Sewage pipe network leakage point location positioning method based on water quality and quantity real-time monitoring data

Abstract

The invention discloses a sewage pipe network leakage point location positioning method based on water quality and quantity real-time monitoring data, and belongs to the technical field of water quality and quantity monitoring. According to the method, the problem of low efficiency of existing manual inspection is solved, and real-time water quality and quantity data are analyzed by constructing the neural network model, so that point location information of possible leakage of the sewage pipe network is accurately predicted; the accuracy and efficiency of leakage detection of the sewage pipe network are improved, the position of leakage of the sewage pipe network can be judged more accurately, and errors caused by human factors in a traditional method are avoided; meanwhile, according to the method, a neural network model is combined with a clustering analysis algorithm, possible leakage point positions can be intelligently predicted, information does not need to be obtained through modes such as manual inspection, the leakage detection process is simplified, and therefore the number of times and cost of leakage repair are reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of water quality and water quantity monitoring, and in particular to a method for locating leakage points in a sewage pipe network based on real-time water quality and water quantity monitoring data. Background Art

[0002] Sewage pipe networks are a vital component of urban infrastructure, significantly impacting a city's environmental quality, quality of life, and economic development. With the accelerating pace of urbanization, sewage pipe networks face numerous challenges and challenges. Currently, leakage in sewage pipe networks is a major contributor to poor urban drainage and water pollution. Rapid and accurate location and timely repair of leakage in sewage pipe networks are crucial.

[0003] In the existing technology, the detection method of sewage pipe network leakage mainly relies on traditional methods such as manual inspection, water level meter, flow meter, etc.; however, in the specific implementation of the above methods, there are still problems such as time-consuming and labor-intensive, low accuracy, and inability to achieve real-time monitoring, which makes it difficult to accurately locate the leakage point of the sewage pipe network.

[0004] Therefore, it does not meet the existing needs. We propose a sewage pipe network leakage point positioning method based on real-time monitoring data of water quality and quantity. Summary of the Invention

[0005] The purpose of the present invention is to provide a method for locating leakage points in a sewage pipe network based on real-time monitoring data of water quality and water quantity. By constructing a neural network model to analyze real-time water quality and water quantity data, the method can accurately predict the point information where leakage may occur in the sewage pipe network. This not only improves the accuracy and efficiency of sewage pipe network leakage detection, but also enables it to more accurately determine the location of sewage pipe network leakage, avoiding errors caused by human factors in traditional methods. At the same time, the method automatically calculates possible leakage points by combining a neural network model with a cluster analysis algorithm, eliminating the need to obtain information through manual inspections, simplifying the leakage detection process, thereby reducing the number and cost of leakage repairs, and solving the problems raised in the above-mentioned background technology.

[0006] To achieve the above object, the present invention provides the following technical solutions:

[0007] A method for locating leakage points in a sewage pipe network based on real-time monitoring data of water quality and quantity, comprising the following steps:

[0008] Step 1: Define the target area, collect historical water quality and quantity data of the sewage pipe network in the target area and information on leakage points, clean and organize the collected information, and then use the organized information as a data sample;

[0009] Step 2: Construct a neural network model and use it to extract characteristic information about water quality and quantity from the data sample. Analyze historical water quality and quantity data using the neural network model, compare historical water quality and quantity data over different time periods, and identify patterns of change in historical water quality and quantity.

[0010] Step 3: Construct the drainage system structure within the target area and train it using a neural network model. Based on the historical changes in water quality and quantity, and the characteristics of the drainage system structure within the target area, predict points where sewage pipe network leakage may occur. The predicted results are then compared with the actual sewage pipe network leakage points to determine the accuracy of the neural network model prediction results.

[0011] Step 4: Set up multiple monitoring points in the target sewage network to monitor water quality and quantity in real time to obtain real-time water quality and quantity data;

[0012] Step 5: Use the real-time water quality and quantity data as the input set of the neural network model, use the neural network model to analyze the real-time water quality and quantity data, and output the change pattern of the real-time water quality and quantity; then combine the characteristics of the drainage system structure to predict the points where sewage pipe network leakage may occur; and feed back the prediction results to the client for early warning, so that the points can be inspected and repaired in time.

[0013] Furthermore, in step 1, the collected information is cleaned and organized, specifically including the following steps:

[0014] Remove duplicate and erroneous data from historical water quality and quantity data, remove damaged and incomplete data, handle outliers, and fill in missing data;

[0015] Adjust the format and structure of historical water quality and quantity data to make them conform to the input requirements of the neural network model;

[0016] The historical water quality and quantity data after cleaning are then sorted out, which includes: classifying the data categories, dividing the data according to time series, and setting output labels for each leakage point information.

[0017] Furthermore, in step 2, a neural network model is constructed, and characteristic information about water quality and quantity in the data sample is extracted using the neural network model, which specifically includes the following steps:

[0018] Construct a neural network model consisting of an input layer, a hidden layer, and an output layer. The input layer includes all feature variables, the number of neurons in the hidden layer and the output layer is adaptively adjusted based on the amount of data and the learning rate, and the output layer is used to output information about possible locations of sewage pipe network leakage.

[0019] Divide the data samples into training set and test set, and normalize the training set, test set and output labels;

[0020] Input the training set data into the neural network model for parameter training to obtain the trained neural network model;

[0021] The trained neural network model is then applied to the test set to extract characteristic information about water quality and quantity, and the extracted characteristic information is combined to generate analysis results.

[0022] Furthermore, in step 2, the historical water quality and quantity data are analyzed by a neural network model, and the historical water quality and quantity data in different time periods are compared to find out the changing pattern of the historical water quality and quantity. Specifically, the following steps are included:

[0023] After extracting the characteristic information of historical water quality and quantity data through the neural network model, the water quality and quantity data in different time periods are input into the neural network model for comparative analysis. By calculating and comparing the differences in historical water quality and quantity data in different time periods, the changing patterns of historical water quality and quantity are obtained. Among them, the differences in historical water quality and quantity data are calculated by using the law of conservation of mass in fluid mechanics. The calculation formula is as follows:

[0024] m1+m2=m3+m4;

[0025] Among them, m1 and m2 represent the mass of the substance before the reaction; m3 and m4 represent the mass of the substance produced after the reaction.

[0026] Furthermore, in step three, based on the change pattern, the points where there may be leakage in the sewage pipe network are calculated, which specifically includes the following steps:

[0027] After obtaining the historical change patterns of water quality and quantity, the drainage system structure is constructed according to the structure and characteristics of the urban drainage system and geographical characteristics;

[0028] The drainage system structure is combined with the collected historical water quality and quantity data and input into the neural network model for training to learn the potential relationship between the historical water quality and quantity data and the drainage system structure;

[0029] Based on the characteristics of the drainage system structure in the target area, predict the locations where sewage pipe network leakage may occur;

[0030] Compare the predicted location with the actual leakage point information in the historical water quality and quantity data to determine whether the prediction result of the neural network model is accurate; if it is accurate, mark the leakage point;

[0031] Among them, the accuracy of the neural network model can be calculated by the following formula:

[0032]

[0033] Among them, y obs，i is the (i)th actual observation value; y prde，i is the predicted value of the (i)th model; is the average of all actual observations; is the sum of all samples.

[0034] Furthermore, after predicting the location where the sewage pipe network may leak, the following steps are included:

[0035] The changing patterns of historical water quality and quantity corresponding to the possible leakage points in the data sample are converted into a format suitable for cluster analysis, and the cluster analysis algorithm is used to analyze the relationship between the changing patterns of historical water quality and quantity data and the possible leakage points;

[0036] Input the data sample into the clustering algorithm to obtain the probability that each sample belongs to a different cluster; based on this probability, divide the data sample into K clusters;

[0037] Analyze the obtained K clusters to see if there is any correlation between the K clusters;

[0038] According to the analysis results, it is speculated which of the K clusters corresponds to water quality changes caused by pipe network leakage;

[0039] Further analysis is then performed on this cluster to check the geographical location and pipeline direction corresponding to this cluster and identify possible leakage points.

[0040] Furthermore, in step 4, the water quality and quantity are monitored in real time to obtain real-time water quality and quantity data, which specifically includes the following steps:

[0041] Multiple monitoring points are deployed based on the characteristics of the target sewage network; water quality sensors and water quantity sensors are installed at each monitoring point to collect real-time water quality and quantity data;

[0042] According to actual needs, a timed or real-time collection mechanism is set for the water quality sensors and water quantity sensors in each monitoring point. According to the preset collection time and frequency, water quality and quantity data are collected regularly or in real time, and fed back to the neural network model in a timely manner for prediction.

[0043] Furthermore, in step five, a neural network model is used to analyze real-time water quality and quantity data, and combined with the characteristics of the drainage system structure, the points where sewage pipe network leakage may occur are predicted, which specifically includes the following steps:

[0044] Use the trained model to analyze the real-time water quality and quantity data to obtain the changing patterns of real-time water quality and quantity;

[0045] Analyze the changing patterns of water quality and quantity in real time based on the characteristics of the drainage system structure, and predict possible leakage points in the sewage pipe network;

[0046] The relationship between the prediction results and the corresponding water quality and quantity change patterns is analyzed using the cluster analysis algorithm to infer whether the water quality change at the point is caused by pipe network leakage;

[0047] And check the geographical location and pipeline direction corresponding to the leakage point, find out the existing leakage point, and obtain the prediction results of the neural network model.

[0048] Furthermore, in step five, the information is fed back to the client for early warning, and the point is promptly inspected and repaired, which specifically includes the following steps:

[0049] Feedback the prediction results of the neural network model to the client, reminding the client to check for leakage points and take timely measures;

[0050] And provide appropriate suggestions to the client based on the location and condition of the leakage point.

[0051] Furthermore, after receiving feedback information about the leakage point, the client checks the specific location of the leakage point to determine whether the prediction result of the neural network model is accurate;

[0052] If the prediction result is accurate, the point will be inspected immediately; if the prediction result is accurate, it will be immediately fed back to the neural network model and the prediction will be made again.

[0053] Compared with the prior art, the present invention has the following beneficial effects:

[0054] The present invention analyzes real-time water quality and quantity data by constructing a neural network model, thereby accurately predicting the point information of possible leakage in the sewage pipe network; it not only improves the accuracy and efficiency of sewage pipe network leakage detection, making it possible to more accurately determine the location of sewage pipe network leakage, and avoids errors caused by human factors in traditional methods; at the same time, the method combines the neural network model with the cluster analysis algorithm, and can intelligently predict possible leakage points, eliminating the need to obtain information through manual inspections and the like, simplifying the leakage detection process, and thus reducing the number and cost of leakage repairs. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] Figure 1 This is a flow chart of a method for locating leakage points in a sewage pipe network based on real-time monitoring data of water quality and quantity according to the present invention;

[0056] Figure 2This is a logic diagram of a method for locating leakage points in a sewage pipe network based on real-time monitoring data of water quality and quantity according to the present invention;

[0057] Figure 3 It is the prediction effect diagram of the present invention;

[0058] Figure 4 This is a result diagram of the present invention under heavy rain weather. DETAILED DESCRIPTION

[0059] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0060] In order to solve the technical problems of the existing technology that the detection method of pipe network leakage mainly relies on manual inspection, water level meter, flow meter and other traditional methods, which are time-consuming and labor-intensive, with low accuracy and inability to achieve real-time monitoring, please refer to Figures 1-4 , this embodiment provides the following technical solutions:

[0061] A method for locating leakage points in a sewage pipe network based on real-time monitoring data of water quality and quantity, comprising the following steps:

[0062] Step 1: Define the target area, collect historical water quality and quantity data and leakage location information of the sewage network in the target area, clean and organize the collected information, and then use the organized information as a data sample to cover various situations and patterns. The specific steps include:

[0063] Since the collected historical water quality and quantity data and the information on the points where leakage occurred usually contain a large number of missing values or outliers, during the cleaning and processing stage of the historical water quality and quantity data, it is usually necessary to remove duplicate data and erroneous data, remove damaged and incomplete data, process outliers, and fill in missing data to make the remaining data as complete and accurate as possible; adjust the format and structure of the historical water quality and quantity data so that the format of the historical water quality and quantity data meets the input requirements of the neural network model; and then organize the cleaned historical water quality and quantity data, which includes: classifying the data categories, dividing the data according to time series, and setting output labels for each point information where leakage occurred, so that the data samples can be smoothly accepted and processed by the neural network model.

[0064] Step 2: Construct a neural network model and use it to extract characteristic information about water quality and quantity from the data sample. Analyze historical water quality and quantity data using the neural network model, compare historical water quality and quantity data over different time periods, and identify the changing patterns of historical water quality and quantity. This specifically includes the following steps:

[0065] A neural network model is constructed, including an input layer, a hidden layer, and an output layer, wherein the input layer includes all feature variables, the number of neurons in the hidden layer and the output layer is adaptively adjusted according to the amount of data and the learning rate, and the output layer is used to output information on points where sewage pipe network leakage may occur; the data samples are divided into a training set and a test set, such as the first N data of each month as the training set and the last N data as the test set, and the training set, the test set, and the output label are normalized to reduce noise and improve the generalization ability of the model; the training set data is input into the neural network model for parameter training, and the loss function is minimized through the back propagation algorithm to obtain the optimal model parameters and the trained neural network model at the same time; the trained neural network model is then applied to the test set to extract characteristic information about water quality and quantity, and the extracted characteristic information is then combined to generate a more comprehensive analysis result.

[0066] Secondly, after extracting the characteristic information of historical water quality and quantity data through the neural network model, the water quality and quantity data in different time periods are input into the neural network model for comparative analysis. By calculating and comparing the differences in historical water quality and quantity data in different time periods, the changing patterns of historical water quality and quantity are obtained. Among them, the differences in historical water quality and quantity data are calculated by using the law of conservation of mass in fluid mechanics. The calculation formula is as follows:

[0067] m1+m2=m3+m4;

[0068] Among them, m1 and m2 represent the mass of the substance before the reaction; m3 and m4 represent the mass of the substance produced after the reaction;

[0069] In this example, suppose there is a river or water body containing two substances: A and B. At the initial moment, the masses of A and B are m and m respectively. A and m B ; After a period of change, the masses of A and B are m' A and m' B ; Based on this, the above information is used to describe the changing patterns of water quality and water quantity.

[0070] As we all know, according to the law of conservation of mass, at any moment, the total mass of the entire system remains unchanged; in other words, the total mass of A is equal to the total mass of B, that is: m A =m B; Therefore, this equation can be used to express the material balance relationship at the initial moment.

[0071] As time goes by, A and B will undergo a certain transformation; suppose that after a certain period of time t, the mass of A becomes m' A , the mass of B becomes m' B Based on this, a new material balance equation can be obtained at this time point, namely:

[0072] m' A =m A +(m B -m')(1-r) / / A undergoes a certain proportion of transformation and changes into B;

[0073] m' B =m B -(m A -m B (1-r) / / B undergoes a certain proportion of transformation and changes to A

[0074] Where r represents the rate at which A is converted to B;

[0075] The above mathematical expressions can be used to calculate the historical water quality and quantity data and summarize the changing patterns of water quality and quantity.

[0076] Step 3: Construct the drainage system structure within the target area and use the neural network model to train the drainage system structure. Based on the historical changes in water quality and quantity, and combined with the characteristics of the drainage system structure in the target area, predict the points where sewage pipe network leakage may occur. The predicted results are then compared with the actual sewage pipe network leakage points to determine the accuracy of the neural network model prediction results. The specific steps include:

[0077] After obtaining the historical patterns of change in water quality and quantity, the drainage system structure is constructed based on the structure and characteristics of the urban drainage system and geographical features, such as topography, population distribution, and traffic conditions. The drainage system structure is combined with the collected historical water quality and quantity data and input into a neural network model for training to learn the potential relationship between the historical water quality and quantity data and the drainage system structure. Based on the characteristics of the drainage system structure in the target area, the location of possible sewage pipe network leakage is predicted. The predicted location is compared with the actual leakage point information in the historical water quality and quantity data to determine whether the prediction result of the neural network model is accurate. If it is accurate, the leakage point is marked.

[0078] Among them, the accuracy of the neural network model can be calculated by the following formula:

[0079]

[0080] Among them, yobs，i is the (i)th actual observation value; y prde，i is the predicted value of the (i)th model; is the average of all actual observations; is the sum of all samples.

[0081] After predicting the possible locations of sewage pipe network leakage, the historical water quality and quantity change patterns corresponding to the possible leakage points in the data sample are converted into a format suitable for cluster analysis, such as: the possible leakage points and the corresponding historical water quality and quantity change patterns are divided into K clusters, and a cluster analysis algorithm is used to analyze the relationship between the change patterns of historical water quality and quantity data and the possible leakage points; the data samples are input into the clustering algorithm to obtain the probability of each sample belonging to a different cluster; based on the probability, the data samples are divided into K clusters; the obtained K clusters are analyzed to analyze whether there is any correlation between the K clusters; for example: if the characteristics of some clusters are obviously different from those of other clusters, it may indicate that the cluster represents certain specific water quality change patterns; based on the analysis results, it is inferred which part of the K clusters corresponds to water quality changes caused by pipe network leakage; further analysis is then performed on these clusters to check the geographical location and pipe network direction corresponding to these clusters to find possible leakage points; and the leakage points found are marked so that the data samples can be more accurately understood by the neural network model, thereby optimizing the neural network model.

[0082] Step 4: Set up multiple monitoring points in the target sewage network to monitor water quality and quantity in real time to obtain real-time water quality and quantity data; specifically, the following steps are included:

[0083] According to the characteristics of the target sewage network, multiple monitoring points are set up; and water quality sensors and water quantity sensors are installed at each monitoring point to collect real-time water quality and quantity data; in the specific implementation, the selection of monitoring points must fully consider the characteristics of the sewage network, such as factors such as pipe length, diameter, and pipe section location, so as to facilitate the layout of appropriate monitoring points in the target sewage network; for example, the main intersections and important nodes of the network can be selected for monitoring to ensure that the entire process of the network is covered; at the same time, the monitoring points can also be reasonably set up in combination with the characteristics of the network, such as length, diameter, etc.; for example, more monitoring points can be set up for long networks to improve the accuracy of monitoring; for networks with larger diameters, monitoring points can be set up inside them to ensure that the water quality and quantity in the pipes are monitored in real time.

[0084] Secondly, according to actual needs, a timed or real-time collection mechanism is set for the water quality sensors and water quantity sensors in each monitoring point. According to the preset collection time and frequency, water quality and quantity data are collected regularly or in real time, and promptly fed back to the neural network model for prediction. In the specific implementation, by collecting real-time water quality and quantity data within a certain time range, the drainage system structural characteristics data at the corresponding time are also collected, such as: pipe material, pipe diameter and valve position, etc., and this part of the data is integrated into a unified data framework as the input set of the neural network model.

[0085] Step 5: Use the real-time water quality and quantity data as the input set of the neural network model, use the neural network model to analyze the real-time water quality and quantity data, and output the change pattern of the real-time water quality and quantity; then combine the characteristics of the drainage system structure to predict the points where sewage pipe network leakage may occur; and feed back the prediction results to the client for early warning, so that the points can be repaired in time. The specific steps include:

[0086] The trained model is used to analyze the real-time water quality and quantity data to obtain the changing patterns of real-time water quality and quantity; the changing patterns of real-time water quality and quantity are analyzed in combination with the characteristics of the drainage system structure to predict possible sewage pipe network leakage points; the relationship between the predicted results and the corresponding changing patterns of water quality and quantity are analyzed based on the cluster analysis algorithm to infer whether the water quality change at the point is caused by pipe network leakage; and the geographical location and pipe network direction corresponding to the leakage point are checked to find the existing leakage point and obtain the prediction results of the neural network model.

[0087] See also Figure 2 In one embodiment, a neural network model is constructed as a water quality and quantity dynamic model, and the implementation process of the water quality and quantity dynamic model is divided into a global scale model, a local segment model, and leakage point location simulation training;

[0088] Global scale model: First, a dynamic model of water quality and quantity is established to simulate the flow rate and analyze the actual liquid level to determine the severity of the leakage. If it is not serious, the process is terminated; if it is serious, it is continued.

[0089] Local segment model: Analyze based on the statistical model of water quality changes, and judge whether it is a minor leakage or a serious leakage based on the analysis results.

[0090] Leakage location simulation training: Through pipe network generalization and SWMM hydrological simulation software analysis, Matlab training sample sets are used to input the upstream pipeline level and flow into the hydraulic model. Simulation data sets are set for different leak locations and sizes. Based on online monitoring of liquid level and flow fitting and inversion, the location of the leaking pipe section and the average length and height of the leaked object are determined, and the final results are obtained.

[0091] See also Figure 3 : Liquid level data (black lines): The figure shows the liquid level changes of four monitoring points (monitoring points 1, 2, 3, and 4). The liquid level data are expressed as lines of different grayscale colors in the figure, reflecting the liquid level height of each monitoring point at different time points.

[0092] Rainfall data (gray bar graph): The gray bar graph in the figure represents rainfall, showing the rainfall situation during the same time period. The rainfall data is displayed as vertical bar graphs, reflecting the amount of rainfall at each point in time.

[0093] Liquid level data is present when rainfall is zero: This indicates that even when there is no rainfall, the monitoring point still has liquid level data. This may be due to other reasons causing the liquid level change. Based on this, the figure shows the relationship between liquid level and rainfall at different monitoring points during a specific time period. By analyzing this data, we can help identify and solve problems in urban drainage systems.

[0094] Leakage caused by external water: In certain periods of time, the mixing of sewage and rainwater due to rainfall may cause leakage in the pipeline.

[0095] See also Figure 4 : It is a time series graph that shows the liquid level and rainfall data at different monitoring points within a specific time period, as well as the situation of pipe network leakage and rainwater and sewage mixing.

[0096] Liquid level data (black lines): Liquid level data is shown in the figure as lines of different grayscale colors, reflecting the liquid level height of each monitoring point at different time points.

[0097] Rainfall data (gray bar graph): The gray bar graph in the figure represents rainfall, showing the rainfall situation during the same time period. The rainfall data is displayed as vertical bar graphs, reflecting the amount of rainfall at each point in time.

[0098] Leakage in mixed rainwater and sewage pipe sections during heavy rain weather: This means that during heavy rain weather, during certain time periods, the mixing of rainwater and sewage may cause the mixing of sewage and rainwater, which may cause pipeline leakage.

[0099] Anomalies: The graph shows "Anomaly 1" and "Anomaly 2." These points may indicate unusual changes in liquid level data, requiring further analysis and investigation. This graph illustrates the relationship between liquid level and rainfall at different monitoring points over a specific time period, as well as possible pipe network leakage and rainwater-sewage mixing. Analyzing this data can help identify and resolve issues within urban drainage systems.

[0100] Secondly, the prediction results of the neural network model are fed back to the client, reminding the client to check for the presence of leakage points and take timely measures; and according to the location and condition of the leakage points, appropriate suggestions are provided to the client; for example: First, pipeline repair: after the location of the leakage point is detected, the pipeline in the vicinity of the leakage point is first cut, a larger-sized pipeline bracket is installed at the location, and the cut pipeline is then connected and bound; and cement or other solidifying materials are wrapped around the coil to ensure the firmness of the pipeline connection and prevent the leakage point from reappearing. Second, pipe section replacement: when the pipeline has a large leak or serious aging problem, the section of pipeline and its nearby accessories are first disassembled, and then a new galvanized pipe is replaced, and the new pipe fittings are connected to the surrounding pipes to help the client take timely action to avoid or reduce the impact of leakage on the environment and human health. Secondly, after receiving the feedback information of the leakage point, the client checks the specific location of the leakage point to determine whether the prediction result of the neural network model is accurate; if the prediction result is accurate, the point will be immediately inspected; if the prediction result is accurate, it will be immediately fed back to the neural network model and re-prediction will be performed.

[0101] The technical effect of the above technical solution is: through the above operation, not only the accuracy and efficiency of sewage pipe network leakage detection are improved, it can also make it possible to more accurately determine the location of sewage pipe network leakage, avoiding the errors caused by human factors in traditional methods; at the same time, this method combines the neural network model with the clustering analysis algorithm, and can intelligently predict possible leakage points. There is no need to obtain information through manual inspections, etc., which simplifies the leakage detection process and reduces the number and cost of leakage repairs.

[0102] Working principle: The neural network model is trained and tested using historical water quality and quantity data and information on leakage points to ensure the accuracy of the neural network model's prediction ability; multiple monitoring points are set up in the target sewage network to obtain real-time water quality and quantity data; the real-time water quality and quantity data are analyzed using the neural network model, and combined with the characteristics of the drainage system structure, points where sewage network leakage may occur are predicted; and the prediction results are fed back to the client for early warning, so that the points can be inspected and repaired in time.

[0103] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and alterations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. A method for locating leakage points in a sewage pipe network based on real-time water quality and quantity monitoring data, comprising the following steps: Step 1: Collect historical water quality and quantity data and leakage point information of the sewage pipe network in the target area, clean and organize the collected information, and then use the organized information as a data sample; Step 2: Build a neural network model and use it to extract characteristic information about water quality and quantity from data samples; and use the neural network model to analyze historical water quality and quantity data to find out the changing patterns of historical water quality and quantity; Step 3: Construct the drainage system structure within the target area. The neural network model uses the historical variation patterns of water quality and quantity, combined with the characteristics of the drainage system structure within the target area, to predict points where sewage pipe network leakage may occur. The predicted results are then compared with the actual sewage pipe network leakage points to determine the accuracy of the neural network model prediction results. Step 4: Set up multiple monitoring points in the target sewage network to monitor water quality and quantity in real time to obtain real-time water quality and quantity data; Step 5: Use the real-time water quality and quantity data as the input set of the neural network model, use the neural network model to analyze the real-time water quality and quantity data, and combine the characteristics of the drainage system structure to predict the points where sewage pipe network leakage may occur; then feed back the prediction results to the client for early warning and timely inspection and maintenance of the points.

2. A method for locating leakage points in a sewage pipe network based on real-time water quality and quantity monitoring data according to claim 1, characterized in that: In step 1, the collected information is cleaned and organized, which specifically includes the following steps: Remove duplicate and erroneous data from historical water quality and quantity data, remove damaged and incomplete data, handle outliers, and fill in missing data; The format and structure of historical water quality and quantity data are adjusted, and then the cleaned historical water quality and quantity data are sorted. The sorting includes: classifying the data categories, dividing the data according to time series, and setting output labels for each leakage point information.

3. The method for locating leakage points in a sewage pipe network based on real-time water quality and quantity monitoring data according to claim 1, characterized in that: In the second step, a neural network model is constructed and used to extract characteristic information about water quality and quantity from the data sample, specifically including the following steps: Construct a neural network model consisting of an input layer, a hidden layer, and an output layer. The input layer includes all feature variables, the number of neurons in the hidden layer and the output layer is adaptively adjusted based on the amount of data and the learning rate, and the output layer is used to output information about possible locations of sewage pipe network leakage. Divide the data samples into training set and test set, and normalize the training set, test set and output labels; Input the training set data into the neural network model for parameter training to obtain the trained neural network model; The trained neural network model is then applied to the test set to extract characteristic information about water quality and quantity, and the extracted characteristic information is combined to generate analysis results.

4. The method for locating leakage points in a sewage pipe network based on real-time water quality and quantity monitoring data according to claim 1, characterized in that: In the second step, the historical water quality and quantity data are analyzed by a neural network model, and the historical water quality and quantity data in different time periods are compared to find out the changing pattern of the historical water quality and quantity. Specifically, the following steps are included: After extracting the characteristic information of historical water quality and quantity data through the neural network model, the water quality and quantity data in different time periods are input into the neural network model for comparative analysis. By calculating and comparing the differences in historical water quality and quantity data in different time periods, the changing patterns of historical water quality and quantity are obtained. Among them, the differences in historical water quality and quantity data are calculated by using the law of conservation of mass in fluid mechanics. The calculation formula is as follows: m1+m2=m3+m4; Among them, m1 and m2 represent the mass of the substance before the reaction; m3 and m4 represent the mass of the substance produced after the reaction.

5. The method for locating leakage points in a sewage pipe network based on real-time water quality and quantity monitoring data according to claim 1, characterized in that: In the third step, based on the change pattern, the points where there may be leakage in the sewage pipe network are calculated, which specifically includes the following steps: After obtaining the historical change patterns of water quality and quantity, the drainage system structure is constructed according to the structure and characteristics of the urban drainage system and geographical characteristics; The drainage system structure is combined with the collected historical water quality and quantity data and input into the neural network model for training to learn the potential relationship between the historical water quality and quantity data and the drainage system structure; Based on the characteristics of the drainage system structure in the target area, predict the locations where sewage pipe network leakage may occur; Compare the predicted location with the actual leakage point information in the historical water quality and quantity data to determine whether the prediction result of the neural network model is accurate; if it is accurate, mark the leakage point; Among them, the accuracy of the neural network model can be calculated by the following formula: Among them, y obs，i is the (i)th actual observation value; y prde，i is the predicted value of the (i)th model; is the average of all actual observations; is the sum of all samples.

6. The method for locating leakage points in a sewage pipe network based on real-time water quality and quantity monitoring data according to claim 1, characterized in that: After predicting the possible location of sewage pipe network leakage, the following steps are included: The changing patterns of historical water quality and quantity corresponding to the possible leakage points in the data sample are converted into a format suitable for cluster analysis, and the cluster analysis algorithm is used to analyze the relationship between the changing patterns of historical water quality and quantity data and the possible leakage points; Input the data sample into the clustering algorithm to obtain the probability that each sample belongs to a different cluster; based on this probability, divide the data sample into K clusters; Analyze the obtained K clusters to see if there is any correlation between the K clusters; According to the analysis results, it is speculated which of the K clusters corresponds to water quality changes caused by pipe network leakage; Further analysis is then performed on this cluster to check the geographical location and pipeline direction corresponding to this cluster and identify possible leakage points.

7. The method for locating leakage points in a sewage pipe network based on real-time water quality and quantity monitoring data according to claim 1, characterized in that: In the fourth step, the water quality and quantity are monitored in real time to obtain real-time water quality and quantity data, which specifically includes the following steps: Multiple monitoring points are deployed based on the characteristics of the target sewage network; water quality sensors and water quantity sensors are installed at each monitoring point to collect real-time water quality and quantity data; According to actual needs, a timed or real-time collection mechanism is set for the water quality sensors and water quantity sensors in each monitoring point. According to the preset collection time and frequency, water quality and quantity data are collected regularly or in real time, and fed back to the neural network model in a timely manner for prediction.

8. The method for locating leakage points in a sewage pipe network based on real-time water quality and quantity monitoring data according to claim 1, characterized in that: In step 5, the real-time water quality and quantity data are analyzed using a neural network model, and combined with the characteristics of the drainage system structure, the points where sewage pipe network leakage may occur are predicted. Specifically, the following steps are included: Use the trained model to analyze the real-time water quality and quantity data to obtain the changing patterns of real-time water quality and quantity; Analyze the changing patterns of water quality and quantity in real time based on the characteristics of the drainage system structure, and predict possible leakage points in the sewage pipe network; The relationship between the prediction results and the corresponding water quality and quantity change patterns is analyzed using the cluster analysis algorithm to infer the reasons for the changes in water quality and quantity at the point; Check the geographical location and pipeline direction corresponding to the leakage point, find the leakage point, and obtain the prediction results of the neural network model.

9. The method for locating leakage points in a sewage pipe network based on real-time water quality and quantity monitoring data according to claim 1, characterized in that: In step 5, the information is fed back to the client for early warning, and the point is repaired in time, which specifically includes the following steps: Feedback the prediction results of the neural network model to the client, reminding the client to check for leakage points and take timely measures; And provide appropriate suggestions to the client based on the location and condition of the leakage point.

10. The method for locating leakage points in a sewage pipe network based on real-time water quality and quantity monitoring data according to claim 1, characterized in that: After receiving feedback on the leakage point, the client checks the specific location of the leakage point to determine whether the prediction result of the neural network model is accurate; If the prediction result is accurate, the point will be inspected immediately; if the prediction result is accurate, it will be immediately fed back to the neural network model and the prediction will be made again.

Images