A municipal drainage pipeline flow monitoring method based on intelligent sensors

Abstract

This invention discloses a method for monitoring the flow of municipal drainage pipelines based on intelligent sensors, comprising the following steps: S1, deploying intelligent sensors at node wells and pipe segments of the drainage pipeline to generate pipe segment monitoring data; S2, generating a pipeline network monitoring map; S3, determining the upstream and downstream relationships of pipe segments and generating hydraulic characteristic data; S4, inputting the hydraulic characteristic data into an improved StemGNN model, introducing a liquid vein coherence spectrum mechanism to generate initial flow results; S5, using the CMA-ES algorithm, introducing a hydraulic trajectory self-consistency mechanism to generate flow correction parameters; S6, performing flow correction; S7, updating the flow correction parameters to generate continuous flow results. This invention achieves high-precision flow calculation, continuous expression of flow distribution, and consistency correction of inflow and outflow at node wells in municipal drainage networks under complex hydraulic conditions, improving the stability of flow monitoring results and their adaptability to engineering applications.

Description

Technical Field

[0001] This invention relates to the field of municipal drainage network monitoring technology, and in particular to a method for monitoring the flow of municipal drainage pipelines based on intelligent sensors. Background Technology

[0002] With the continuous expansion of urban drainage systems and the increasing demand for refined operation and maintenance, continuous monitoring and dynamic evaluation technologies for flow status in municipal drainage networks have received widespread attention. Existing methods for monitoring drainage pipeline flow mainly rely on single-point velocity meter measurements or indirect estimation methods based on empirical formulas for flow calculation. However, these methods commonly suffer from the following problems in practical applications: The flow state within drainage pipes exhibits significant non-steady-state characteristics due to the influence of rainfall, sewage discharge, and pipe network structure. The coupling relationship between velocity, water level, and pressure signals is complex, and traditional single-parameter measurement methods struggle to accurately reflect actual flow changes, resulting in large fluctuations and poor stability in flow calculation results. Hydraulic conduction between multiple node wells exhibits significant spatial correlation characteristics. Existing methods typically ignore the topological connections between node wells and the upstream-downstream transmission effects, leading to an inability to effectively characterize the distribution and transmission process of flow within the pipe network, thus affecting the overall accuracy of pipe network monitoring. During multi-source sensor data acquisition, velocity, water level, and pressure data suffer from inconsistent sampling frequencies and time stamp deviations. Existing data alignment methods cannot guarantee the synchronization and continuity between different types of data, thereby affecting the accuracy of subsequent flow calculations. For the nonlinear flow characteristics under complex pipe network structures, traditional flow calculation methods based on empirical formulas or simple regression models are ill-suited to varying operating conditions, easily leading to the accumulation of local errors that are amplified at node wells, resulting in an unbalanced overall flow distribution and reducing the reliability of drainage system operational status assessments.

[0003] Therefore, how to provide a method for monitoring the flow of municipal drainage pipelines based on intelligent sensors is a problem that urgently needs to be solved by those skilled in the art. Summary of the Invention

[0004] One objective of this invention is to propose a method for monitoring the flow of municipal drainage pipelines based on intelligent sensors. This invention fully utilizes intelligent sensors deployed at node wells and pipe sections to acquire pipe section monitoring data, constructs a pipe network monitoring map by combining the pipe network topology, generates hydraulic characteristic data through water level difference, upstream and downstream relationship of pipe sections, and pipe section fullness, inputs the hydraulic characteristic data into an improved StemGNN model and introduces a liquid vein coherent conduction spectrum mechanism to generate initial flow results, and simultaneously combines the CMA-ES algorithm and introduces a hydraulic trajectory self-consistency mechanism to generate flow correction parameters, corrects the initial flow results, and completes flow deviation calculation and updating by combining the pipe network topology to form continuous flow results. It has the advantages of high flow calculation accuracy, strong flow distribution continuity, and good hydraulic conduction consistency.

[0005] A method for monitoring the flow of municipal drainage pipelines based on smart sensors according to an embodiment of the present invention includes the following steps: S1. Install smart sensors at the node wells and pipe sections of the drainage pipeline to collect the flow velocity, water level and pressure of the pipe section, perform alignment processing, and generate pipe section monitoring data; S2. Obtain the pipe segment connection relationship, node well location, pipe diameter, slope and pipe bottom elevation of the drainage pipe network, construct the pipe network topology, and map the pipe segment monitoring data to the corresponding pipe segment to generate a pipe network monitoring map; S3. Calculate the water level difference between adjacent node wells based on the pipeline network monitoring map, determine the upstream and downstream relationship of the pipeline segment according to the pipeline segment connection relationship, calculate the pipeline segment filling degree in combination with the pipe diameter and slope, and generate hydraulic characteristic data. S4. Input the hydraulic characteristic data into the improved StemGNN model, introduce the liquid vein coherent spectrum guidance mechanism in the spectrum construction module, generate the liquid flow coherent relationship based on the water level difference, pipe diameter, slope and upstream and downstream relationship of the pipe section, guide and correct the hydraulic propagation relationship between node wells, and generate the initial flow result. S5. Based on the initial flow rate result and the hydraulic characteristic data, the CMA-ES algorithm is used to introduce a hydraulic trajectory self-consistency mechanism and generate flow correction parameters through consistency calculation. S6. Apply the flow correction parameters to the initial flow result, calculate the water flow capacity of the pipe section based on the pipe section cross-sectional area, pipe section filling degree and water level difference, and perform flow correction in combination with the upstream and downstream relationship of the pipe section to generate the corrected flow rate of the pipe section. S7. Input the corrected flow rate of the pipe segment into the pipe network topology, calculate the flow deviation according to the inflow and outflow of the node well, and if there is a flow deviation, update the flow correction parameters to generate continuous flow results.

[0006] Optionally, S1 specifically includes: Flow velocity sensors, water level sensors, and pressure sensors are installed on the well wall or inside the well casing of the node well, and flow velocity sensors are installed inside the pipe section. The flow velocity sensor, water level sensor, and pressure sensor are synchronously sampled at a uniform sampling frequency to obtain the flow velocity data, water level data, and pressure data of the corresponding node well and pipe section. The flow velocity data, water level data, and pressure data are time-stamped and matched. The flow velocity data, water level data, and pressure data corresponding to the same time stamp are combined to form the pipe section monitoring data.

[0007] Optionally, the step of obtaining the pipe segment connection relationships, node well locations, pipe diameters, slopes, and pipe bottom elevations of the drainage pipe network, and constructing the pipe network topology, specifically involves: Retrieve the node well number, node well spatial coordinates, and pipe segment number from the pipeline network database; Establish the connection relationship between pipe segments based on the starting point node well and the ending point node well of the pipe segment, and form the connection relationship between node wells; Read the pipe diameter, slope, and bottom elevation of the corresponding pipe segment, and associate them according to the pipe segment number; The topology between node wells and pipe segments is established based on the spatial coordinates of the node wells and the connection relationship of the pipe segments, thus generating the pipeline topology.

[0008] Optionally, the step of mapping pipe segment monitoring data to corresponding pipe segments to generate a pipe network monitoring map specifically involves: The monitoring data of the pipe section is identified and matched according to the node well number and pipe section number, and the flow velocity data, water level data and pressure data are matched with the corresponding pipe section number; The flow velocity, water level, and pressure data at the same time point are aligned according to the time identifier and associated with the corresponding pipe segment number. The flow velocity data, water level data, and pressure data corresponding to the pipe segment number are loaded into the corresponding pipe segment location in the pipe network topology to form a pipe network monitoring map containing pipe segment monitoring data.

[0009] Optionally, S3 specifically includes: Extract the water level data corresponding to the starting node well and the ending node well of the same pipe segment from the pipeline monitoring map, and calculate the water level difference between the node wells by subtracting the water level data of the ending node well from the water level data of the starting node well. The starting point node well and the ending point node well are determined based on the connection relationship of the pipe segments, and the upstream and downstream relationship of the pipe segments is determined according to the positive and negative correspondence of the slope. The pipe section cross-sectional height is calculated based on the pipe diameter, and the water depth is obtained by calculating the difference between the node well water level data and the pipe bottom elevation. The pipe section filling degree is obtained by calculating the ratio of the water depth to the pipe section cross-sectional height. Hydraulic characteristic data are generated by combining the water level difference, the upstream and downstream relationship of the pipe segment, and the filling degree of the pipe segment according to the pipe segment number.

[0010] Optionally, the improved StemGNN model specifically includes a feature encoding module, a graph construction module, a spectral domain operation module, and a flow output module; The feature encoding module, based on hydraulic feature data, sorts the water level difference, upstream and downstream relationship of the pipe segment, pipe segment fullness, pipe diameter and slope according to the pipe segment number. It arranges the water level difference, upstream and downstream relationship of the pipe segment, pipe segment fullness, pipe diameter and slope corresponding to the same pipe segment number in a fixed order to form a pipe segment feature vector. Then, it arranges the pipe segment feature vectors row by row according to the pipe segment number order to generate a pipe segment feature matrix. The graph construction module establishes an initial adjacency matrix based on the pipeline network topology. The row and column positions of the initial adjacency matrix correspond to the node well numbers, and the connection values ​​of the initial adjacency matrix correspond to the pipe segment connection relationships between the node wells. A fluid flow coherence guide spectrum mechanism is introduced in the spectrum construction module. The water level direction is determined based on the sign of the water level difference, the slope direction based on the sign of the slope, and the pipe segment transmission volume based on the upstream and downstream relationship of the pipe segment. The water level direction, the slope direction, and the pipe segment transmission volume are multiplied to generate a propagation direction term. The pipe segment cross-sectional area is calculated based on the pipe diameter, and the cross-sectional area is multiplied by the pipe segment filling degree to generate a flow rate term. The propagation direction term and the flow rate term are normalized respectively, and the normalized propagation direction term is multiplied by the normalized flow rate term to generate a fluid flow coherence relation. The fluid flow coherence relation is arranged according to the node well number to generate a coherence matrix, and the coherence matrix is ​​multiplied by the connection values ​​at corresponding positions in the initial adjacency matrix to generate a guide spectrum adjacency matrix. The spectral domain operation module calculates the degree matrix based on the guided spectrum adjacency matrix, subtracts the degree matrix from the guided spectrum adjacency matrix to generate a graph Laplacian matrix, performs eigenvalue decomposition on the graph Laplacian matrix to generate a graph spectral basis, and multiplies the pipe segment feature matrix with the graph spectral basis to generate pipe segment spectral domain features. The flow output module converts the pipe segment spectral domain characteristics into pipe segment flow values ​​and generates initial flow results by arranging them according to the pipe segment number.

[0011] Optionally, S5 specifically includes: The initial flow rate results and the hydraulic characteristic data are arranged according to the pipe segment number to generate a flow rate data matrix and a hydraulic characteristic matrix; A set of parameter vectors is constructed based on the row vectors of the traffic data matrix, and the mean vector and covariance matrix are calculated to generate the parameter distribution. Multiple sets of parameter vectors are generated based on the parameter distribution, and each set of parameter vectors is calculated in correspondence with the initial flow result to generate multiple sets of flow results; The flow difference corresponding to the water level difference, the flow transmission difference corresponding to the upstream and downstream relationship of the pipe section, and the flow change difference corresponding to the pipe section fullness are calculated based on multiple sets of flow results and hydraulic feature matrix. The flow difference, the flow transmission difference, and the flow change difference are then weighted and summed to generate a consistent value. The multiple sets of parameter vectors are sorted based on the consistency values, and the top N sets of parameter vectors are selected to update the mean vector and covariance matrix, where N is a preset positive integer. The flow correction parameters are generated based on the updated mean vector and the updated covariance matrix.

[0012] Optionally, S6 specifically includes: Arrange the flow correction parameters according to the pipe segment number and the corresponding initial flow result to generate a flow correction matrix; Calculate the cross-sectional area of ​​the pipe section based on the pipe diameter, and multiply the cross-sectional area of ​​the pipe section by the filling degree of the pipe section to generate the effective water flow area; Multiply the effective water flow area by the water level difference to generate the water flow capacity of the pipe section; multiply the flow correction matrix by the initial flow result to generate the corrected flow result; The flow direction sign of the corrected flow rate result is determined based on the upstream and downstream relationship of the pipe section, and the flow direction sign is multiplied by the water carrying capacity of the pipe section to generate the hydraulic correction amount; The corrected flow rate result is added to the hydraulic correction amount to generate the corrected flow rate for the pipe section.

[0013] Optionally, the step of inputting the corrected flow rate of the pipe segment into the pipe network topology and calculating the flow deviation according to the inflow and outflow rates of the node wells specifically involves: In the pipeline network topology, the corrected flow rate of the pipe segment is calibrated according to the pipe segment number, and the connection correspondence between the node well and the pipe segment is established; Using the node well number as an index, the pipe segments connected to the same node well are divided into inflow direction pipe segments and outflow direction pipe segments; The corrected flow rates of the pipe segments corresponding to the inflow direction under the same time identifier are added one by one to generate the node well inflow rate; the corrected flow rates of the pipe segments corresponding to the outflow direction under the same time identifier are added one by one to generate the node well outflow rate; the node well inflow rate is subtracted from the node well outflow rate to obtain the flow deviation.

[0014] Optionally, if a flow deviation exists, the flow correction parameters are updated to generate continuous flow results, specifically as follows: The flow deviation is allocated to the corresponding pipe segment according to the node well connection relationship to generate the pipe segment deviation amount; Using the pipe segment number as an index, establish the correspondence between the pipe segment deviation and the flow correction parameters, and generate the parameter correction amount; The flow correction parameters are added to the parameter correction values ​​one by one to generate the updated flow correction parameters; The updated flow correction parameters are multiplied one by one with the initial flow result to generate the updated pipe section corrected flow rate; The updated pipeline flow rate is arranged in chronological order based on the updated pipeline segment correction flow rate to generate continuous flow rate results.

[0015] The beneficial effects of this invention are: By deploying smart sensors at node wells and pipe sections to acquire pipe section monitoring data, and combining the pipe network topology to construct a pipe network monitoring map, the flow velocity data, water level data, and pressure data can be accurately matched under a unified time identifier, solving the problem of asynchronous multi-source data and improving the continuity and consistency of pipe section monitoring data. Hydraulic characteristic data are generated based on water level difference, upstream and downstream relationship of pipe segment and pipe segment fullness. A liquid vein coherent guide spectrum mechanism is introduced into the improved StemGNN model to embed the hydraulic conduction direction and water flow capacity into the guide spectrum adjacency matrix, thereby enhancing the ability to express the hydraulic propagation relationship between node wells and improving the accuracy of the initial flow rate results. Based on the initial flow results and hydraulic characteristic data, the CMA-ES algorithm is adopted and a hydraulic trajectory self-consistency mechanism is introduced to dynamically optimize the flow correction parameters, reduce the accumulation of errors caused by flow difference, flow transmission difference and flow change difference, and improve the stability of flow calculation results. By applying flow correction parameters to the initial flow result, a corrected flow rate for the pipe section is generated. The flow deviation is calculated in conjunction with the pipe network topology, and the flow correction parameters are continuously updated to form a continuous flow result, thereby improving the continuity of the overall flow distribution and hydraulic consistency of the drainage pipe network. Attached Figure Description

[0016] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings: Figure 1 This is a flowchart of a municipal drainage pipeline flow monitoring method based on intelligent sensors proposed in this invention; Figure 2 This is a schematic diagram of the improved StemGNN model proposed in this invention; Figure 3 This is a data flow diagram of a municipal drainage pipeline flow monitoring method based on intelligent sensors proposed in this invention. Detailed Implementation

[0017] The present invention will now be described in further detail with reference to the accompanying drawings. These drawings are simplified schematic diagrams, illustrating only the basic structure of the invention, and therefore only show the components relevant to the invention.

[0018] refer to Figures 1-3 A method for monitoring the flow of municipal drainage pipelines based on intelligent sensors includes the following steps: S1. Install smart sensors at the node wells and pipe sections of the drainage pipeline to collect the flow velocity, water level and pressure of the pipe section, perform alignment processing, and generate pipe section monitoring data; S2. Obtain the pipe segment connection relationship, node well location, pipe diameter, slope and pipe bottom elevation of the drainage pipe network, construct the pipe network topology, and map the pipe segment monitoring data to the corresponding pipe segment to generate a pipe network monitoring map; S3. Calculate the water level difference between adjacent node wells based on the pipeline network monitoring map, determine the upstream and downstream relationship of the pipeline segment according to the pipeline segment connection relationship, calculate the pipeline segment filling degree in combination with pipe diameter and slope, and generate hydraulic characteristic data; S4. Input the hydraulic characteristic data into the improved StemGNN model, introduce the liquid vein coherent spectrum guidance mechanism in the spectrum construction module, generate the liquid flow coherent relationship based on the water level difference, pipe diameter, slope and upstream and downstream relationship of the pipe section, guide and correct the hydraulic propagation relationship between node wells, and generate the initial flow result. S5. Based on the initial flow results and hydraulic characteristic data, the CMA-ES algorithm is adopted, and a hydraulic trajectory self-consistency mechanism is introduced to generate flow correction parameters through consistency calculation. S6. Apply the flow correction parameters to the initial flow result, calculate the water flow capacity of the pipe section based on the pipe section cross-sectional area, pipe section filling degree and water level difference, and perform flow correction in combination with the upstream and downstream relationship of the pipe section to generate the corrected flow rate of the pipe section. S7. Input the corrected flow rate of the pipe section into the pipe network topology, calculate the flow deviation according to the inflow and outflow of the node well, and update the flow correction parameters if there is a flow deviation to generate continuous flow results.

[0019] In this embodiment, S1 specifically refers to: A flow velocity sensor, a water level sensor, and a pressure sensor are installed on the inner side of the node well wall. Cable fasteners and waterproof connectors are installed on the inner side of the node well shaft. A flow velocity sensor is installed in the pipe section near the center streamline of the pipe. Establish a correspondence between node well number, pipe segment number, sensor number and installation location, and write the correspondence into the sensor acquisition terminal; The flow velocity sensor collects the water flow velocity in the pipe section, the water level sensor collects the liquid level height in the node well, and the pressure sensor collects the water pressure in the node well. The sensor acquisition terminal collects flow velocity data, water level data, and pressure data at the same sampling frequency, and writes a time stamp generated by the same clock source for each sampling result. Flow velocity data, water level data, and pressure data are arranged according to node well number, pipe section number, sensor number, and time identifier; flow velocity data, water level data, and pressure data under the same time identifier are combined into a single monitoring record; multiple monitoring records corresponding to the same pipe section number are arranged in chronological order according to the time identifier to generate pipe section monitoring data.

[0020] In this embodiment, the pipe segment connection relationships, node well locations, pipe diameters, slopes, and pipe bottom elevations of the drainage pipe network are obtained to construct the pipe network topology, specifically as follows: Obtain the node well number, node well spatial coordinates, pipe segment number, pipe segment starting node well number, pipe segment ending node well number, pipe diameter, slope, and pipe bottom elevation from the pipeline network database; Arrange the node well numbers and their spatial coordinates to form node well location data; arrange the pipe segment numbers, the starting node well numbers, and the ending node well numbers to form pipe segment connection relationships; arrange the pipe segment numbers, pipe diameters, slopes, and bottom elevations to form pipe segment attribute data. Connect the pipe segment numbers to the corresponding well numbers according to the starting and ending well numbers of the pipe segment; determine the spatial location of the well in the drainage network according to the spatial coordinates of the well; determine the connection path between wells according to the connection relationship of the pipe segments; and mark the pipe diameter, slope and bottom elevation to the corresponding pipe segment numbers according to the pipe segment attribute data. The pipeline topology is formed by the node well number, node well spatial coordinates, pipe segment number, pipe segment connection relationship and pipe segment attribute data.

[0021] In this embodiment, the monitoring data of the pipe segment is mapped to the corresponding pipe segment to generate a pipe network monitoring map, specifically as follows: The node well number, pipe segment number, time stamp, flow velocity data, water level data, and pressure data are obtained from the pipe segment monitoring data. The node well number is matched one-to-one with the node well number in the pipe network topology, and the pipe segment number is matched one-to-one with the pipe segment number in the pipe network topology. The flow velocity data, water level data, and pressure data corresponding to the same pipe segment number are grouped according to the time stamp, with each group of data corresponding to one time stamp. The flow velocity data, water level data, and pressure data under the same time identifier are arranged in the order of flow velocity data first, water level data in the middle, and pressure data last to form a pipe section monitoring record; Multiple time markers correspond to pipe segment monitoring records arranged in chronological order to form pipe segment time-series monitoring data; the pipe segment time-series monitoring data is then labeled to the corresponding pipe segment location in the pipeline network topology. By combining the pipeline topology, node well number, pipe segment number, and pipe segment time-series monitoring data, a pipeline monitoring map is formed.

[0022] In this embodiment, S3 specifically refers to: Obtain the pipe segment number, starting node well number, ending node well number, starting node well water level data, ending node well water level data, pipe diameter, slope and pipe bottom elevation corresponding to the same pipe segment from the pipeline network monitoring map; Subtract the water level data of the endpoint node well from the water level data of the starting node well to obtain the water level difference between adjacent node wells; In the pipe segment connection relationship, the node well located at the beginning of the pipe segment is called the starting node well, and the node well located at the end of the pipe segment is called the ending node well. When the slope is positive, the starting node well corresponds to the upstream end of the pipe segment, and the ending node well corresponds to the downstream end of the pipe segment; when the slope is negative, the ending node well corresponds to the upstream end of the pipe segment, and the starting node well corresponds to the downstream end of the pipe segment; when the pipe diameter is the inner diameter of a circular pipe segment, the height of the pipe segment section is taken as the pipe diameter value. Subtract the pipe bottom elevation from the node well water level data to obtain the water depth; divide the water depth by the pipe section cross-sectional height to obtain the pipe section fill degree. The water level difference, upstream and downstream relationships of the pipe section, and pipe section filling degree are arranged according to the pipe section number to generate hydraulic characteristic data.

[0023] In this embodiment, the improved StemGNN model specifically includes a feature encoding module, a graph construction module, a spectral domain operation module, and a flow output module; After the hydraulic feature data enters the feature encoding module, a one-to-one correspondence is established between the water level difference, upstream and downstream relationship of the pipe segment, pipe segment fullness, pipe diameter, and slope according to the pipe segment number. The water level difference, upstream and downstream relationship of the pipe segment, pipe segment fullness, pipe diameter, and slope corresponding to the same pipe segment number are arranged in the following order: water level difference first, upstream and downstream relationship of the pipe segment second, pipe segment fullness in the middle, pipe diameter second, and slope last, forming a pipe segment feature vector. Multiple pipe segment feature vectors are arranged row by row in ascending order of pipe segment number, forming a pipe segment feature matrix. After the pipeline topology is entered into the map construction module, the node well numbers are arranged in the order of the numbers as matrix row numbers and matrix column numbers; two node wells with pipe segment connection relationship are assigned a value of 1 at the corresponding position in the initial adjacency matrix, and two node wells without pipe segment connection relationship are assigned a value of 0 at the corresponding position in the initial adjacency matrix, thus forming the initial adjacency matrix; The spectrum construction module introduces a liquid vein coherent spectrum mechanism. When the water level difference is greater than 0, the water level direction measurement is 1; when the water level difference is equal to 0, the water level direction measurement is 0; and when the water level difference is less than 0, the water level direction measurement is -1. Similarly, when the slope is greater than 0, the slope direction measurement is 1; when the slope is equal to 0, the slope direction measurement is 0; and when the slope is less than 0, the slope direction measurement is -1. When the upstream and downstream relationship of a pipe segment is consistent with the direction of the water level difference, the pipe segment transmission quantity is 1; when the upstream and downstream relationship of a pipe segment is opposite to the direction of the water level difference, the pipe segment transmission quantity is -1; and when there is no water level difference in the corresponding pipe segment, the pipe segment transmission quantity is 0. Multiplying the water level direction quantity, the slope direction quantity, and the pipe segment transmission quantity yields the propagation direction term. Multiplying the pipe diameter by the pipe diameter yields the square of the pipe diameter; multiplying the square of the pipe diameter by pi and dividing by 4 yields the cross-sectional area of ​​the pipe segment; multiplying the cross-sectional area of ​​the pipe segment by the pipe diameter... The effective water flow area is obtained by calculating the fullness of the pipe section. At the same sampling time, the effective water flow areas of all pipe sections are arranged according to their numerical values. The effective water flow area minus the minimum effective water flow area is divided by the difference between the maximum and minimum effective water flow areas to obtain the normalized water flow scale term. If the maximum and minimum effective water flow areas are equal, the normalized water flow scale term is set to 1. The propagation direction term is used as the normalized propagation direction term. The normalized propagation direction term is multiplied by the normalized water flow scale term to obtain the fluid flow coherence relation. The fluid flow coherence relation is arranged according to the node well number and pipe section number to the matrix position between the corresponding node wells to form the coherence matrix. The value at the corresponding position in the coherence matrix is ​​multiplied by the connection value at the corresponding position in the initial adjacency matrix to obtain the lead spectrum adjacency matrix. After receiving the guided spectrum adjacency matrix, the spectral domain operation module adds the connection values ​​of each row of the guided spectrum adjacency matrix to obtain the degree value of the corresponding node well. The degree value of the node well is written into the main diagonal position of the matrix according to the node well number, and the non-main diagonal positions are assigned a value of 0 to form the degree matrix. The degree matrix is ​​subtracted from the guided spectrum adjacency matrix to obtain the graph Laplacian matrix. The graph Laplacian matrix is ​​decomposed into eigenvalues ​​and eigenvectors. The eigenvectors are arranged in ascending order of the corresponding eigenvalues ​​to form the graph basis. The pipe segment feature matrix is ​​multiplied with the graph basis to obtain the pipe segment spectral domain feature. The pipe segment spectral domain feature is multiplied with the flow output coefficient matrix in the flow output module and superimposed with the flow output bias vector to obtain the pipe segment flow value. The pipe segment flow value is arranged according to the pipe segment number and time identifier to generate the initial flow result.

[0024] In this embodiment, both the improved StemGNN model and the StemGNN model adopt a joint modeling approach combining temporal features and graph structures. Both include feature encoding, graph structure construction, and spectral domain computation, and utilize the graph Laplacian matrix to characterize the connectivity between node wells. The improvement process introduces a fluid flow coherence guide spectrum mechanism, which maps the water level difference direction, slope direction, and upstream-downstream relationship of the pipe segment into a unified propagation direction term. This term is then normalized and combined with a flow rate term composed of pipe diameter and pipe segment fill degree to form a fluid flow coherence with both directional and hydraulic scale constraints. The relationship between fluid flow coherence is embedded in the corresponding positions of the adjacency matrix, changing the traditional StemGNN model where the adjacency matrix only represents structural connections, and directly integrating hydraulic propagation characteristics into the graph structure. In the graph construction stage, the initial adjacency matrix is ​​corrected position-by-position through the coherence matrix, ensuring that the connection weights between node wells simultaneously reflect connectivity and hydraulic conduction capacity. In the spectral domain computation stage, the graph Laplacian matrix is ​​constructed through the guided adjacency matrix, ensuring that spectral features simultaneously include structural and hydrodynamic information. In the feature encoding stage, the feature arrangement is constrained by the pipe segment numbering order, ensuring... Water level difference, fullness, pipe diameter, and slope form a stable feature vector structure, avoiding information distortion caused by feature mixing. During the flow-scale processing, a normalization step eliminates dimensional differences, making the hydraulic characteristics of different pipe sections comparable. Symbolic processing during propagation direction modeling avoids the impact of continuous numerical fluctuations on propagation direction judgment. The improved model maintains consistent expression of hydraulic propagation direction under complex pipe network connections, maintains stability in flow estimation even with significant variations in pipe fullness, and effectively distinguishes the impact of inflow and outflow paths on flow results in multi-node well confluence areas, thereby improving the continuity and physical consistency of flow calculation results and reducing the probability of misjudgment in low-velocity or reverse-flow scenarios. By embedding hydraulic features into a graph structure, the dependence on historical data length is reduced, improving the model's adaptability to varying data sampling intervals. Overall, the improvement enhances the expressive power of the graph structure and explicitly introduces hydraulic constraints without changing the basic StemGNN computational framework, thus improving applicability and computational accuracy in municipal drainage network flow monitoring scenarios.

[0025] In this embodiment, S5 specifically refers to: The initial flow rate results are arranged according to pipe segment number and time identifier to form a flow rate data matrix. The hydraulic characteristic data are arranged according to pipe segment number and time identifier to form a hydraulic characteristic matrix. In the flow rate data matrix, the flow rate data corresponding to the same pipe segment number is used as a parameter vector. The parameter vectors corresponding to multiple pipe segment numbers are arranged in the order of pipe segment number to form a parameter vector set. The parameters in the same column of the parameter vector set are added one by one and then divided by the number of parameter vectors to obtain the mean parameter at the corresponding position in the mean vector. Subtract the mean vector from each parameter vector to obtain the parameter deviation vector. Multiply the parameter deviation vector with its transpose to obtain the deviation product matrix. Add all the deviation product matrices together and divide by the number of parameter vectors to obtain the covariance matrix. The parameter distribution is formed by combining the mean vector and the covariance matrix. In the parameter distribution, the parameter center is determined according to the mean vector, and the parameter perturbation direction and parameter perturbation amplitude are determined according to the covariance matrix, generating multiple sets of parameter vectors. Each set of parameter vectors is multiplied by the initial flow result according to the pipe segment number to obtain a corresponding set of flow results. Multiple sets of parameter vectors generate multiple sets of flow results respectively. The flow values ​​corresponding to the node wells at both ends of the same pipe segment in each set of flow results are subtracted to obtain the flow difference corresponding to the water level difference. The flow value of the upstream pipe segment is subtracted from the flow value of the downstream pipe segment in each set of flow results to obtain the flow transmission difference corresponding to the upstream and downstream relationship of the pipe segment. The flow values ​​corresponding to adjacent time markers of the same pipe segment in each set of flow results are subtracted to obtain the flow change difference corresponding to the pipe segment fullness. The absolute values ​​of the flow difference, flow transmission difference, and flow change difference are taken respectively, and the three absolute values ​​are added together to obtain a consistent value. Arrange multiple sets of parameter vectors in ascending order of consistent values. The parameter vectors with the first N positions of the arrangement are selected as the parameter vectors. N is half the number of parameter vectors in the multiple sets, rounded down and not less than 1. The parameters in the same column of the selected parameter vector are added together and then divided by the number of selected parameter vectors to obtain the updated mean parameter at the corresponding position in the updated mean vector. Subtract the updated mean vector from each selected parameter vector to obtain the updated bias vector. Multiply the updated bias vector with the transpose of the updated bias vector to obtain the updated bias product matrix. Add all the updated bias product matrices one by one and divide by the number of selected parameter vectors to obtain the updated covariance matrix. The parameters in the updated mean vector are used as the center values ​​of the flow correction parameters, and the parameters at the main diagonal position of the updated covariance matrix are used as the fluctuation values ​​of the flow correction parameters. The center values ​​and fluctuation values ​​are arranged according to the pipe segment number to generate the flow correction parameters.

[0026] In this embodiment, S6 specifically refers to: A correspondence is established between the flow correction parameters, pipe segment number, and time identifier to form a flow correction matrix; the initial flow results are arranged according to the pipe segment number and time identifier to form an initial flow matrix. The corrected flow value is obtained by multiplying the correction parameter at the corresponding position in the flow correction matrix with the initial flow value at the corresponding position in the initial flow matrix item by item. The corrected flow rates are arranged according to pipe section number and time identifier to form the corrected flow rate results; Calculate the cross-sectional area of ​​the pipe section based on the pipe diameter. Multiply the pipe diameter by the pipe diameter to get the square of the pipe diameter. Multiply the square of the pipe diameter by pi and divide by 4 to get the cross-sectional area of ​​the pipe section. The cross-sectional area of ​​the pipe section is matched with the filling degree of the pipe section according to the pipe section number. The effective water passage area is obtained by multiplying the two items one by one. The water passage capacity of the pipe section is obtained by multiplying the effective water passage area with the absolute value of the water level difference one by one. The flow direction sign is determined based on the upstream and downstream relationship of the pipe section and the direction of the water level difference. If the direction is the same, the sign is 1; if the direction is opposite, the sign is -1; and if the water level difference is 0, the sign is 0. The hydraulic correction is obtained by multiplying the flow direction symbol and the water carrying capacity of the pipe section one by one; the corrected flow rate is then added to the hydraulic correction according to the pipe section number and time identifier to generate the corrected flow rate of the pipe section.

[0027] In this embodiment, the corrected flow rate of the pipe segment is input into the pipe network topology, and the flow deviation is calculated according to the inflow and outflow rates of the node wells, specifically as follows: The pipe segment corrected flow rate is arranged according to the pipe segment number and time identifier, and a corresponding relationship is established with the pipe segment number in the pipe network topology; In the pipeline topology, mark the starting node well number and the ending node well number corresponding to each pipe segment number; Based on the upstream and downstream relationship of the pipe section, determine the node wells into which the corrected flow enters and out of the pipe section; when the corrected flow of the pipe section flows from the upstream node well to the downstream node well, the downstream node well corresponds to the pipe section in the inflow direction, and the upstream node well corresponds to the pipe section in the outflow direction. For the same node well number, the corrected flow rates of the pipe segments corresponding to the inflow direction are added one by one according to the same time identifier to generate the node well inflow rate; for the same node well number, the corrected flow rates of the pipe segments corresponding to the outflow direction are added one by one according to the same time identifier to generate the node well outflow rate. The flow rate deviation of a node well is obtained by subtracting its outflow rate from its inflow rate. The flow rate deviations are then arranged according to the node well number and time identifier to form node well flow rate deviation data.

[0028] In this embodiment, if a flow deviation exists, the flow correction parameters are updated to generate continuous flow results, specifically as follows: The node well number where the flow deviation is not equal to 0 is used as the deviation node well number; The inflow and outflow pipe segments corresponding to the deviation node well numbers are determined by the pipe network topology, pipe segment connection relationships, and upstream and downstream relationships of the pipe segments. When the inflow rate of a node well is greater than the outflow rate of a node well, the flow deviation is evenly distributed to the pipe section in the outflow direction to obtain the pipe section deviation amount corresponding to the outflow direction pipe section. When the inflow rate of a node well is less than the outflow rate of a node well, the absolute value of the flow deviation is evenly distributed to the pipe section in the inflow direction to obtain the pipe section deviation amount corresponding to the inflow direction. Establish a correspondence between pipe segment deviation and flow correction parameters according to pipe segment number and time identifier; When the initial flow rate result is not equal to 0, the pipe segment deviation is divided by the initial flow rate result corresponding to the same pipe segment number and the same time identifier to obtain the parameter correction amount; when the initial flow rate result is equal to 0, the pipe segment deviation is divided by the average value of the initial flow rate results corresponding to adjacent time identifiers of the same pipe segment number to obtain the parameter correction amount. The flow correction parameters are added one by one to the parameter correction values ​​corresponding to the same pipe section number and the same time identifier to generate updated flow correction parameters; The updated flow correction parameters are multiplied one by one with the initial flow results corresponding to the same pipe segment number and the same time identifier to generate the updated pipe segment corrected flow; the updated pipe segment corrected flow is arranged in chronological order according to the pipe segment number and time identifier to generate continuous flow results.

[0029] Example 1: To verify the feasibility of this invention in practice, it was applied to a flow monitoring scenario in a combined sewer system in the old urban area of ​​a city. This area includes 12 node wells and 11 sections of municipal drainage pipes, with pipe diameters ranging from 600mm to 1200mm and slopes from 0.18% to 0.72%. Some sections suffer from long-term sedimentation, low water levels, and short-term backflow after rain. Existing monitoring methods primarily relied on single-point flow meters and manual inspection records. There was a time discrepancy between flow velocity and water level data, and the inflow and outflow from the node wells were frequently unbalanced. Flow fluctuations were particularly pronounced within two hours after rain, making it difficult for drainage dispatchers to determine whether a sudden increase in flow was caused by actual confluence, sensor drift, localized sedimentation, or an imbalance in hydraulic transmission between upstream and downstream areas.

[0030] Water level and pressure sensors are installed inside the well walls of node wells, and flow velocity sensors are installed within the pipe sections. Flow velocity, water level, and pressure data are collected at a uniform sampling frequency of 30 seconds, and time-stamped to form pipe section monitoring data. The system retrieves node well numbers, node well spatial coordinates, pipe section numbers, pipe diameters, slopes, and pipe bottom elevations from the pipe network database to construct the pipe network topology and maps the pipe section monitoring data to the corresponding pipe sections to form a pipe network monitoring map. Subsequently, hydraulic characteristic data is generated based on the water level difference between adjacent node wells, the upstream and downstream relationship of the pipe section, and the pipe section's fill degree, and input into the improved StemGNN model. The improved StemGNN model introduces a fluid vein coherence guidance mechanism in the map construction module, transforming the water level difference sign, slope sign, upstream and downstream relationship of the pipe section, and flow rate scale term into a fluid flow coherence relationship, guiding and correcting the hydraulic propagation relationship between node wells, and outputting the initial flow rate result. Then, the CMA-ES algorithm, which introduces a hydraulic trajectory self-consistency mechanism, generates flow correction parameters to correct the initial flow results, obtain the corrected flow of the pipe section, and calculate the flow deviation between the inflow and outflow of the node well based on the pipe network topology to form a continuous flow result.

[0031] Table 1 Comparison of Flow Monitoring Accuracy for Different Pipe Sections

[0032] As shown in Table 1, the original method exhibits significant deviations in multiple pipe sections, with errors concentrated between 7.74% and 8.95%. Particularly in large-diameter, high-fill-degree pipe sections such as P07 and P11, the original method is prone to amplifying flow rate results due to local velocity fluctuations. This invention incorporates water level difference, upstream-downstream relationship of the pipe section, and pipe section fill degree into the hydraulic characteristic data. It also corrects the hydraulic propagation relationship between node wells through a liquid vein coherent spectrum mechanism, ensuring that the initial flow rate result no longer depends solely on a single-point velocity. After generating flow correction parameters using a hydraulic trajectory self-consistency mechanism and the CMA-ES algorithm, the error of this invention is reduced to 1.31% to 1.73%, indicating that the pipe section corrected flow rate is closer to the manually verified flow rate, and the continuous flow rate results have better accuracy and engineering reliability.

[0033] Table 2 Comparison of Flow Balance and Anomaly Identification Effects in Node Wells

[0034] As shown in Table 2, the original method did not adequately utilize the balance between the inflow and outflow of the node wells. Large inflow-outflow deviations were observed in wells J04, J06, J08, and J12, and the anomaly identification results were inconsistent with manual verification results. This invention inputs the corrected flow rate of the pipe section into the pipeline topology, calculates the flow deviation according to the inflow and outflow of the node wells, and updates the flow correction parameters when flow deviations exist. This reduces the inflow-outflow deviation of the node wells from 28.4 m³·h⁻¹ to 61.3 m³·h⁻¹ to 5.7 m³·h⁻¹ to 10.5 m³·h⁻¹. The deviation reduction rate exceeds 79%, indicating that this invention can effectively suppress the accumulation of local errors at the node wells and has a clearer ability to distinguish between backflow conditions, siltation disturbances, and normal fluctuations.

[0035] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A method for monitoring the flow of municipal drainage pipelines based on intelligent sensors, characterized in that, Includes the following steps: S1. Install smart sensors at the node wells and pipe sections of the drainage pipeline to collect the flow velocity, water level and pressure of the pipe section, perform alignment processing, and generate pipe section monitoring data; S2. Obtain the pipe segment connection relationship, node well location, pipe diameter, slope and pipe bottom elevation of the drainage pipe network, construct the pipe network topology, and map the pipe segment monitoring data to the corresponding pipe segment to generate a pipe network monitoring map; S3. Calculate the water level difference between adjacent node wells based on the pipeline network monitoring map, determine the upstream and downstream relationship of the pipeline segment according to the pipeline segment connection relationship, calculate the pipeline segment filling degree in combination with the pipe diameter and slope, and generate hydraulic characteristic data. S4. Input the hydraulic characteristic data into the improved StemGNN model, introduce the liquid vein coherent spectrum guidance mechanism in the spectrum construction module, generate the liquid flow coherent relationship based on the water level difference, pipe diameter, slope and upstream and downstream relationship of the pipe section, guide and correct the hydraulic propagation relationship between node wells, and generate the initial flow result. S5. Based on the initial flow rate result and the hydraulic characteristic data, the CMA-ES algorithm is used to introduce a hydraulic trajectory self-consistency mechanism and generate flow correction parameters through consistency calculation. S6. Apply the flow correction parameters to the initial flow result, calculate the water flow capacity of the pipe section based on the pipe section cross-sectional area, pipe section filling degree and water level difference, and perform flow correction in combination with the upstream and downstream relationship of the pipe section to generate the corrected flow rate of the pipe section. S7. Input the corrected flow rate of the pipe segment into the pipe network topology, calculate the flow deviation according to the inflow and outflow of the node well, and if there is a flow deviation, update the flow correction parameters to generate continuous flow results.

2. The method for monitoring the flow of municipal drainage pipelines based on intelligent sensors according to claim 1, characterized in that, Specifically, S1 is: Flow velocity sensors, water level sensors, and pressure sensors are installed on the well wall or inside the well casing of the node well, and flow velocity sensors are installed inside the pipe section. The flow velocity sensor, water level sensor, and pressure sensor are synchronously sampled at a uniform sampling frequency to obtain the flow velocity data, water level data, and pressure data of the corresponding node well and pipe section. The flow velocity data, water level data, and pressure data are time-stamped and matched. The flow velocity data, water level data, and pressure data corresponding to the same time stamp are combined to form the pipe section monitoring data.

3. The method for monitoring the flow of municipal drainage pipelines based on intelligent sensors according to claim 1, characterized in that, The process of obtaining the pipe segment connection relationships, node well locations, pipe diameters, slopes, and pipe bottom elevations of the drainage pipe network, and constructing the pipe network topology, specifically involves: Retrieve the node well number, node well spatial coordinates, and pipe segment number from the pipeline network database; Establish the connection relationship between pipe segments based on the starting point node well and the ending point node well of the pipe segment, and form the connection relationship between node wells; Read the pipe diameter, slope, and bottom elevation of the corresponding pipe segment, and associate them according to the pipe segment number; The topology between node wells and pipe segments is established based on the spatial coordinates of the node wells and the connection relationship of the pipe segments, thus generating the pipeline topology.

4. The method for monitoring the flow of municipal drainage pipelines based on intelligent sensors according to claim 1, characterized in that, The process of mapping pipe segment monitoring data to corresponding pipe segments to generate a pipe network monitoring map specifically involves: The monitoring data of the pipe section is identified and matched according to the node well number and pipe section number, and the flow velocity data, water level data and pressure data are matched with the corresponding pipe section number; The flow velocity, water level, and pressure data at the same time point are aligned according to the time identifier and associated with the corresponding pipe segment number. The flow velocity data, water level data, and pressure data of the corresponding pipe segment number are loaded into the corresponding pipe segment location in the pipe network topology to form a pipe network monitoring map containing pipe segment monitoring data.

5. The method for monitoring the flow of municipal drainage pipelines based on intelligent sensors according to claim 1, characterized in that, Specifically, S3 is: Extract the water level data corresponding to the starting node well and the ending node well of the same pipe segment from the pipeline monitoring map, and calculate the water level difference between the node wells by subtracting the water level data of the ending node well from the water level data of the starting node well. The starting point node well and the ending point node well are determined based on the connection relationship of the pipe segments, and the upstream and downstream relationship of the pipe segments is determined according to the positive and negative correspondence of the slope. The pipe section cross-sectional height is calculated based on the pipe diameter, and the water depth is obtained by calculating the difference between the node well water level data and the pipe bottom elevation. The pipe section filling degree is obtained by calculating the ratio of the water depth to the pipe section cross-sectional height. Hydraulic characteristic data are generated by combining the water level difference, the upstream and downstream relationship of the pipe segment, and the filling degree of the pipe segment according to the pipe segment number.

6. The method for monitoring the flow of municipal drainage pipelines based on intelligent sensors according to claim 1, characterized in that, The improved StemGNN model specifically includes a feature encoding module, a graph construction module, a spectral domain operation module, and a flow output module; The feature encoding module, based on hydraulic feature data, sorts the water level difference, upstream and downstream relationship of the pipe segment, pipe segment fullness, pipe diameter and slope according to the pipe segment number. It arranges the water level difference, upstream and downstream relationship of the pipe segment, pipe segment fullness, pipe diameter and slope corresponding to the same pipe segment number in a fixed order to form a pipe segment feature vector. Then, it arranges the pipe segment feature vectors row by row according to the pipe segment number order to generate a pipe segment feature matrix. The graph construction module establishes an initial adjacency matrix based on the pipeline network topology. The row and column positions of the initial adjacency matrix correspond to the node well numbers, and the connection values ​​of the initial adjacency matrix correspond to the pipe segment connection relationships between the node wells. The graph construction module introduces a liquid vein coherent spectrum mechanism, which determines the water level direction based on the sign of the water level difference, the slope direction based on the sign of the slope, and the pipe segment transmission quantity based on the upstream and downstream relationship of the pipe segment. The water level direction, the slope direction, and the pipe segment transmission quantity are multiplied to generate a propagation direction term. The pipe segment cross-sectional area is calculated based on the pipe diameter, and the pipe segment cross-sectional area is multiplied by the pipe segment filling degree to generate a water flow scale term. The propagation direction term and the water flow scale term are normalized respectively, and the normalized propagation direction term and the normalized water flow scale term are multiplied to generate the fluid flow coherence relation; the fluid flow coherence relation is arranged according to the node well number to generate a coherence matrix, and the coherence matrix is ​​multiplied with the connection value at the corresponding position of the initial adjacency matrix to generate the lead spectrum adjacency matrix. The spectral domain operation module calculates the degree matrix based on the guided spectrum adjacency matrix, subtracts the degree matrix from the guided spectrum adjacency matrix to generate a graph Laplacian matrix, performs eigenvalue decomposition on the graph Laplacian matrix to generate a graph spectral basis, and multiplies the pipe segment feature matrix with the graph spectral basis to generate pipe segment spectral domain features. The flow output module converts the pipe segment spectral domain characteristics into pipe segment flow values ​​and generates initial flow results by arranging them according to the pipe segment number.

7. The method for monitoring the flow of municipal drainage pipelines based on intelligent sensors according to claim 1, characterized in that, Specifically, S5 is: The initial flow rate results and the hydraulic characteristic data are arranged according to the pipe segment number to generate a flow rate data matrix and a hydraulic characteristic matrix; A set of parameter vectors is constructed based on the row vectors of the traffic data matrix, and the mean vector and covariance matrix are calculated to generate the parameter distribution. Multiple sets of parameter vectors are generated based on the parameter distribution, and each set of parameter vectors is calculated in correspondence with the initial flow result to generate multiple sets of flow results; The flow difference corresponding to the water level difference, the flow transmission difference corresponding to the upstream and downstream relationship of the pipe section, and the flow change difference corresponding to the pipe section fullness are calculated based on multiple sets of flow results and hydraulic feature matrix. The flow difference, the flow transmission difference, and the flow change difference are then weighted and summed to generate a consistent value. The multiple sets of parameter vectors are sorted based on the consistency values, and the top N sets of parameter vectors are selected to update the mean vector and covariance matrix, where N is a preset positive integer. The flow correction parameters are generated based on the updated mean vector and the updated covariance matrix.

8. The method for monitoring the flow of municipal drainage pipelines based on intelligent sensors according to claim 1, characterized in that, Specifically, S6 is: Arrange the flow correction parameters according to the pipe segment number and the corresponding initial flow result to generate a flow correction matrix; Calculate the cross-sectional area of ​​the pipe section based on the pipe diameter, and multiply the cross-sectional area of ​​the pipe section by the filling degree of the pipe section to generate the effective water flow area; Multiply the effective water flow area by the water level difference to generate the water flow capacity of the pipe section; multiply the flow correction matrix by the initial flow result to generate the corrected flow result; The flow direction sign of the corrected flow rate result is determined based on the upstream and downstream relationship of the pipe section, and the flow direction sign is multiplied by the water carrying capacity of the pipe section to generate the hydraulic correction amount; The corrected flow rate result is added to the hydraulic correction amount to generate the corrected flow rate for the pipe section.

9. A method for monitoring the flow of municipal drainage pipelines based on intelligent sensors according to claim 1, characterized in that, The step of inputting the corrected flow rate of the pipe segment into the pipe network topology and calculating the flow deviation according to the inflow and outflow rates of the node wells is specifically as follows: In the pipeline network topology, the corrected flow rate of the pipe segment is calibrated according to the pipe segment number, and the connection correspondence between the node well and the pipe segment is established; Using the node well number as an index, the pipe segments connected to the same node well are divided into inflow direction pipe segments and outflow direction pipe segments; The corrected flow rates of the pipe segments corresponding to the inflow direction under the same time identifier are added one by one to generate the node well inflow rate; the corrected flow rates of the pipe segments corresponding to the outflow direction under the same time identifier are added one by one to generate the node well outflow rate. The flow deviation is obtained by subtracting the flow rate from the flow rate at the node well inlet.

10. A method for monitoring the flow of municipal drainage pipelines based on intelligent sensors according to claim 1, characterized in that, If a flow deviation exists, the flow correction parameters are updated to generate continuous flow results, specifically as follows: The flow deviation is allocated to the corresponding pipe segment according to the node well connection relationship to generate the pipe segment deviation amount; Using the pipe segment number as an index, establish the correspondence between the pipe segment deviation and the flow correction parameters, and generate the parameter correction amount; The flow correction parameters are added to the parameter correction values ​​one by one to generate the updated flow correction parameters; The updated flow correction parameters are multiplied one by one with the initial flow result to generate the updated pipe section corrected flow rate; The updated pipeline flow rate is arranged in chronological order based on the updated pipeline segment correction flow rate to generate continuous flow rate results.

Images