment corresponding to the target anomaly event in the flow analysis time series at that monitoring point can be determined as the ratio of the peak value of the time series segment corresponding to the target anomaly event in the flow analysis time series at that monitoring point corresponding to the candidate anomaly event in the time series analysis of the flow analysis time series at that monitoring point corresponding to the candidate anomaly event. Here, the peak value can be either the maximum or minimum value, and can be flexibly set according to actual usage requirements; this disclosure does not impose specific limitations on it.

[0122] Based on this, a weighted sum of all peak-to-peak ratios is performed to determine the amplitude similarity between the target anomaly and the candidate anomaly. The method for determining the peak-to-peak ratio can be flexibly set according to actual usage requirements, and this disclosure does not impose specific limitations on it.

[0123] In one example, the peak signal ratio (PSR) can be expressed as formula (4):

[0124]

[0125] Wherein, PS1 represents the peak value of the time series segment corresponding to the target abnormal event in the time series analysis of the monitoring values ​​of at least one water quality indicator at any monitoring point; PS2 represents the peak value of the time series segment corresponding to the candidate abnormal event in the time series analysis of the monitoring values ​​of the water quality indicator at any monitoring point corresponding to the candidate abnormal event.

[0126] For any target anomaly event at a monitoring point, the temporal similarity between the target anomaly event and any corresponding candidate anomaly event can represent the degree of matching between the actual occurrence time difference between the target anomaly event and the candidate anomaly event and the theoretical propagation time of the monitoring point and the monitoring point corresponding to the candidate anomaly event. Its specific form can be flexibly set according to actual usage requirements, and this disclosure does not impose specific limitations on it. The method for determining the theoretical propagation time can be referred to the preceding description and will not be repeated here.

[0127] In one example, for any target anomaly event at a monitoring point, the temporal similarity between the target anomaly event and any corresponding candidate anomaly event can be represented as the Time Matching Assessment Results (TMAR). The Time Matching Assessment Results can be expressed as formula (5):

[0128]

[0129] TMAR represents the time-matching evaluation result between a target anomaly event at any monitoring point and any corresponding candidate anomaly event; Time a Indicates the actual time difference between the target anomaly and the candidate anomaly; Time t This indicates the theoretical propagation time between the monitoring point and the monitoring point corresponding to the candidate abnormal event.

[0130] For any target anomaly event at a monitoring point, the confidence level evaluation result for each candidate anomaly event can be determined based on the multi-dimensional similarity analysis results between the target anomaly event and each corresponding candidate anomaly event. The specific format of the confidence level evaluation result can be flexibly set according to actual usage requirements, and this disclosure does not impose specific limitations on it.

[0131] In one example, for any target anomaly event at a monitoring point, the waveform similarity, amplitude similarity, and time similarity between the target anomaly event and any corresponding candidate anomaly event can be weighted and summed. The summation result is then used as the confidence evaluation result of the candidate anomaly event.

[0132] If the confidence evaluation results of all candidate anomalies fail to meet the confidence threshold, it indicates that there is no correlation between these candidate anomalies and the target anomaly. Therefore, it can be determined that there is no monitoring point upstream of the monitoring point that could have triggered the target anomaly; in other words, the monitoring point itself is the source monitoring point of the target anomaly. The specific value of the confidence threshold can be flexibly set according to actual usage requirements, and this disclosure does not impose specific limitations on it.

[0133] If the confidence evaluation result of at least one candidate anomaly meets the confidence threshold, it indicates that there is a relatively obvious correlation between at least one candidate anomaly and the target anomaly, and it can be determined that there may be a monitoring point upstream of the monitoring point that triggered the target anomaly. Therefore, based on all candidate anomalies that meet the confidence threshold, the target candidate anomaly can be determined, and the monitoring point corresponding to the target candidate anomaly can be determined as the source monitoring point of the target anomaly. The specific method for determining the target candidate anomaly based on all candidate anomalies that meet the confidence threshold can be flexibly set according to actual usage requirements, and this disclosure does not impose specific limitations on it.

[0134] In one example, for any target anomaly event at a monitoring point, if the confidence evaluation result of at least one candidate anomaly event of the target anomaly event meets the confidence threshold, the candidate anomaly event with the largest confidence evaluation result among all candidate anomalies that can meet the confidence threshold is determined and identified as the target candidate anomaly event.

[0135] Through the above process, it is possible to achieve comprehensive analysis of anomaly source tracing in both time and space dimensions by using the number of candidate anomalies corresponding to the target anomaly at any monitoring point as a basis. This fully leverages the advantages of online monitoring of drainage pipe networks in terms of time frequency and spatial density, thereby improving the efficiency of tracing the source of the target anomaly at that monitoring point. Furthermore, by utilizing multi-dimensional similarity analysis, the accuracy and reliability of anomaly source tracing can be improved.

[0136] In one possible implementation, the method further includes: determining the rainfall event analysis results of the target drainage network based on the initial rainfall time series of the target drainage network, wherein the rainfall event analysis results are used to indicate the rainfall time and rainfall type of the target drainage network; and for any target abnormal event at any monitoring point, performing a causal analysis on the target abnormal event based on the rainfall event analysis results, and determining the causal analysis results of the target abnormal event.

[0137] The initial rainfall time series corresponding to the target drainage network can reflect the change in rainfall over time within the geographical area corresponding to the target drainage network. Specific methods for determining the initial rainfall time series corresponding to the target drainage network can refer to implementation methods in related technologies, such as real-time monitoring using rain gauges, etc., and this disclosure does not specifically limit this method.

[0138] Based on the initial rainfall time series of the target drainage network, the rainfall event analysis results of the target drainage network can be determined to indicate the rainfall time and rainfall type of the target drainage network. The rainfall time can include the start and end times of rainfall for each rainfall period of the target drainage network, etc., and its content can be flexibly set according to actual usage needs. This disclosure does not make specific limitations on this. The rainfall type can represent the rainfall rating based on preset standards for each rainfall period. For example, it can include light rain, moderate rain, heavy rain, rainstorm, heavy rainstorm, and extremely heavy rainstorm, etc., and its content can be flexibly set according to actual usage needs. This disclosure does not make specific limitations on this.

[0139] For any target anomaly event at any monitoring point, a causal analysis can be performed based on the rainfall event analysis results. The causal analysis results indicate whether the target anomaly event at the monitoring point is caused by rainfall within the geographical area corresponding to the target drainage network. For example, if heavy rain occurs within the geographical area corresponding to the target drainage network, surface pollutants may be washed into the target drainage network, leading to an anomaly event at at least one monitoring point. The specific method for performing causal analysis on a target anomaly event at any monitoring point based on the rainfall event analysis results can be flexibly set according to actual usage requirements; this disclosure does not impose specific limitations on it.

[0140] In one example, for any target anomaly event at a monitoring point, the analysis can be performed directly based on the time range of the target anomaly event and the rainfall time of the target drainage network indicated by the rainfall event analysis results. This analysis determines whether the target anomaly event and the rainfall have a delayed synchronicity. If the target anomaly event and the rainfall have a delayed synchronicity, the cause analysis result of the target anomaly event is determined to be affected by rainfall; if the target anomaly event and the rainfall do not have a delayed synchronicity, the cause analysis result of the target anomaly event is determined to be unaffected by rainfall.

[0141] The specific method for analyzing whether the target abnormal event and rainfall have delayed synchronization can refer to the implementation methods in related technologies. For example, a first time difference between the start time of rainfall and the start time of the target abnormal event, and a second time difference between the end time of rainfall and the end time of the target abnormal event can be determined. If the error value between the first time difference and the second time difference is less than or equal to a preset error threshold, it is determined that the target abnormal event and rainfall have delayed synchronization. If the error value between the first time difference and the second time difference is greater than the preset error threshold, it is determined that the target abnormal event and rainfall do not have delayed synchronization. This disclosure does not make specific limitations in this regard.

[0142] In one possible implementation, for any target anomaly event at a monitoring point, a causal analysis is performed on the target anomaly event based on the rainfall event analysis results to determine the causal analysis results. This includes: for any target anomaly event at a monitoring point, determining the rainfall correlation analysis results of the target anomaly event based on the water quality analysis time series and flow analysis time series of the monitoring point and the rainfall event analysis results, or based on the water quality analysis time series and flow analysis time series of the source monitoring point of the target anomaly event and the rainfall event analysis results. The rainfall correlation analysis results indicate the degree of correlation between the target anomaly event and rainfall. Based on the rainfall correlation analysis results of the target anomaly event, the causal analysis results of the target anomaly event are determined.

[0143] For any target anomaly event at any monitoring point, a correlation analysis can be performed between the water quality analysis time series and flow analysis time series at that monitoring point and the rainfall time series analysis results. This analysis examines the correlation between the target anomaly event and rainfall from the perspective of whether rainfall directly affects the monitoring point, thus determining the rainfall correlation analysis result for the target anomaly event. The specific methods of the correlation analysis and the specific format of the rainfall correlation analysis results can be flexibly set according to actual usage requirements. For example, a correlation analysis can be performed between the water quality analysis time series and flow analysis time series at any monitoring point and the corresponding rainfall time series, and the calculated correlation coefficient can be determined as the rainfall correlation analysis result. This disclosure does not impose specific limitations on this.

[0144] On the other hand, for any target anomaly event at any monitoring point, a correlation analysis can be performed between the water quality analysis time series and flow analysis time series of the source monitoring point of the target anomaly event and the rainfall time analysis results. This allows for analysis of whether rainfall affected the source monitoring point of the target anomaly event, the magnitude of the correlation between the target anomaly event and rainfall, and the determination of the rainfall correlation analysis results for the target anomaly event.

[0145] For any target anomaly event at any monitoring point, if the rainfall correlation analysis result of the target anomaly event meets the seventh preset condition, the causal analysis result of the target anomaly event can be determined to be affected by rainfall; if the rainfall correlation analysis result of the target anomaly event does not meet the seventh preset condition, the causal analysis result of the target anomaly event can be determined to be unaffected by rainfall. The specific content of the seventh preset condition can be flexibly set according to actual usage needs and depends on the actual form of the rainfall correlation analysis result; this disclosure does not impose specific limitations on it.

[0146] In one example, given the correlation coefficient between the rainfall correlation analysis results of a target anomalous event at any monitoring point, including the water quality analysis time series and flow analysis time series at that monitoring point, and the rainfall time series corresponding to the rainfall time analysis results, the seventh preset condition may include a correlation coefficient threshold. If the correlation coefficient between the water quality analysis time series and flow analysis time series at that monitoring point and the rainfall time is greater than or equal to the correlation coefficient threshold, it can be determined that the rainfall correlation analysis results of the target anomalous event meet the seventh preset condition; if the correlation coefficient between the water quality analysis time series and flow analysis time series at that monitoring point and the rainfall time is less than the correlation coefficient threshold, it can be determined that the rainfall correlation analysis results of the target anomalous event do not meet the seventh preset condition.

[0147] In this embodiment of the disclosure, for any monitoring point in the target drainage network, the water quality analysis time series and flow analysis time series of the monitoring point can be determined based on the initial water quality monitoring time series and the initial flow time series, which include the initial monitoring values ​​of at least one water quality indicator. Based on the water quality analysis time series and flow analysis time series of the monitoring point, multi-dimensional statistical analysis can be performed on the monitoring values ​​of at least one water quality indicator in the water quality analysis time series and the monitoring values ​​of the flow in the flow analysis time series, combining instantaneous state dimensions and evolutionary state dimensions. This fully utilizes the data characteristics of the water quality analysis time series and flow analysis time series of the monitoring point to comprehensively capture water quality anomalies and flow anomalies at the monitoring point, thereby identifying target abnormal events where both water quality and flow anomalies occur simultaneously at the monitoring point, and improving the accuracy and reliability of abnormal event identification. Based on the time range of the target anomaly at the monitoring point and the topological relationship of the monitoring points in the target drainage network, the expected anomaly time window for each monitoring point upstream of the current monitoring point can be determined. This narrows the time range required for upstream source tracing of the target anomaly at any given monitoring point, enabling comprehensive analysis of anomaly source tracing in both time and spatial dimensions. It fully leverages the advantages of the water quality and flow analysis time series at each monitoring point in terms of time frequency, as well as the spatial density of monitoring points in the target drainage network, thereby improving the efficiency of anomaly source tracing. Based on the expected anomaly time window of each monitoring point upstream of the current monitoring point, the source monitoring point of the target anomaly at that monitoring point can be determined. On this basis, with the source monitoring point of the target anomaly at the current monitoring point as the center, targeted and rapid pollution source tracing can be carried out within the monitoring range of the current monitoring point, further improving the efficiency and accuracy of anomaly source tracing. Compared with the common analysis methods for drainage pipe network monitoring data in the prior art, the embodiments of this disclosure can give full play to the advantages of online monitoring of drainage pipe networks in terms of time frequency and spatial density. It is not only suitable for identifying and tracing abnormal events in the long-term operation of the target drainage pipe network, but also can provide a reliable basis for real-time pipeline health diagnosis of the target drainage pipe network.

[0148] It is understood that the various method embodiments mentioned above in this disclosure can be combined with each other to form combined embodiments without violating the principle and logic. Due to space limitations, this disclosure will not elaborate further. Those skilled in the art will understand that in the above methods of specific implementation, the specific execution order of each step should be determined by its function and possible internal logic.

[0149] This disclosure also provides an electronic device, including a memory, a processor, and a computer program stored in the memory, wherein the processor executes the computer program to implement the steps of the above method.

[0150] This disclosure also provides a non-volatile computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the steps of the above-described method.

[0151] Figure 3 A block diagram of an electronic device according to an embodiment of the present disclosure is shown. For example, electronic device 1900 may be provided as a server or terminal device. (Refer to...) Figure 3 The electronic device 1900 includes a processing component 1922, which further includes one or more processors, and memory resources represented by memory 1932 for storing instructions, such as application programs, that can be executed by the processing component 1922. The application programs stored in memory 1932 may include one or more modules, each corresponding to a set of instructions. Furthermore, the processing component 1922 is configured to execute instructions to perform the methods described above.

[0152] Electronic device 1900 may also include a power supply component 1926 configured to perform power management of electronic device 1900, a wired or wireless network interface 1950 configured to connect electronic device 1900 to a network, and an input / output interface 1958 (I / O interface). Electronic device 1900 can operate on an operating system, such as Windows Server, stored in memory 1932. TM Mac OS X TM Unix TM Linux TM FreeBSD TM Or similar.

[0153] The various embodiments of this disclosure have been described above. These descriptions are exemplary and not exhaustive, nor are they limited to the disclosed embodiments. Many modifications and variations will be apparent to those skilled in the art without departing from the scope and spirit of the described embodiments. The terminology used herein is chosen to best explain the principles, practical application, or technical improvements to the embodiments in the market, or to enable others skilled in the art to understand the embodiments disclosed herein.

Claims

1. A method for tracing the source of anomalies in a drainage pipe network, characterized in that, include: For any monitoring point in the target drainage network, the water quality analysis time series and flow analysis time series of the monitoring point are determined based on the initial water quality monitoring time series and the initial flow time series of the monitoring point. The initial water quality monitoring time series of the monitoring point includes the initial monitoring value time series of at least one water quality indicator of the monitoring point. Based on the water quality analysis time series and flow analysis time series of the monitoring point, multi-dimensional statistical analysis is performed to determine the target abnormal event of the monitoring point. The multi-dimensional statistical analysis includes: combining the instantaneous state dimension and the evolution state dimension to perform statistical analysis on the monitoring values ​​of at least one water quality indicator in the water quality analysis time series and the monitoring values ​​of flow in the flow analysis time series. The target abnormal event of any monitoring point indicates that the monitoring point is simultaneously experiencing water quality abnormality and flow abnormality. Based on the time range of the target abnormal event at the monitoring point and the topological relationship of the monitoring points in the target drainage network, determine the expected abnormal time window of each monitoring point upstream of the monitoring point. Based on the expected abnormal time window of each monitoring point upstream of the current monitoring point, determine the source monitoring point of the target abnormal event for the current monitoring point.

2. The method according to claim 1, characterized in that, For any monitoring point in the target drainage network, the water quality analysis time series and flow analysis time series for that monitoring point are determined based on the initial water quality monitoring time series and initial flow time series, including: For any given monitoring point, data preprocessing is performed based on the initial water quality monitoring time series and initial flow time series of that monitoring point to determine the processed water quality monitoring time series and processed flow time series of that monitoring point. Based on the treated water quality monitoring time series of the monitoring point, time series decomposition and periodic term removal are performed to determine the water quality analysis time series of the monitoring point. Based on the processed flow monitoring time series of the monitoring point, time series decomposition and periodic term removal are performed to determine the flow analysis time series of the monitoring point.

3. The method according to claim 1 or 2, characterized in that, The process involves performing multi-dimensional statistical analysis based on the water quality analysis time series and flow analysis time series at the monitoring point to determine the target abnormal events at that monitoring point, including: For any given monitoring point, based on the water quality analysis time series of that monitoring point, instantaneous state anomaly analysis and evolution state anomaly analysis are performed to determine the water quality anomaly data in the water quality analysis time series of that monitoring point. Based on the flow analysis time series of the monitoring point, instantaneous state anomaly analysis and evolution state anomaly analysis are performed to determine the abnormal flow data in the flow analysis time series of the monitoring point. Based on the abnormal water quality and flow data at the monitoring point, the target abnormal event at the monitoring point is determined.

4. The method according to claim 3, characterized in that, For any given monitoring point, based on the water quality analysis time series of that monitoring point, instantaneous state anomaly analysis and evolutionary state anomaly analysis are performed to determine the water quality anomaly data in the water quality analysis time series of that monitoring point, including: For any monitoring point, if there are multiple water quality indicators in the water quality analysis time series of the monitoring point that meet the first preset number of monitoring values, and the monitoring value at any time in the analysis time series of the multiple water quality indicators meets the first preset condition, then the monitoring value of the multiple water quality indicators at that time is determined as the first candidate abnormal data in the water quality analysis time series of the monitoring point. Based on the water quality analysis time series at the monitoring point, determine the evaluation result of the rate of change of each water quality indicator at any time at the monitoring point; If, at any given moment, the rate of change of multiple water quality indicators at the monitoring point meets the second preset condition, the monitoring values ​​of the multiple water quality indicators at that moment are determined as the second candidate abnormal data in the water quality analysis time series of the monitoring point. Based on the first and second candidate anomaly data of the monitoring point, the water quality anomaly data in the water quality analysis time series of the monitoring point are determined.

5. The method according to claim 3, characterized in that, The step of performing instantaneous state anomaly analysis and evolution state anomaly analysis based on the flow analysis time series of the monitoring point to determine the abnormal flow data in the flow analysis time series of the monitoring point includes: For any monitoring point, if the monitoring value at any time in the flow analysis time series of that monitoring point meets the third preset condition, the monitoring value at that time is determined as the third candidate abnormal data in the flow analysis time series of that monitoring point. Based on the flow analysis time series of the monitoring point, determine the evaluation result of the flow change rate at any time of the monitoring point; If the evaluation result of the rate of change of the flow at any time at the monitoring point meets the fourth preset condition, the monitoring value at that time is determined as the fourth candidate abnormal data in the flow analysis time series of the monitoring point. Based on the third and fourth candidate anomaly data of the monitoring point, the abnormal flow data in the flow analysis time series of the monitoring point are determined.

6. The method according to claim 3, characterized in that, The process of determining the target abnormal event at the monitoring point based on the abnormal water quality and abnormal flow data includes: For any given monitoring point, the target abnormal data for that monitoring point is determined based on the abnormal water quality data and abnormal flow data at that monitoring point. If there are multiple target anomaly data points at a monitoring point that are continuous over time and exceed a preset number, the target anomaly event at the monitoring point is determined based on the multiple target anomaly data points.

7. The method according to claim 1 or 2, characterized in that, The step of determining the source monitoring point of the target abnormal event at the monitoring point based on the expected abnormal time window of each monitoring point upstream of the monitoring point includes: Based on the expected abnormal time window of each monitoring point upstream of the monitoring point, determine the number of candidate abnormal events corresponding to the target abnormal event of the monitoring point. Based on the number of candidate abnormal events corresponding to the target abnormal event at the monitoring point, the source monitoring point of the target abnormal event at the monitoring point is determined.

8. The method according to claim 7, characterized in that, The step of determining the source monitoring point of the target abnormal event based on the number of candidate abnormal events corresponding to the target abnormal event at the monitoring point includes: For any target abnormal event at a monitoring point, if the number of candidate abnormal events corresponding to the target abnormal event meets the fifth preset condition, the monitoring point is determined as the source monitoring point of the target abnormal event. If the number of candidate anomalies corresponding to the target anomaly meets the sixth preset condition, a multi-dimensional similarity analysis is performed on the target anomaly and all its corresponding candidate anomalies to determine the source monitoring point of the target anomaly. The multi-dimensional similarity analysis includes: combining the data waveform dimension, data amplitude dimension and time dimension to perform a similarity analysis on the target anomaly at any monitoring point and each of its corresponding candidate anomalies.

9. The method according to claim 8, characterized in that, When the number of candidate anomalies corresponding to the target anomaly meets the sixth preset condition, a multi-dimensional similarity analysis is performed on the target anomaly and all its corresponding candidate anomalies to determine the source monitoring point of the target anomaly, including: For any target anomaly event at a monitoring point, if the number of candidate anomalies corresponding to the target anomaly event meets the sixth preset condition, the multi-dimensional similarity analysis results between the target anomaly event and each of its corresponding candidate anomalies are determined respectively. Based on the multi-dimensional similarity analysis results between the target anomaly and each of its corresponding candidate anomalies, the confidence evaluation result of each candidate anomaly is determined. If the confidence evaluation results of all candidate anomalies do not meet the confidence threshold, the monitoring point is determined as the source monitoring point of the target anomaly. If the confidence evaluation result of at least one candidate abnormal event meets the confidence threshold, a target candidate abnormal event is determined based on all candidate abnormal events that meet the confidence threshold, and the monitoring point corresponding to the target candidate abnormal event is determined as the source monitoring point of the target abnormal event.

10. The method according to claim 1 or 2, characterized in that, The method further includes: Based on the initial rainfall time series of the target drainage network, the rainfall event analysis results of the target drainage network are determined, wherein the rainfall event analysis results are used to indicate the rainfall time and rainfall type of the target drainage network; For any target anomaly event at any monitoring point, based on the rainfall event analysis results, a causal analysis is performed on the target anomaly event to determine the causal analysis results.

11. The method according to claim 10, characterized in that, For any target anomaly event at any monitoring point, based on the rainfall event analysis results, a causal analysis is performed on the target anomaly event to determine the causal analysis results, including: For any target abnormal event at a monitoring point, the rainfall correlation analysis result of the target abnormal event is determined based on the water quality analysis time series and flow analysis time series of the monitoring point and the rainfall event analysis result, or based on the water quality analysis time series and flow analysis time series of the source monitoring point of the target abnormal event and the rainfall event analysis result. The rainfall correlation analysis result indicates the degree of correlation between the target abnormal event and rainfall. Based on the rainfall correlation analysis results of the target anomaly event, the causal analysis results of the target anomaly event are determined.

12. An electronic device comprising a memory, a processor, and a computer program stored in the memory, characterized in that, The processor executes the computer program to implement the steps of the method according to any one of claims 1 to 11.

Images