A water supply pipe network pollution source positioning method based on dynamic optimization sampling of water quality
By optimizing the location of sampling points and dynamically updating laboratory test results, the method of identifying suspected pipelines has solved the problems of high cost and low efficiency in locating pollution sources in water supply networks, achieving low-cost and high-efficiency pollution source location.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2021-12-31
- Publication Date
- 2026-03-27
AI Technical Summary
In existing technologies, methods for locating pollution sources in water supply networks rely on high-cost online water quality sensors and complex water quality models, which makes practical applications difficult and results in low efficiency in selecting sampling points, making it difficult to quickly and accurately locate pollution sources.
By optimizing the selection of sampling point locations in each cycle, the pipeline network is divided into multiple sub-regions. Suspected pipelines are dynamically updated using laboratory test results. By minimizing the objective function to select sampling points, the range of pollution sources is gradually narrowed down, achieving efficient location of pollution sources.
It enables low-cost, rapid, and accurate location of pollution sources in water supply networks, avoiding sensor installation and maintenance costs, and is independent of water quality models, thus improving location efficiency and reliability.
Smart Images
Figure CN114459810B_ABST
Abstract
Description
TECHNICAL FIELD
[0001] The present application belongs to the field of urban water supply, and particularly relates to a pollution source positioning technology in an urban water supply network. BACKGROUND
[0002] Water quality guarantee in a water supply network is an important work. The water quality guarantee work in the water supply network can be generally divided into three parts, including: (1) water quality risk monitoring of the network, (2) positioning of the pollution source, and (3) control and removal of the found pollution source. Among them, the positioning of the pollution source refers to that when a pollution event occurs in the network, various measurement and simulation means are used to narrow down the range of the pollution source as much as possible, so as to facilitate the subsequent pollution source control and removal work. The commonly used pollution source positioning method is to use the data of the online water quality sensor installed in the network, couple a water quality model, and use various simulation fitting means to speculate the position of the pollution source. However, at present, due to the high installation and maintenance cost of the sensor, the limited types of water quality indicators that can be monitored by the sensor technology, and the difficulty in obtaining an accurate water quality model, this method of collecting data by using the online water quality sensor often exists in the theoretical derivation and laboratory verification stage, and lacks the conditions for large-scale practical application.
[0003] In addition to installing online sensors, water quality sampling is another method for measuring water quality in a water supply network. Water quality sampling refers to a method of manually taking a water sample at a certain point in the network and bringing it back to the laboratory for detection. Through laboratory detection, detailed water quality indicators of the water sample can be obtained, and the detection results of the sample can fully reflect the water quality of the sampling point and the upstream pipeline. The main factor affecting the efficiency of the pollution source positioning method of water quality sampling is the selection of the sampling point position; there are many points in the network that can be sampled, but the number of samples that can be simultaneously sampled and detected is limited due to the detection capacity of the laboratory and the personnel arrangement; on the other hand, the selection of the sampling point needs to consider the detection results of the detected samples. This means that the position of the sampling point needs to be optimized and dynamically selected. SUMMARY
[0004] The technical problem to be solved by the present application is to provide a water quality dynamic optimization sampling based pollution source positioning method for a water supply network, which optimizes the selection of the sampling point position in each cycle, thereby improving the efficiency of positioning the pollution source.
[0005] The technical scheme adopted by the present application is as follows:
[0006] A water quality dynamic optimization sampling based pollution source positioning method for a water supply network, the steps of which are as follows:
[0007] S11: Determine the sampling points in the pipe network according to the actual situation. For example, for a pipe network, water samples can be conveniently taken at fire hydrants, and samples can also be taken from residents' homes or other places with faucets.
[0008] Determine the number of sampling points in each cycle n , n The pipe network is divided into T sub-regions by K sampling points:
[0009] T=2 n 1-1,
[0010] wherein the sub-region is a set of pipes having the same upstream and downstream relationship with a group of sampling points. For example, when K = 1, the pipes upstream of the sampling point in the pipe network collectively form a sub-region, and the remaining pipes not upstream of the sampling point form another sub-region. When K = 2, the pipe network is divided into four sub-regions: a set of pipes upstream of both sampling points, a set of pipes upstream of only the first sampling point, a set of pipes upstream of only the second sampling point, and a set of pipes not upstream of both sampling points. Similarly, n K sampling points divide the pipe network into n sub-regions. n T
[0011] S12: All pipes in the pipe network are marked as suspicious pipes.
[0012] S21: Select sampling points by minimizing the objective function F(A), denoted as sampling group A, sample and detect.
[0013] 1-2,
[0014] wherein, L all is the length sum of all suspicious pipes at present, L i is the length sum of suspicious pipes in the Kth sub-region. i
[0015] After obtaining the sample, water quality detection is needed to obtain the detection result of the water sample. A suitable method is to send it back to the laboratory for detection, because the current on-site detection technology is difficult to obtain complete water quality information and will lead to inaccurate detection results. The purpose of water quality detection is to obtain the detection result of whether the sample is contaminated or not. The detection method itself is not the innovation point of the present application and will not be described here.
[0016] S22: update the suspicious pipeline according to the sample detection result, the specific updating method is: (1) if all samples are contaminated: the suspicious pipeline is updated to the intersection of the upstream pipelines of the sampling points; (2) if some samples are contaminated: the suspicious pipeline is updated to the intersection of the upstream pipelines of the contaminated sampling points minus the union of the upstream pipelines of the uncontaminated sampling points, that is, the suspicious pipeline Q = the intersection of the upstream pipelines of the contaminated sampling points Q1' minus the union of the upstream pipelines of the uncontaminated sampling points Q2; (3) if all samples are uncontaminated: the suspicious pipeline is updated to the complement set of the union of the upstream pipelines of the sampling points, where the suspicious pipeline refers to the set of suspicious pipelines, for example, the suspicious pipeline Q = {pipeline 1, pipeline 3, pipeline 4}.
[0017] S31: if the number of contaminated samples in S21 is less than two, skip S3.
[0018] S32: if the updated suspicious pipeline in S22 is empty, update the suspicious pipeline to the union of the upstream pipelines of the contaminated sampling points minus the union of the upstream pipelines of the uncontaminated sampling points, that is, the suspicious pipeline Q = the union of the upstream pipelines of the contaminated sampling points Q1 minus the union of the upstream pipelines of the uncontaminated sampling points Q2, and skip S3.
[0019] S33: determine one sampling point position as the end of the current suspicious pipeline, and the remaining n -1 sampling points are determined according to the method of S21, and sampling and detection are performed on the determined n sampling point positions.
[0020] S34: if the detection result of the suspicious pipeline end sampling point in S33 is uncontaminated, update the suspicious pipeline to the union of the upstream pipelines of the contaminated sampling points minus the union of the upstream pipelines of the uncontaminated sampling points minus the upstream pipeline of the suspicious pipeline end sampling point in S33 in S21, that is, the suspicious pipeline Q = the union of the upstream pipelines of the contaminated sampling points Q1 minus the union of the upstream pipelines of the uncontaminated sampling points Q2 minus the upstream pipeline of the suspicious pipeline end sampling point Q3; otherwise, update the suspicious pipeline according to the method in S22.
[0021] In order to solve the problems in the background art, the application proposes a water quality dynamic optimization sampling water supply network pollution source positioning method, which does not rely on water quality model and online sensor data, but through water quality sampling and laboratory detection to position the pollution source. In the method, the water supply network is divided into several sub-regions according to the selected sampling point position and the current network flow, and the sampling group optimizes the selection through the proposed objective function to minimize the sum of squares of the lengths of each sub-region. According to the detection results of the samples, the pollution source can be locked into one of the sub-regions, and the range of the sub-region where the pollution source is located (suspected pipeline) is continuously narrowed through iterative cycles, until there is no hydrant (sampling point) in the suspected pipeline that can continue to be sampled, and the water quality dynamic optimization sampling is realized through the above method. The application has originality, proposes a new water supply network pollution source positioning method, has a wide application scenario, and has great significance for the safety guarantee work in the field of urban water supply network.
[0022] Compared with the prior art, the application has the following advantages:
[0023] (1) Compared with the traditional pollution source positioning method based on online water quality sensor data, the water quality dynamic optimization sampling method proposed by the application does not use water quality sensors, which means that the installation and maintenance costs of online water quality sensors are avoided, indicating that the method has lower cost and wider application space.
[0024] (2) The water quality dynamic optimization sampling pollution source positioning method proposed by the application only uses the flow direction information of the network, without the help of any network water quality model, which can more conveniently realize the positioning purpose and has high reliability.
[0025] (3) The application proposes a new idea for water supply network pollution source positioning, which can effectively solve the problems of immature sensor technology and insufficient reliability of water quality model in traditional positioning methods, can quickly and accurately locate the pollution source in the network at a lower cost, and has a wide application scenario. BRIEF DESCRIPTION OF DRAWINGS
[0026] Figure 1 is the specific implementation roadmap of the application.
[0027] Figure 2 is a schematic diagram of the network structure used in the embodiment.
[0028] Figure 3 is a positioning process schematic diagram of one pollution case in the embodiment.
[0029] Figure 4 is a lock area percentage probability density distribution diagram of 347 pollution cases in the embodiment.
[0030] Figure 5is a distribution diagram of the number of cycles required for positioning of 347 pollution cases in the example. DETAILED DESCRIPTION
[0031] The present application is specifically described below by means of the accompanying drawings and examples, so that those skilled in the art can better understand the essence of the present application.
[0032] Reference Figure 1 A water quality dynamic optimization sampling-based water supply network pollution source positioning method, the steps of which are as follows:
[0033] S11: Determine the points that can be sampled in the network, and determine the number of sampling points in each cycle n , n The T sampling points divide the network into T sub-regions:
[0034] T=2 n 1-1 ，
[0035] In this embodiment, fire hydrants are used as sampling points, and the selection principle of sampling points is mainly to facilitate sampling of each pipe in the network. In step S113, a sub-region is a set of pipes that have the same upstream and downstream relationship with a group of sampling points. For example, when n =1, the sub-regions include two, one is composed of pipes upstream of the sampling points in the network, and the other sub-region is composed of the remaining pipes (not upstream of the sampling points). When n =2, the network is divided into four sub-regions: one sub-region is a set of pipes that are upstream of both sampling points, the second sub-region is a set of pipes that are upstream of only the first sampling point, the third sub-region is a set of pipes that are upstream of only the second sampling point, and the fourth sub-region is a set of pipes that are not upstream of both sampling points, and so on, n T sampling points divide the network into T T sub-regions.
[0036] S12: All pipes in the network are marked as suspicious pipes.
[0037] S21: Select sampling points by minimizing the objective function F(A), denoted as sampling group A, sample and detect, and after sampling, send the sample to the laboratory for water quality detection, determine whether the sample is contaminated, and mark yes / no.
[0038] 1-2,
[0039] Wherein, L all is the length of all suspicious pipes at present, L iFor the first i length and.
[0040] S22: Update the suspicious pipeline according to the sample detection results, the updating method is: (1) if all samples are contaminated: update the suspicious pipeline to the intersection of the upstream pipelines of the sampling points; (2) if some samples are contaminated: update the suspicious pipeline to the intersection of the upstream pipelines of the contaminated sampling points minus the union of the upstream pipelines of the uncontaminated sampling points, that is, the suspicious pipeline Q = the intersection of the upstream pipelines of the contaminated sampling points Q1' minus the union of the upstream pipelines of the uncontaminated sampling points Q2; (3) if all samples are uncontaminated: update the suspicious pipeline to the complement of the union of the upstream pipelines of the sampling points, where the suspicious pipeline refers to the pipeline set of the suspicious pipeline, for example, the suspicious pipeline Q = {pipeline 1, pipeline 3, pipeline 4}.
[0041] S31: If the number of contaminated samples in S21 is less than two, skip S3.
[0042] S32: If the updated suspicious pipeline in S22 is empty, update the suspicious pipeline to the union of the upstream pipelines of the contaminated sampling points minus the union of the upstream pipelines of the uncontaminated sampling points, that is, the suspicious pipeline Q = the union of the upstream pipelines of the contaminated sampling points Q1 minus the union of the upstream pipelines of the uncontaminated sampling points Q2, and skip S3.
[0043] S33: Determine one sampling point position as the end of the current suspicious pipeline, and determine the remaining n -1 sampling point according to the method of S21. Sample and detect the determined n sampling point position.
[0044] S34: If the detection result of the suspicious pipeline end sampling point in S33 is uncontaminated, update the suspicious pipeline to the union of the upstream pipelines of the contaminated sampling points minus the union of the upstream pipelines of the uncontaminated sampling points minus the upstream pipeline of the suspicious pipeline end sampling point in S33, that is, the suspicious pipeline Q = the union of the upstream pipelines of the contaminated sampling points Q1 minus the union of the upstream pipelines of the uncontaminated sampling points Q2 minus the upstream pipeline of the suspicious pipeline end sampling point Q3; otherwise, update the suspicious pipeline according to the method in S22.
[0045] Next, based on the method, it is combined with a specific embodiment to show its specific technical effect. Embodiment
[0046] The method proposed in the application will be verified on a pipe network from a real water supply system in a certain city. As Figure 2 shown, there are 149 pipelines and 78 sampling points in the pipe network, and the total length of the pipelines is 58.7 km.
[0047] As Figure 3 shown, two cases with two pollution sources existing simultaneously are selected to demonstrate the implementation process of the model construction method proposed by the present application:
[0048] 1) As shown in (a), there is a pollution source in the pipe network, and the monitoring system issues a warning (through water quality monitoring at the end of the pipe network or through resident complaints), so the first round of positioning process is started. In this case, the sampling capacity is set to 2 (i.e. up to two sampling positions are selected each time). Figure 3 2) As shown in (b), in the first sampling, both sampling points detect pollution, and the suspected pipe is updated to the common upstream pipe of the two sampling points.
[0049] Figure 3 3) As shown in (c), since the number of sampling points with pollution in the last sampling is greater than or equal to two, one sampling point is placed at the end of the suspected pipe, and the other sampling point position is calculated according to formula 1-2. The detection result of the end sampling point is no pollution, so the suspected pipe is updated to the union of the upstream pipes of the sampling points with pollution in the last sampling minus the upstream of the end sampling point.
[0050] 4) As shown in (d), in the third sampling, the sampling point position is determined according to formula 1-2. The detection result is that only one sampling point has pollution, and the suspected pipe is updated to the individual upstream pipe of the sampling point. Figure 3 5) As shown in (e), in the fourth sampling, the sampling point position is determined according to formula 1-2. The detection result is that both sampling points have pollution, and the suspected pipe is updated to the common upstream pipe of the two sampling points.
[0051] Figure 3 6) As shown in (f), since the number of sampling points with pollution in the last sampling is greater than or equal to two, one sampling point is placed at the end of the suspected pipe, and the other sampling point position is calculated according to formula 1-2. Both sampling points detect pollution, so the suspected pipe is updated to the common upstream pipe of the two sampling points.
[0052] 7) As shown in (g), in the sixth sampling, there is only one sampling point left in the suspected pipe. The sampling result is no pollution, and the suspected pipe is updated to the pipe that is not the upstream of the sampling point. There is no sampling point in the updated suspected pipe, so the first round of positioning process ends, and the pollution source 1 is successfully located. Figure 3 8) As shown in (h), in the second round of positioning, the suspected pipe is updated to the pipe that is not the upstream of the sampling point in the first round of positioning. The sampling result is no pollution, so the suspected pipe is updated to the pipe that is not the upstream of the sampling point in the first round of positioning. There is no sampling point in the updated suspected pipe, so the second round of positioning process ends, and the pollution source 2 is successfully located.
[0053] Figure 3 9) As shown in (i), in the third round of positioning, the suspected pipe is updated to the pipe that is not the upstream of the sampling point in the second round of positioning. The sampling result is no pollution, so the suspected pipe is updated to the pipe that is not the upstream of the sampling point in the second round of positioning. There is no sampling point in the updated suspected pipe, so the third round of positioning process ends, and the pollution source 3 is successfully located.
[0054] 10) As shown in (j), in the fourth round of positioning, the suspected pipe is updated to the pipe that is not the upstream of the sampling point in the third round of positioning. The sampling result is no pollution, so the suspected pipe is updated to the pipe that is not the upstream of the sampling point in the third round of positioning. There is no sampling point in the updated suspected pipe, so the fourth round of positioning process ends, and the pollution source 4 is successfully located. Figure 3 11) As shown in (k), in the fifth round of positioning, the suspected pipe is updated to the pipe that is not the upstream of the sampling point in the fourth round of positioning. The sampling result is no pollution, so the suspected pipe is updated to the pipe that is not the upstream of the sampling point in the fourth round of positioning. There is no sampling point in the updated suspected pipe, so the fifth round of positioning process ends, and the pollution source 5 is successfully located.
[0055] Figure 3 (h) (k) After locating and resolving the pollution source 1, the pipe network monitoring system still reports a pollution source. So the same positioning method is implemented again, i.e. entering a second round of positioning process, and finally the pollution source 2 is successfully determined through four sampling cycles.
[0056] In order to verify the positioning effect of the method of the present application on pollution cases with different positions and different numbers of pollution sources under different sampling capabilities, all single pollution events (a total of 149) and 100 random two pollution source events and 100 random three pollution source events in the pipe network are tested by using the method of the present application, so as to prove the effectiveness and high efficiency of the present application in a statistical sense.
[0057] Figure 4 is the probability density distribution diagram of the percentage of the locked area of the total 347 pollution cases in the embodiment. As can be seen from the diagram, for most of the pollution events, the method of the present application can lock the pollution source to a small range of pipe with a length less than 5% of the total length of the pipe network, which shows that the method of the present application has good positioning performance and can greatly narrow down the possible position of the pollution source. At the same time, for some cases in the pipe network, the length of the area finally locked is 10-15% of the total length of the pipe network, which is due to the sparse distribution of the sampleable hydrants in some areas of the pipe network, which shows that the positioning performance of the method of the present application will be affected by the distribution of the sampleable hydrants in the pipe network.
[0058] Figure 5 is the distribution diagram of the number of cycles required for positioning of the total 347 pollution cases in the embodiment. As can be seen from the diagram, the method of the present application can complete the positioning of the pollution source in a relatively small number of cycles, which shows the high efficiency of the present application in positioning the pollution source. In addition, the results of single pollution source events, two pollution source events and three pollution source events show that with the increase of the sampling capability, the number of sampling cycles required for positioning the pollution event is constantly decreasing, but the decreasing amplitude is constantly slowing down. When the sampling capability reaches 3-4, the increase of the sampling capability has little effect on the number of sampling cycles. Considering that the increase of the sampling capability is accompanied by the increase of the cost of the whole positioning process, there may be an optimal sampling capability in the application of the present application in the pipe network, which can balance the cost and efficiency.
[0059] As can be seen from the above results, the method of the present application for positioning the pollution source in the water supply pipe network based on optimized manual sampling can efficiently and accurately position the pollution source at each position in the water supply pipe network. In addition, the method of the present application for positioning the pollution source in the water supply pipe network is not limited to the pipe network model used in the embodiment, but can be widely applied to any other pipe network, and has wide application prospect and good popularization and practical application value.
Claims
1. A water supply network pollution source positioning method based on dynamic optimization sampling of water quality, characterized in that, The steps are as follows: S1: initialize sub-regions and suspicious pipes according to steps S11-S12; S11: Determine the sampling points in the pipe network, determine the number of sampling points in each cycle n , n The pipe network is divided into T sub-regions by the sampling points T=2 n 1-1， wherein a sub-region is a set of pipes that have the same upstream and downstream relationship with a set of sample points, when n = 1, the pipes in the network that are upstream of the sample point collectively form one sub-region, and the rest of the pipes that are not upstream of the sample point form another sub-region; when n = 2, the network is divided into four sub-regions: the set of pipes that are upstream of both sample points, the set of pipes that are upstream of only the first sample point, the set of pipes that are upstream of only the second sample point, and the set of pipes that are not upstream of either sample point; and similarly, n = n, the network is divided into T sub-regions. S12: all pipes in the pipe network are marked as suspicious pipes; S2: sampling, sample detection and suspicious pipe updating are performed according to steps S21-S22; S21: select sampling points by minimizing the objective function F(A), denoted as sampling set A, and sample and detect; 1-2， wherein, L all is the sum of the lengths of all suspicious pipes in the current region, L i is the sum of the lengths of suspicious pipes in the first i region. S22: update suspicious pipes according to the sample detection results, in the following ways: (1) if all samples are contaminated: update the suspicious pipes as the intersection of the upstream pipes of the sampling points; (2) if some samples are contaminated: update the suspicious pipes as the intersection of the upstream pipes of the contaminated sampling points minus the union of the upstream pipes of the uncontaminated sampling points, i.e., suspicious pipes Q = intersection of the upstream pipes of the contaminated sampling points Q1' minus union of the upstream pipes of the uncontaminated sampling points Q2; (3) if all samples are uncontaminated: update the suspicious pipes as the complement of the union of the upstream pipes of the sampling points; S3: perform operations according to the detection results of the last sampling according to steps S31-S34; S31: if the number of contaminated samples in S21 is less than two, skip S3; S32: if the updated suspicious pipes in S22 are empty, update the suspicious pipes as the union of the upstream pipes of the contaminated sampling points minus the union of the upstream pipes of the uncontaminated sampling points, i.e., suspicious pipes Q = union of the upstream pipes of the contaminated sampling points Q1 minus union of the upstream pipes of the uncontaminated sampling points Q2, and skip S3; S33: one sample point position is determined as the end of the current suspicious pipe, and the rest of the n -1 sample point is determined according to the method of S21, and sampling and detection are performed on the determined n sample point position. S34: if the detection result of the end sampling point of the suspicious pipes in S33 is uncontaminated, update the suspicious pipes as the union of the upstream pipes of the contaminated sampling points minus the union of the upstream pipes of the uncontaminated sampling points minus the upstream pipes of the end sampling point of the suspicious pipes in S33, i.e., suspicious pipes Q = union of the upstream pipes of the contaminated sampling points Q1 minus union of the upstream pipes of the uncontaminated sampling points Q2 minus upstream pipes of the end sampling point of the suspicious pipes Q3, otherwise update the suspicious pipes according to the method in S22; S4: repeat the operations of S2-S3 until there is no sampling point in the suspicious pipes, and the positioning is completed.