Low-cost rapid pollution source tracing method based on pipe network topological relation analysis

Abstract

The invention discloses a low-cost rapid pollution source tracing method based on pipe network topological relation analysis, and relates to the technical field of environmental engineering, pipe network topological modeling, a temperature-corrected conductivity-water quality correlation model and an improved correlation algorithm are fused, data are collected through a low-cost detection terminal, pollution abrupt change nodes are rapidly screened out, and the pollution source tracing efficiency is improved. The method has the advantages that a directed graph model is constructed based on a pipe network topological relation to define node association and water flow logic, a conductivity-water quality index association model is established in combination with synchronous detection of a total tail end of the pipe network, and a low-cost detection terminal is arranged at a key node to realize real-time data acquisition; and through data preprocessing and conductivity amplitude variation, correlation analysis is improved to quickly screen abrupt change nodes, so that the problems of high cost, slow response, low efficiency and difficulty in popularization of the traditional technology are avoided, and the method does not need to depend on a high-price traceability system and laboratory complex detection.

Description

Technical Field

[0001] This invention relates to the field of environmental engineering technology, specifically to a low-cost and rapid method for tracing pollution sources based on pipeline network topology analysis. Background Technology

[0002] As a core component of urban infrastructure and industrial production systems, pipeline systems encompass various types, including urban drainage networks and dedicated industrial park networks. They undertake crucial functions such as water resource transportation, sewage collection, and discharge. Their operational status directly impacts water resource security, ecological environment quality, and public health. With the acceleration of urbanization and the expansion of industrial production, the coverage of pipeline networks continues to expand, and their structures are becoming increasingly complex. The emergence of various sudden or persistent pollution sources can not only disrupt the normal operation of pipeline networks but also trigger a chain of problems such as downstream water pollution and ecological damage. Therefore, tracing pollution sources has become a key aspect of pipeline network operation and maintenance management and environmental governance.

[0003] Existing pipeline pollution source tracing technologies mainly rely on the construction of expensive tracing systems, offline laboratory testing, or manual on-site investigation to locate pollution sources. These technologies have certain drawbacks. First, they depend on expensive equipment and complex testing processes, requiring a large amount of manpower, resulting in high costs, slow response, low efficiency, and difficulty in widespread adoption. Second, they do not fully integrate pipeline topology relationships, and the testing parameters are fixed and lack targeted judgment rules, leading to large tracing deviations and difficulty in adapting to pipeline networks of different sizes. Therefore, we propose a low-cost and rapid pollution source tracing method based on pipeline topology relationship analysis. Summary of the Invention

[0004] The purpose of this invention is to provide a low-cost and rapid method for tracing pollution sources based on pipeline network topology analysis.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a low-cost and rapid pollution source tracing method based on pipeline network topology analysis, comprising the following steps:

[0006] Step 1: Collect and verify basic data on the pipeline network in the target area, and construct a directed graph topology model. An adjacency matrix is ​​used to store and form a callable topology database;

[0007] Step 2: Simultaneously and continuously monitor conductivity, COD, and ammonia nitrogen concentrations at the end of the pipeline network, construct and calibrate a conductivity-water quality index correlation model incorporating a temperature correction coefficient;

[0008] Step 3: Based on the topology model, select key detection nodes in the pipeline network, deploy low-cost conductivity detection terminals and configure detection parameters to achieve real-time data transmission;

[0009] Step 4: Perform integrity verification, interpolation, and noise reduction on the received detection data to construct a standardized time-series dataset;

[0010] Step 5: Based on the upstream and downstream relationships in the topology model, calculate the hydraulic propagation time, and screen abrupt change nodes and associated nodes through relative conductivity variation and correlation analysis;

[0011] Step 6: Verify the mutation nodes and associated nodes according to the preset judgment criteria, locate the pollution source nodes and generate a location report;

[0012] Step 7: Conduct on-site sampling and verification of pollution source nodes, calibrate the model and correction report based on the verification results, and form a closed loop for source tracing.

[0013] As a further aspect of the present invention: In step one, the basic pipeline data includes pipeline as-built drawings, pipe diameter, pipe slope, node location coordinates, pipe segment length, design flow direction and actual operation flow direction records, and the location and specifications of pipeline ancillary facilities. The collected data is validated, and damaged or outdated invalid data is removed. In the directed graph topology model, the node set... Each pipeline node contains five parameters: a unique identifier (ID), pipe diameter (d), slope (s), location coordinates, and node type (branch node, main node, or confluence node). The edge set... Define three attributes for each pipe segment: water flow direction (from upstream node to downstream node), pipe segment length L, and pipe segment material. Use an adjacency matrix to store the directed graph data, forming a topology database that can be called in real time and whose parameters can be updated.

[0014] As a further aspect of the present invention: In step two, a comprehensive detection device is installed at the outlet of the main end of the pipeline network. The detection frequency is uniformly set to 10-15 minutes per detection, continuously collecting basic data for no less than 72 hours to ensure that the data covers different periods of normal operation of the pipeline network. The data collection adopts a 3... Outliers were removed and the data were standardized. COD and ammonia nitrogen concentration data were converted into time-series datasets that perfectly matched the conductivity detection time. Based on the processed time-series data, a linear correlation model between conductivity and water quality indicators was constructed, incorporating a temperature correction coefficient. The model formula is as follows:

[0015] ;

[0016] ;

[0017] in, This is the calculated COD concentration. This is the calculated value of ammonia nitrogen concentration. , These are the calibration coefficients of the COD association model (obtained through linear regression fitting, goodness of fit). ≥0.9), , The calibration coefficients (goodness of fit) of the ammonia-nitrogen correlation model are respectively. ≥0.88), This is the actual measured conductivity value. This is the conductivity temperature correction factor (value range 0.02-0.025 / ℃, determined according to the temperature range of the water in the pipe network). The actual water temperature at the time of testing. The baseline temperature is 25℃, and the results are verified through 30-50 sets of on-site sampling.

[0018] As a further aspect of the present invention: In step three, the criteria for screening key detection nodes in the pipeline network include: nodes 30-50m before the branch pipe merges into the main pipe, adjacent nodes before and after the main pipe connects to the branch pipe (with a spacing of 10-20m, used to compare the impact of the branch pipe merging into the main pipe on the water quality), nodes where 3 or more pipelines converge, and pipeline bends and diameter change nodes with a pipe diameter change range of ≥50mm.

[0019] On-site inspections were conducted on the selected key nodes to confirm that the nodes met the conditions for installing detection terminals. Nodes that did not meet the installation conditions were removed and replaced with the nearest qualified upstream and downstream nodes to ensure that the coverage rate of key nodes was ≥90%.

[0020] Low-cost conductivity detection terminals are installed at key nodes. The accuracy of the terminals is set to ±1μS / cm. The detection frequency is configured according to the scale of the pipeline network: 5-15 min / time for small and medium-sized pipelines with a total length ≤10km, and 20-60 min / time for large-scale pipelines with a total length >10km. The detection terminals transmit real-time detection data (including detection timestamp, node ID, conductivity value, and detection terminal working status) to the data processing center through a wireless communication module.

[0021] As a further aspect of the present invention: In step four, after receiving the data, the data processing center first performs data integrity verification. Data with missing timestamps, ambiguous node IDs, or conductivity values ​​exceeding reasonable ranges (0-1500 μS / cm for domestic sewage networks, 0-5000 μS / cm for industrial wastewater networks) are marked as invalid data. Missing valid data is supplemented using interpolation. The verified conductivity data undergoes noise reduction processing, and a moving average method is used to smooth data fluctuations. The moving window size is... Determined based on the detection frequency, when the detection frequency is 5-15 minutes / time. =5, 20-60 min / time =3, the calculation formula is as follows:

[0022] ;

[0023] in, For the first Smoothed conductivity value for each detection cycle For the first The raw conductivity values ​​of each detection cycle are used to sort the smoothed conductivity data of each key node by timestamp and associate them one by one with the node IDs in the topology model to construct a time-series dataset containing node ID, detection time, smoothed conductivity value, and water temperature.

[0024] As a further aspect of the present invention: in step five, the upstream and downstream relationships of each key node and the hydraulic parameters of the pipe segment are extracted from the topology database, and the hydraulic propagation time between the upstream and downstream nodes is calculated. The calculation formula is:

[0025] ;

[0026] in: The water flow velocity within the pipe section is determined by the pipe diameter. Calculated using Manning's formula For pipe segment length, based on time series dataset and Calculate the correlation between the conductivity variation of each node and the upstream and downstream nodes:

[0027] Conductivity variation calculation: Nodes are calculated using a sliding time window. In the The relative change in conductivity at time t is given by the formula:

[0028] ;

[0029] in, For nodes exist The relative amplitude at any given moment, The duration of the sliding window. For nodes exist The smoothed conductivity value at time t. ) is a node exist The smoothed conductivity value at any given time;

[0030] Upstream and downstream node correlation calculation: The correlation degree is used to calculate the upstream node. With downstream nodes The conductivity correlation is used to introduce the hydraulic propagation time correction time difference, and the formula is:

[0031] ;

[0032] in, For nodes and The degree of correlation, The number of valid data sets, For nodes The mean smooth conductivity, For nodes The average smooth conductivity is used to set the amplitude threshold according to the pipeline network type. (Domestic sewage pipe network) =15%, industrial wastewater pipeline network =25%) and correlation threshold Filter out mutation nodes and their correlation with mutation nodes The associated nodes form a node change relationship matrix.

[0033] As a further aspect of the present invention: In step six, based on the obtained node change relationship matrix and topological model, a criterion for determining pollution source nodes is set:

[0034] (1) The node is a mutation node, and the mutation time of all its downstream related nodes is later than that of the node;

[0035] (2) There are no mutation nodes among the upstream associated nodes of this node that meet the above conditions;

[0036] (3) The pollution concentration calculated by the correlation model exceeds the preset range of the corresponding pipeline network:

[0037] The location rules for different node types are as follows: the branch pipe corresponding to the node before the branch pipe merges is the branch pipe into which the pollution source merges; the main pipe merging node is the core pollution source node if it meets the conditions; and the turning and diameter change nodes are pollution source nodes within the pipe section if they meet the conditions.

[0038] The location report must clearly state the node ID, specific location, time of pollution occurrence, estimated pollution concentration, and scope of impact.

[0039] As a further aspect of the present invention: In step seven, 2-3 sets of water samples are collected on-site for verification. The detection method conforms to the standards GB / T11914-1989 and HJ535-2009. If the relative error is ≤15%, the verification is passed. If it is greater than 15%, the correlation model is recalibrated and steps five and six are repeated. After verification, the source of pollution is confirmed by manual investigation, and supplementary treatment measures are taken to form a closed loop of location-verification-source tracing.

[0040] Compared with the prior art, the beneficial effects of the present invention by adopting the above technical solution are as follows:

[0041] 1. This invention constructs a directed graph model based on the pipeline network topology to clarify node associations and water flow logic. It combines synchronous detection at the end of the pipeline network to establish a conductivity-water quality index correlation model. Low-cost detection terminals are deployed at key nodes to achieve real-time data acquisition. After data preprocessing and conductivity amplitude and improved correlation analysis, nodes with abrupt changes are quickly screened. With the help of on-site verification and closed-loop calibration, it avoids the problems of high cost, slow response, low efficiency and difficulty in popularization of traditional technologies. It does not require reliance on expensive traceability systems and complex laboratory testing, nor does it require a large amount of manual investigation. Data acquisition and analysis can be completed without professional operation.

[0042] 2. This invention clarifies the hydraulic correlation and water flow propagation logic of each node based on the pipeline network topology model. It combines improved correlation analysis to correct the time difference between upstream and downstream nodes, sets targeted judgment rules according to node type, and verifies and calibrates the model and location report through on-site sampling. At the same time, it flexibly configures the detection frequency according to the scale of the pipeline network. This avoids the problems of large source tracing deviation and difficulty in adapting to different scale pipeline networks caused by traditional technologies that ignore the topological relationship of nodes and have fixed detection parameters. It does not require a unified detection scheme and experience-based investigation, nor does it require a large amount of data supplementation later. It can achieve accurate correlation and node location of pipeline network pollution data in different scenarios, and ultimately improve the credibility and scenario adaptability of the source tracing results. Attached Figure Description

[0043] Figure 1 This is a schematic diagram of the method steps in an embodiment of the present invention. Detailed Implementation

[0044] The specific embodiments of the present invention will be further described below with reference to the accompanying drawings. It should be noted that the description of these embodiments is for the purpose of helping to understand the present invention, but does not constitute a limitation of the present invention.

[0045] Furthermore, the technical features involved in the various embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0046] Please see the appendix Figure 1 This invention discloses a low-cost and rapid pollution source tracing method based on pipeline network topology analysis, comprising the following steps:

[0047] Step 1: Collect and verify basic data on the pipeline network in the target area, and construct a directed graph topology model. An adjacency matrix is ​​used to store and form a callable topology database;

[0048] Step 2: Simultaneously and continuously monitor conductivity, COD, and ammonia nitrogen concentrations at the end of the pipeline network, construct and calibrate a conductivity-water quality index correlation model incorporating a temperature correction coefficient;

[0049] Step 3: Based on the topology model, select key detection nodes in the pipeline network, deploy low-cost conductivity detection terminals and configure detection parameters to achieve real-time data transmission;

[0050] Step 4: Perform integrity verification, interpolation, and noise reduction on the received detection data to construct a standardized time-series dataset;

[0051] Step 5: Based on the upstream and downstream relationships in the topology model, calculate the hydraulic propagation time, and screen abrupt change nodes and associated nodes through relative conductivity variation and correlation analysis;

[0052] Step 6: Verify the mutation nodes and associated nodes according to the preset judgment criteria, locate the pollution source nodes and generate a location report;

[0053] Step 7: Conduct on-site sampling and verification of pollution source nodes, calibrate the model and correction report based on the verification results, and form a closed loop for source tracing.

[0054] Example 1, please refer to the appendix. Figure 1 "Directed graph topology model" The "Construction" process involves importing as-built drawings of a town's pipeline network into AutoCAD Map3D software. The "Pipeline Network Data Extraction Tool" automatically extracts pipe diameters, pipe segment lengths, and manhole coordinates. The flow direction is manually marked (based on the pipeline slope: 0.5‰-3‰, with downward slope indicating flow direction). Node sets are then defined within the software. Each inspection well is a node, assigned an ID, and associated with pipe diameter d, slope s, coordinates (x,y) and node type;

[0055] edge set The corresponding edges of the pipe segments are stored, including the flow direction (e.g., J5→J6), the pipe segment length L, and the material (concrete pipe is labeled "C30"). Finally, the result is exported as an adjacency matrix (25×25), with the matrix elements being the pipe segment length (0 for no connection), forming a topology database.

[0056] Example 2, please refer to the appendix. Figure 1 The "correlation model calibration" involved installing a multi-parameter detector at the end of the pipeline network (the inlet of the downstream wastewater treatment plant), collecting data for 72 hours at 10-minute intervals (432 sets in total). Invalid data (e.g., conductivity > 2000 μS / cm, indicating instrument malfunction) was filtered out using Excel. The remaining 418 sets of data were used to perform linear regression using Origin software: in the COD correlation model... =0.25、 =8.3 ( =0.92), in the ammonia nitrogen correlation model =0.06、 =2.1( =0.89), temperature correction factor A value of 0.022 / ℃ was used (because the local pipe network water temperature is 5-30℃, the average deviation of conductivity at different temperatures is 0.022). 30 sets of water samples were tested in the laboratory. For example, if the measured COD of a certain set is 85mg / L, the model calculated value is 83.2mg / L, with a relative error of 2.1%, which meets the requirements.

[0057] Example 3, please refer to the appendix. Figure 1 In the "critical node screening" process, open the topology database visualization interface and screen according to "30-50m before branch pipe merges into main pipe" and "pipe diameter change ≥ 50mm": for example, 40m before a branch pipe DN200 merges into a main pipe DN500 (node ​​J12), and where DN500 changes to DN600 (node ​​J18), screening a total of 10 nodes. On-site, use a tape measure to measure the depth of the inspection well (ensuring space for terminal installation), eliminate J15 which is submerged in water, and replace it with upstream J14. Install a low-cost terminal (cost 185 yuan, accuracy ±1μS / cm), and fix it to the well wall 15cm from the water surface with cable ties. The module (800MHz band) is bound to a data center and tests the frequency every 10 minutes. After powering on, the self-test shows that the data transmission is normal.

[0058] Specifically, after receiving the data, the data processing center first verifies the following: data with missing timestamps (e.g., due to terminal network outage), blurred node IDs (e.g., worn ID labels), or conductivity outside the range (domestic sewage > 1500 μS / cm) is marked as invalid. Missing data is interpolated linearly: for example, if node J12 has no data between 10:00 and 10:20, the data from 9:50 is used. =850μS / cm) and 10:20 ( =880μS / cm) Calculated at 10:00 as 860μS / cm and at 10:10 as 870μS / cm. Noise reduction was achieved using a moving average. Detection frequency was 10 min / time, window n=5, as shown in groups 1-5. The values ​​are 850, 852, 848, 855, and 851. After smoothing, the result is (850+852+848+855+851) / 5 = 851.2 μS / cm. The final result is "Node ID-Time-". -Water Temperature dataset.

[0059] Specifically, the process for calculating the hydraulic propagation time is as follows: Extract the pipe segment parameters from node J12 (upstream) to J18 (downstream) from the topology database: L=60m, d=500mm (concrete pipe). =0.013), slope s=2‰(0.002), calculate the flow velocity using Manning's formula. Hydraulic radius R = d / 4 = 0.125m ≈0.62m / s, then calculate : =(60×1000) / (0.62×60)≈16.13min, rounded to 16min, is used to correct the time difference in subsequent correlation analysis (e.g., if J12 abruptly changes at 10:00, J18 should abruptly change at around 10:16).

[0060] Specifically, the criteria for identifying pollution sources include first checking whether the node is a sudden change node. >15%, domestic sewage): such as J12 at 10:00 =22% (mutation), downstream J18 at 10:16 =18% (mutation, later than J1216 minutes, consistent with...) Then check upstream: J12 upstream J10 =5% (non-mutated), no upstream mutation nodes, finally calculate the pollution concentration: =1200μS / cm, T=20℃, COD=0.25×1200×(1+0.022×(20-25))+8.3≈275.3mg / L (normal ≤300mg / L, determined to be a pollution source node, the location report marks J12 as the branch pipe entering the pollution source node.

[0061] Working principle:

[0062] First, basic data on the target area's pipe network is collected and verified. A directed graph topology model containing node parameters and pipe segment attributes, along with a topology database stored in an adjacency matrix, is constructed. Then, conductivity, COD, and ammonia nitrogen concentrations are simultaneously detected at the end of the pipe network. After data processing, a conductivity-water quality index correlation model with a temperature correction coefficient is constructed and calibrated. Subsequently, key nodes are selected based on the topology model, and low-cost conductivity detection terminals are deployed to collect and transmit real-time data. The received data undergoes integrity verification, interpolation, and noise reduction to form a standardized time-series dataset. The hydraulic propagation time of nodes is calculated using the topology model. Abrupt nodes and associated nodes are selected through conductivity relative amplitude and correlation analysis. Pollution source nodes are identified and located according to preset standards. Finally, on-site sampling and verification of pollution source nodes are conducted, and the model is calibrated or a correction report is issued, forming a closed loop of "modeling-detection-analysis-judgment-verification" for source tracing. At this point, the entire workflow is complete.

[0063] Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make possible changes and modifications without departing from the spirit and scope of the present invention. Therefore, any modifications, equivalent changes and alterations made to the above embodiments based on the technical essence of the present invention without departing from the content of the technical solution of the present invention shall fall within the protection scope defined by the claims of the present invention.

Claims

1. A low-cost, rapid pollution source tracing method based on pipeline network topology analysis, comprising a tracing method, characterized in that, The source tracing method includes the following steps: Step 1: Collect and verify basic data on the pipeline network in the target area, and construct a directed graph topology model. An adjacency matrix is ​​used to store and form a callable topology database; Step 2: Simultaneously and continuously monitor conductivity, COD, and ammonia nitrogen concentrations at the end of the pipeline network, construct and calibrate a conductivity-water quality index correlation model incorporating a temperature correction coefficient; Step 3: Based on the topology model, select key detection nodes in the pipeline network, deploy low-cost conductivity detection terminals and configure detection parameters to achieve real-time data transmission; Step 4: Perform integrity verification, interpolation, and noise reduction on the received detection data to construct a standardized time-series dataset; Step 5: Based on the upstream and downstream relationships in the topology model, calculate the hydraulic propagation time, and screen abrupt change nodes and associated nodes through relative conductivity variation and correlation analysis; Step 6: Verify the mutation nodes and associated nodes according to the preset judgment criteria, locate the pollution source nodes and generate a location report; Step 7: Conduct on-site sampling and verification of pollution source nodes, calibrate the model and correction report based on the verification results, and form a closed loop for source tracing.

2. The low-cost and rapid pollution source tracing method based on pipeline network topology analysis according to claim 1, characterized in that: In step one, the basic data of the pipeline network includes as-built drawings, pipe diameter, pipe slope, node location coordinates, pipe segment length, design flow direction and actual operation flow direction records, and the location and specifications of pipeline network ancillary facilities. In the directed graph topology model, the node set... Each pipeline node contains five parameters: a unique identifier (ID), pipe diameter (d), slope (s), location coordinates, and node type. The edge set... Define three attributes for a pipe segment: water flow direction, pipe segment length L, and pipe segment material. Use an adjacency matrix to store the directed graph data.

3. The low-cost and rapid pollution source tracing method based on pipeline network topology analysis according to claim 2, characterized in that: In step two, a comprehensive detection device is installed at the outlet of the main pipeline network. The detection frequency is uniformly set at 10-15 minutes per detection, continuously collecting basic data for no less than 72 hours. The collected data uses 3... Outliers were removed and the data were standardized. COD and ammonia nitrogen concentration data were converted into time-series datasets that perfectly matched the conductivity detection time. Based on the processed time-series data, a linear correlation model between conductivity and water quality indicators was constructed, incorporating a temperature correction coefficient. The model formula is as follows: ； ； in, This is the calculated COD concentration. This is the calculated value of ammonia nitrogen concentration. , These are the calibration coefficients for the COD correlation model. , These are the calibration coefficients for the ammonia-nitrogen correlation model. This is the actual measured conductivity value. This is the temperature correction factor for conductivity. The actual water temperature at the time of testing. The baseline temperature is 25℃, and the results are verified through 30-50 sets of on-site sampling.

4. The low-cost and rapid pollution source tracing method based on pipeline network topology analysis according to claim 3, characterized in that: In step three, the criteria for screening key inspection nodes in the pipeline network include: nodes 30-50m before a branch pipe merges into a main pipe, adjacent nodes before and after a main pipe connects to a branch pipe, nodes where three or more pipelines converge, and pipeline bends and diameter change nodes with a pipe diameter change range of ≥50mm. Low-cost conductivity detection terminals are installed at key nodes. The accuracy of the terminals is set to ±1μS / cm. The detection frequency is configured according to the scale of the pipeline network: 5-15 min / time for small and medium-sized pipelines with a total length ≤10km, and 20-60 min / time for large-scale pipelines with a total length >10km. The real-time detection data is transmitted to the data processing center.

5. A low-cost, rapid pollution source tracing method based on pipeline network topology analysis according to claim 4, characterized in that: In step four, after receiving the data, the data processing center first performs data integrity verification. Data with missing timestamps, ambiguous node IDs, or conductivity values ​​exceeding a reasonable range are marked as invalid data. Missing valid data is supplemented using interpolation. The verified conductivity data is then subjected to noise reduction processing, and a moving average method is used to smooth data fluctuations. The moving window size is... Determined based on the detection frequency, when the detection frequency is 5-15 minutes / time. =5, 20-60 min / time =3, the calculation formula is as follows: ； in, For the first Smoothed conductivity value for each detection cycle For the first The raw conductivity values ​​of each detection cycle are used to sort the smoothed conductivity data of each key node by timestamp and associate them one by one with the node IDs in the topology model to construct a time-series dataset containing node ID, detection time, smoothed conductivity value, and water temperature.

6. The low-cost and rapid pollution source tracing method based on pipeline network topology analysis according to claim 1, characterized in that: In step five, the upstream and downstream relationships of each key node and the hydraulic parameters of the pipe segment are extracted from the topology database, and the hydraulic propagation time between upstream and downstream nodes is calculated. The calculation formula is: ； in: The water flow velocity within the pipe section is determined by the pipe diameter. Calculated using Manning's formula For pipe segment length, based on time series dataset and Calculate the correlation between the conductivity variation of each node and the upstream and downstream nodes: Conductivity variation calculation: Nodes are calculated using a sliding time window. In the The relative change in conductivity at time t is given by the formula: ； in, For nodes exist The relative amplitude at any given moment, The duration of the sliding window. For nodes exist The smoothed conductivity value at time t. ) is a node exist The smoothed conductivity value at any given time; Upstream and downstream node correlation calculation: The correlation degree is used to calculate the upstream node. With downstream nodes The conductivity correlation is used to introduce the hydraulic propagation time correction time difference, and the formula is: ； in, For nodes and The degree of correlation, The number of valid data sets, For nodes The mean smooth conductivity, For nodes The average smooth conductivity is used to set the amplitude threshold according to the pipeline network type. With correlation threshold Filter out mutation nodes and their correlation with mutation nodes The associated nodes form a node change relationship matrix.

7. The low-cost and rapid pollution source tracing method based on pipeline network topology analysis according to claim 1, characterized in that: In step six, based on the obtained node change relationship matrix and topological model, the criteria for determining pollution source nodes are set: (1) The node is a mutation node, and the mutation time of all its downstream related nodes is later than that of the node; (2) There are no mutation nodes among the upstream associated nodes of this node that meet the above conditions; (3) The pollution concentration calculated by the correlation model exceeds the preset range of the corresponding pipeline network: The location rules for different node types are as follows: the branch pipe corresponding to the node before the branch pipe merges is the branch pipe into which the pollution source merges; the main pipe merging node is the core pollution source node if it meets the conditions; and the turning and diameter change nodes are pollution source nodes within the pipe section if they meet the conditions. The location report must clearly state the node ID, specific location, time of pollution occurrence, estimated pollution concentration, and scope of impact.

8. The low-cost and rapid pollution source tracing method based on pipeline network topology analysis according to claim 1, characterized in that: In step seven, 2-3 sets of water samples are collected on-site for verification. The detection methods comply with standards GB / T11914-1989 and HJ535-2009. If the relative error is ≤15%, the verification is passed. If it is greater than 15%, the correlation model is recalibrated and steps five and six are repeated. After verification, the source of pollution is confirmed by manual investigation, and supplementary treatment measures are taken to form a closed loop of location-verification-source tracing.

Images