CN109555979B - Water supply pipe network leakage monitoring method - Google Patents
Water supply pipe network leakage monitoring method Download PDFInfo
- Publication number
- CN109555979B CN109555979B CN201811507342.7A CN201811507342A CN109555979B CN 109555979 B CN109555979 B CN 109555979B CN 201811507342 A CN201811507342 A CN 201811507342A CN 109555979 B CN109555979 B CN 109555979B
- Authority
- CN
- China
- Prior art keywords
- leakage
- data
- water supply
- pipe
- layer
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 title claims abstract description 77
- 238000000034 method Methods 0.000 title claims abstract description 39
- 238000012544 monitoring process Methods 0.000 title claims abstract description 13
- 238000013528 artificial neural network Methods 0.000 claims abstract description 33
- 238000012549 training Methods 0.000 claims description 17
- 238000007477 logistic regression Methods 0.000 claims description 12
- 238000011176 pooling Methods 0.000 claims description 12
- 230000005236 sound signal Effects 0.000 claims description 12
- 238000012706 support-vector machine Methods 0.000 claims description 12
- 238000012360 testing method Methods 0.000 claims description 12
- 230000000694 effects Effects 0.000 claims description 9
- 230000004927 fusion Effects 0.000 claims description 9
- 238000012545 processing Methods 0.000 claims description 9
- 238000010586 diagram Methods 0.000 claims description 7
- 239000011159 matrix material Substances 0.000 claims description 7
- 239000000284 extract Substances 0.000 claims description 3
- 238000001514 detection method Methods 0.000 abstract description 16
- 238000004364 calculation method Methods 0.000 abstract description 5
- 230000009977 dual effect Effects 0.000 description 5
- 230000005540 biological transmission Effects 0.000 description 4
- 238000010606 normalization Methods 0.000 description 4
- 230000008569 process Effects 0.000 description 4
- 241000209094 Oryza Species 0.000 description 3
- 235000007164 Oryza sativa Nutrition 0.000 description 3
- 235000009566 rice Nutrition 0.000 description 3
- 238000010845 search algorithm Methods 0.000 description 3
- 239000008399 tap water Substances 0.000 description 3
- 235000020679 tap water Nutrition 0.000 description 3
- 238000004458 analytical method Methods 0.000 description 2
- 239000012736 aqueous medium Substances 0.000 description 2
- 238000013527 convolutional neural network Methods 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 238000012423 maintenance Methods 0.000 description 2
- 238000007726 management method Methods 0.000 description 2
- 239000002609 medium Substances 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000003062 neural network model Methods 0.000 description 2
- 230000000644 propagated effect Effects 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 230000007797 corrosion Effects 0.000 description 1
- 238000005260 corrosion Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000009440 infrastructure construction Methods 0.000 description 1
- 230000004807 localization Effects 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 238000005065 mining Methods 0.000 description 1
- 239000013307 optical fiber Substances 0.000 description 1
- 239000000700 radioactive tracer Substances 0.000 description 1
- 230000002787 reinforcement Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000035945 sensitivity Effects 0.000 description 1
- 230000007847 structural defect Effects 0.000 description 1
- 230000004083 survival effect Effects 0.000 description 1
- 230000001052 transient effect Effects 0.000 description 1
- 239000002699 waste material Substances 0.000 description 1
Images
Classifications
-
- F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
- F17—STORING OR DISTRIBUTING GASES OR LIQUIDS
- F17D—PIPE-LINE SYSTEMS; PIPE-LINES
- F17D5/00—Protection or supervision of installations
- F17D5/02—Preventing, monitoring, or locating loss
- F17D5/06—Preventing, monitoring, or locating loss using electric or acoustic means
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Acoustics & Sound (AREA)
- Mechanical Engineering (AREA)
- General Engineering & Computer Science (AREA)
- Examining Or Testing Airtightness (AREA)
Abstract
The utility model provides a water supply pipe network leakage monitoring method, which comprises the following steps: s1, acquiring water supply network data; s2, establishing a heterogeneous double-classifier leakage identification model based on a deep neural network; and S3, utilizing the water supply network data and the heterogeneous dual-classifier leakage identification model to identify the leakage of the water supply network. The water supply pipe network leakage monitoring method effectively improves the calculation efficiency of pipeline leakage detection and positioning, enlarges the application range of the pipeline leakage detection method, and can perform high-precision leakage point detection and positioning under the condition of complex working conditions.
Description
Technical Field
The utility model belongs to the field of leakage detection of urban water supply networks, and particularly relates to a leakage monitoring method for a water supply network.
Background
Water resources are the foundation of human survival and development, the total amount of water resources in China is rich and is listed in the fourth position of the world, but the per-capita water resource holding amount is only 2300 cubic meters, which is equivalent to about 1/4 of the world average level, so that China is a country with shortage of water resources. Urban water supply pipe network systems are an important field of urban infrastructure construction, called "lifeline engineering". However, the leakage problem of the water supply network always troubles all large water supply companies in China, not only causes the waste of resources and energy, but also causes secondary disasters such as ground settlement and the like, and influences the urban safety. The survey data of the Ministry of construction shows that the leakage rate of most of cities in China is between 15% and 35%, while the leakage rate of developed countries such as Japan, America and Europe is generally controlled to be about 10% or even lower at the end of the last century, so that the control and management of the leakage of the pipe network in China are in urgent need of reinforcement.
The current water supply network leakage detection methods include a passive detection method, a regional surface mounting method, a surface radar leakage detection method, a tracer detection method, an acoustic detection method, an optical fiber sensing technology method, a negative pressure wave method, a real-time transient model method and the like. The acoustic detection method has the advantages of simplicity, reliability, high detection efficiency, wide application range and the like, and is widely applied to pipeline leakage detection and positioning. However, due to the insufficient recognition of the generation mechanism and characteristics of the leakage acoustic signals of the water supply network and the insufficient consideration of the actual complex topology structure of the water supply network, the existing pipeline leakage detection method is limited in actual application, the pipeline leakage identification accuracy is low, the false alarm rate is high, and especially under the conditions that various fixed interference noises exist in a detection field, the leakage rate of a leakage point is small, or the structure of the pipe network is complex and part of pipeline information is unknown.
Therefore, based on the current situation and the future development trend of urban water supply networks in China, the sound wave transmission theory, the deep neural network model and the local search positioning model are used for detecting the leakage of the pipeline caused by structural defects of the pipeline, internal and external corrosion and the like, and an effective water supply network leakage identification and positioning technical method is established, so that basis and guidance are provided for decision making of maintenance, maintenance and management of a pipe network structure, the leakage rate of the pipe network is reduced, and the safe operation of the water supply network is guaranteed.
Disclosure of Invention
Technical problem to be solved
In view of the above problems, it is a primary object of the present disclosure to provide a water supply pipe network leakage monitoring method to solve at least one of the above problems.
(II) technical scheme
In order to achieve the above object, as one aspect of the present disclosure, there is provided a water supply pipe network leakage monitoring method, including the steps of:
s1, acquiring water supply network data;
s2, establishing a heterogeneous double-classifier leakage identification model based on a deep neural network; and
s3, identifying the leakage of the water supply network by using the water supply network data and the heterogeneous dual-classifier leakage identification model; the water supply network data comprises leakage acoustic signal data transmitted along a water medium, flow data in the water supply pipe, pressure data in the water supply pipe, pipe data and pipe diameter data.
In some embodiments, after the step S3, the method further includes:
s4, establishing a local searching and positioning model based on the topological structure of the water supply network; and
and S5, carrying out leakage positioning by using the water supply pipe network leakage identification result and the local searching and positioning model.
In some embodiments, the deep neural network based heterogeneous dual classifier leakage recognition model comprises a convolutional layer, a max pooling layer, a long-short term neural network layer, a first fully-connected layer, a fusion layer, a second fully-connected layer, a support vector machine classifier, a logistic regression classifier, and a heterogeneous dual classifier.
In some embodiments, the step S3 includes:
receiving leakage acoustic signal data by the convolution layer;
the maximum pooling layer divides the output of the convolutional layer into m sub-regions, extracts the maximum value of each sub-region to form output, and m is a positive integer;
the long-time neural network layer and the short-time neural network layer perform nonlinear data processing on the output of the maximum pooling layer;
the first full-connection layer receives flow data, pressure data, pipe data and pipe diameter data;
the fusion layer receives the outputs of the long-time and short-time neural network layer and the first full connection layer;
a second fully connected layer receives the output of the fused layer;
the second full connection layer is respectively connected with the support vector machine classifier and the logistic regression classifier;
the support vector machine classifier receives the output of the second full connection layer, classifies and identifies the leakage event of the water supply network and outputs a classification vector Y1; the logistic regression classifier receives the output of the second full-connection layer, classifies and identifies the leakage events of the water supply network and outputs a classification vector Y2, wherein the classification vector Y1 and the classification vector Y2 are probability values of each leakage identification result;
the heterogeneous double classifier obtains a leakage identification result Y by calculation according to the formula (1) according to the classification vector Y1 and the classification vector Y2,
y β 1Y 1+ β 2Y 2 formula (1)
Wherein β 1+ β 2 is 1, 0 < β 1 <1, 0 < β 2< 1.
In some embodiments, the leakage acoustic signal data is acquired with a hydrophone sensor, the flow data is acquired with a flow meter sensor, and the pressure data is acquired with a pressure gauge sensor.
In some embodiments, the sensor S corresponding to the occurrence of the leakage event is determined by a leakage recognition modelkThe number of the corresponding sensors is more thanOr equal to 2, the leak localization is performed.
In some embodiments, the step S5 includes:
with the sensor SkForming a closed loop on the pipe network diagram by using the shortest path as a base point, wherein the closed loop comprises i pipeline nodes, and the number of the corresponding pipeline node is ri(ii) a j tubes of pipeline with corresponding pipeline number lj(ii) a k sensors with corresponding sensor number SkI, j and k are positive integers, and the shortest path is through the sensor SkThe shortest circumference of the closed loop of (a);
using an objective function fiSearching the pipeline node closest to the leakage point in the closed loop, and sequentially calculating f of the i pipeline nodesiValue, choose the smallest fiThe pipeline node c corresponding to the value is taken as a virtual leakage point v nearest to the leakage pointc；
With virtual missing points vcAs a center, search a path with a virtual leak point vcThe connected pipelines are provided with 1 virtual leakage point v every z metersgN, and sequentially calculating f of the n virtual leakage pointsiValue, choose the smallest fiAnd taking the virtual leakage point corresponding to the value as a final leakage positioning point.
In some embodiments, the objective function is as shown in equation (2),
fi＝∑a≠b(|ta-tb|-|ωia-ωib|)2formula (2)
In the formula (2), fiError square value, t, representing the time difference between arrival of a leaky acoustic signal at the ith node at different sensorsaRepresenting the time, t, required for the leakage acoustic signal to reach the sensor abRepresenting the time, ω, required for the leakage acoustic signal to reach the sensor biaRepresenting the time, ω, required for the leakage acoustic signal to reach sensor a from node iibRepresenting the time required for the leaky acoustic signal to reach sensor b from node i, which may be a pipe node or a virtual leak point.
In some embodiments, the step S2 includes:
acquiring a leakage sound signal data sequence S (t), a flow data sequence Q (t) and a pressure data sequence P (t) of a single sensor, and normalizing the leakage sound signal data sequence S (t), the flow data sequence Q (t) and the pressure data sequence P (t);
acquiring pipe diameter and pipe data D (t), carrying out independent thermal coding treatment on the pipe diameter and pipe data, and converting the pipe diameter and pipe data into a sequence only containing numbers of 0 and 1;
dividing the leakage sound signal data sequence S (t), the flow data sequence Q (t), the pressure data sequence P (t) and the corresponding pipe diameter and pipe data D (t) into a training set and a testing set at random;
selecting a cross entropy loss function as a classification target of the model, and training the heterogeneous dual-classifier leakage recognition model based on the deep neural network by using a training set until the cross entropy loss function value is unchanged, and stopping training;
and testing the trained model by using a test set, and evaluating the effect of the model by using a confusion matrix.
(III) advantageous effects
According to the technical scheme, the technical method for monitoring the leakage of the water supply pipe network (including leakage identification and positioning) at least has the following beneficial effects:
(1) the deep neural network and the heterogeneous dual classifiers are combined, the problem of low recognition accuracy of a single classifier is solved, the effect of pipeline leakage detection in the actual process is effectively improved, and particularly recognition under interference of surrounding complex environments can be realized. Meanwhile, the size of the leakage quantity can be judged according to the identification result, and a foundation is provided for the next processing.
(2) The local search positioning model based on the water supply pipe network topological structure is used for positioning the leakage point, and the calculation efficiency of leakage positioning is improved by adopting a local search algorithm based on the pipe network topological structure; the error square value of the time difference of the node reaching different sensors is selected as a target function, so that the sensitivity of the model is improved; the pipe network topological structure in practice is fully considered, the application range of pipeline leakage detection and identification is expanded, and high-precision leakage point positioning can be performed under the condition of complex working conditions.
(3) The data are collected by using the sensors, including leakage acoustic signal data, flow data, pressure data, pipe diameter and pipe data, fusion and mining of multi-dimensional data are achieved through the deep neural network, and the application effect of the data in practice is further improved.
(4) The method obviously improves the identification accuracy of leakage events and the positioning accuracy of leakage points, enlarges the application range of the method in the actual complex pipe network topological structure, facilitates daily managers to find the leakage of the pipe network in time and maintain the pipe network in time, reduces economic loss, saves water resources, and assists a tap water company to make scientific and reasonable decisions.
Drawings
FIG. 1 is a flow chart of a water supply network leakage identification method of the present disclosure.
FIG. 2 is a schematic diagram of a water supply network leakage identification and location method according to the present disclosure.
Fig. 3 is a structural diagram of a leakage identification model of a heterogeneous dual classifier based on a deep neural network according to the present disclosure.
FIG. 4 is a schematic diagram of a water supply network topology in an area according to an embodiment of the disclosure.
Detailed Description
For a better understanding and an enabling description of the present disclosure, reference will now be made in detail to the present disclosure, examples of which are illustrated in the accompanying drawings. It is to be understood that although the embodiments of the present disclosure have been described, it is apparent that the present disclosure is not limited to the above embodiments, and various modifications can be made within a scope not departing from the gist thereof.
The monitoring method aims to solve the defects of the prior art, combines a sound wave transmission theory, a deep neural network model and a local search positioning model, and aims to provide a monitoring method for identifying and positioning leakage of the urban water supply network. The method can improve the accuracy of identifying the leakage of the water supply network and reduce the false alarm rate, so that a daily manager can find the leakage of the water supply network in time and maintain the water supply network in time, the economic loss is reduced, the water resource is saved, and a tap water company is assisted to make a scientific and reasonable decision.
As shown in fig. 1, the leakage monitoring method (leakage identification) for water supply network of the present disclosure includes the following steps:
s1, acquiring water supply network data;
s2, establishing a heterogeneous double-classifier leakage identification model based on a deep neural network; and
and S3, identifying the leakage of the water supply network by using the water supply network data and the heterogeneous double-classifier leakage identification model.
The water supply network data comprises leakage acoustic signal data transmitted along a water medium, flow data in the water supply pipe, pressure data in the water supply pipe, pipe data and pipe diameter data.
Further, the water supply network leakage monitoring method (leakage positioning) of the present disclosure further includes, after the step S3:
s4, establishing a local searching and positioning model based on the topological structure of the water supply network; and
and S5, carrying out leakage positioning by using the water supply pipe network leakage identification result and the local searching and positioning model.
In general, the method disclosed by the invention mainly comprises three parts, namely, data acquisition, leakage identification model establishment and leakage point positioning model establishment, which are sequentially performed, and are shown in fig. 2.
Specifically, the data acquisition process is as follows: arranging a sensor on a water supply network for data acquisition, the sensor comprising: hydrophones, flow meters, pressure gauges. The hydrophone receives the acoustic signal propagated along the aqueous medium, the flow meter measures the flow in the pipe, and the pressure gauge measures the pipe pressure. The sensors transmit the acquired data to the control center through wireless transmission and store the data in the computer. Meanwhile, basic information of the pipeline is obtained from a city water supply network database, wherein the basic information comprises pipe data and pipe diameter data.
The leakage identification model establishment process is as follows: establishing a heterogeneous dual-classifier leakage identification model based on a deep neural network, as shown in fig. 3, the specific steps include:
the convolutional layer is used for receiving the leakage acoustic signal data, belongs to a convolutional neural network and is used for extracting the characteristics of the leakage acoustic signal;
the maximum pooling layer is connected with the convolution layer, the maximum pooling layer divides the output of the convolution layer into m sub-regions, the maximum value of each sub-region is extracted to form the output, and m is a positive integer;
the long and short time neural network layer is connected with the maximum pooling layer, belongs to a variant of a circulating neural network, and has strong capacity of processing nonlinear data;
the first full-connection layer is used for receiving flow data, pressure data, pipe and pipe diameter data, and is connected with the long and short time neural network layer to the fusion layer in a tensor series mode;
the fusion layer is connected to a second fully-connected layer in a tensor series manner;
the second full connection layer is respectively connected with the support vector machine classifier and the logistic regression classifier;
the support vector machine classifier classifies and identifies the pipeline leakage event and outputs a classification vector Y1;
the logistic regression classifier classifies and identifies the pipeline leakage event and outputs a classification vector Y2;
the classification vector Y1 ═ (p1, …, p8) and classification vector Y2 ═ (q1, …, q8) are probability values for the occurrence of each missing recognition result, as shown in table 1;
the heterogeneous double classifier obtains a leakage identification result Y by calculation according to the formula (1) according to the classification vector Y1 and the classification vector Y2, as shown in Table 1.
Y β 1Y 1+ β 2Y 2 formula (1)
Wherein β 1+ β 2 is 1, 0 < β 1 <1, 0 < β 2< 1.
The heterogeneous double-classifier leakage identification model based on the deep neural network is used for modeling a pipeline leakage event, the model is input into a leakage sound signal data sequence S (t), a flow data sequence Q (t), a pressure data sequence P (t), corresponding pipe diameter and pipe data D (t), a classification result Y1 of a support vector machine classifier, a classification result Y2 of a logistic regression classifier and a leakage identification result Y, the model outputs Y1, Y2 and Y are divided into 8 classes according to the size of leakage, and each class corresponds to a probability value as shown in Table 1.
And analyzing each sensor by using the leakage identification model, inputting the data acquired by the sensors into the leakage identification model, and determining that a leakage event occurs in the pipe network when the leakage identification corresponding to the maximum probability value in the identification result Y output by the leakage identification model is classified into 2-8, wherein the corresponding sensor is the sensor corresponding to the leakage event.
TABLE 1 leakage recognition result output
For each sensor, the heterogeneous double-classifier leakage identification model based on the deep neural network is adopted for modeling, and the specific steps are as follows:
1) acquiring a leakage sound signal data sequence S (t), a flow data sequence Q (t) and a pressure data sequence P (t) of a single sensor, and carrying out normalization processing on the leakage sound signal data sequence S (t), the flow data sequence Q (t) and the pressure data sequence P (t), wherein the normalization processing comprises the step of converting the numerical value of original data into a range of [0, 1], and the formula of the normalization processing is shown as a formula (2),
in the formula (2), y represents normalized data, x represents input raw data, and x representsmaxAnd xminRepresenting the maximum and minimum values of the input data, respectively.
And acquiring corresponding pipe diameter and pipe data D (t), carrying out independent thermal coding treatment on the pipe diameter and pipe data D (t), and converting discrete variables such as the pipe diameter and the pipe into a sequence only containing numbers of 0 and 1.
And randomly dividing the leakage sound signal data sequence S (t), the flow data sequence Q (t), the pressure data sequence P (t) and the corresponding pipe diameter and pipe data D (t) into a training set and a testing set.
2) Training the heterogeneous dual-classifier leakage recognition model based on the deep neural network by using a training set, selecting a cross entropy loss function as a classification target of the model, wherein the formula of the cross entropy loss function is shown as a formula (3),
in the formula (3), L represents a cross entropy loss function value, N represents a sample amount, and hpqRepresenting the probability value, y, that a sample p belongs to the class qpqRepresenting the probability value that the model predicts for sample p as belonging to class q.
When the cross entropy loss function value L is not changed any more, the model stops training.
3) And (3) testing the trained model by using a test set, and evaluating the effect of the model by using a confusion matrix, wherein the confusion matrix is shown in a table 2.
TABLE 2 confusion matrix
| Normal events | Leakage event | |
| Model identification as Normal event | TP | FP |
| Model identification as leakageEvent(s) | FN | TN |
In table 2, the TP value indicates an actual normal event, and the model identification result also indicates a normal event; the FP value represents a leakage event in practice, and the model identification result is a normal event; the FN value represents an actual normal event, and the model identification result is a leakage event; the TN value represents the actual leakage event, and the model identification result is also the leakage event.
According to the 4 values, the identification accuracy rate of the calculation model is TN/(FP + TN) and the false alarm rate is FN/(TP + FN), and the higher the accuracy rate is, the lower the false alarm rate is, and the better the model effect is.
The process of establishing the missing point positioning model is as follows: establishing a local search positioning model based on a pipe network topological structure for positioning a leakage point, and specifically comprising the following steps of:
1) and a leakage identification model is established for analyzing the data acquired by each sensor, but a leakage positioning model is established for analyzing only the sensor with the leakage identification result of the leakage event. And when the leakage identification class corresponding to the maximum probability value in the identification result Y output by the leakage identification model is 2-8, determining that the leakage event occurs in the pipe network. Determining the corresponding sensor S when a leakage event occurs through a leakage identification modelkWhen the number of the sensors determined as the leakage event is more than or equal to 2, a leakage positioning model is established for analysis, and the sensors S are used for analysiskAnd forming a closed loop on the pipe network diagram by using the shortest path as a base point. The closed loop comprises i pipeline nodes, and the number of the corresponding pipeline node is ri(ii) a j tubes of pipeline with corresponding pipeline number lj(ii) a k sensors with corresponding sensor number SkAnd i, j and k are positive integers. The shortest path is through the sensor SkThe shortest circumference of the closed loop.
2) A local search algorithm based on a pipe network topological structure is established, an objective function is different from the formula (4),
fi＝∑a≠b(|ta-tb|-|ωia-ωib|)2formula (4)
In the formula (4), fiError square value, t, representing the time difference between arrival of a leaky acoustic signal at the ith node at different sensorsaRepresenting the time, t, required for the leakage acoustic signal to reach the sensor abRepresenting the time, ω, required for the leakage acoustic signal to reach the sensor biaRepresenting the time, ω, required for the leakage acoustic signal to reach sensor a from node iibRepresenting the time required for a leakage acoustic signal to reach sensor b from node i, which may be a pipe node or a virtual leakage point;
3) searching the pipeline node closest to the leakage point in the closed loop, and calculating f of the i pipeline nodes according to the formula (4)iValue, choose the smallest fiThe pipeline node c corresponding to the value is taken as a virtual leakage point v nearest to the leakage pointc；
4) With virtual missing points vcAs a center, search a path with a virtual leak point vcThe connected pipelines are provided with 1 virtual leakage point v every z metersgN, and f of the n virtual leak points is calculated according to equation (4)iValue, choose the smallest fiAnd taking the virtual leakage point corresponding to the value as a final leakage positioning point.
FIG. 4 is a schematic view showing a water supply network topology in a region where sensors are installed at pipe nodes or fire plugs, and the network has 6 sensors S1…S 66 pipe nodes r1…r616 pipelines with the corresponding pipeline number l1…l16True missing point b1。
Arranging sensors on a water supply network for data acquisition, each multi-sensor comprising: hydrophones, flow meters, pressure gauges. The hydrophone receives the acoustic signal propagated along the aqueous medium, the flow meter measures the flow in the pipe, and the pressure gauge measures the pipe pressure. The sensors transmit the acquired data to the control center through wireless transmission and store the data in the computer. Meanwhile, basic information of the pipeline, including pipe and pipe diameter data, is obtained from a city water supply pipe network database.
Aiming at the embodiment, the method for establishing the heterogeneous dual-classifier leakage identification model based on the deep neural network comprises the following steps:
the convolutional layers are used for receiving the data of the leakage acoustic signals, belong to a convolutional neural network and are used for extracting the characteristics of the leakage acoustic signals;
connecting 1 maximum pooling layer with the convolution layer, wherein the maximum pooling layer divides the output of the convolution layer into 20 sub-regions, and extracts the maximum value of each sub-region to form output;
1 long and short time neural network layer is connected with the maximum pooling layer, belongs to a variant of a circulating neural network, and has strong capacity of processing nonlinear data;
the first full-connection layer is used for receiving flow data, pressure data, pipe and pipe diameter data, and is connected with the long and short time neural network layer to 1 fusion layer in a tensor series mode;
the fusion layer is connected to a second fully-connected layer in a tensor series manner;
the second full connection layer is respectively connected with the support vector machine classifier and the logistic regression classifier;
the support vector machine classifier classifies and identifies the pipeline leakage event and outputs a classification vector Y1;
the logistic regression classifier classifies and identifies the pipeline leakage event and outputs a classification vector Y2;
the classification vector Y1 ═ (p1, …, p8) and classification vector Y2 ═ (q1, …, q8) are probability values for the occurrence of each missing recognition result, as shown in table 1;
and the heterogeneous dual classifier obtains a leakage identification result Y- β 1Y 1+ β 2Y 2 according to the formula (1) according to the classification vector Y1 and the classification vector Y2, wherein β 1 is 0.4, and β 2 is 0.6.
The heterogeneous double-classifier leakage identification model based on the deep neural network is used for modeling a pipeline leakage event, the model is input into a leakage sound signal data sequence S (t), a flow data sequence Q (t), a pressure data sequence P (t), corresponding pipe diameter and pipe data D (t), a classification result Y1 of a support vector machine classifier, a classification result Y2 of a logistic regression classifier and a leakage identification result Y, the model outputs Y1, Y2 and Y are divided into 8 classes according to the size of leakage, and each class corresponds to a probability value as shown in Table 1.
For 6 sensors, the heterogeneous dual-classifier leakage identification model based on the deep neural network is adopted for modeling, and here, the sensor 1 is taken as an embodiment, and the specific steps are as follows:
1) a leakage acoustic signal data sequence S (t), a flow rate data sequence Q (t), and a pressure data sequence P (t) in the sensor 1 are acquired, and normalization processing is performed according to equation (2).
And acquiring corresponding pipe diameter and pipe data D (t), carrying out independent thermal coding treatment on the pipe diameter and pipe data D (t), and converting discrete variables such as the pipe diameter and the pipe into a sequence only containing numbers of 0 and 1.
The leakage acoustic signal data sequence S (t), the flow data sequence Q (t), the pressure data sequence P (t), the pipe diameter and pipe data D (t) are randomly divided into a training set and a testing set, wherein the training set comprises 32000 samples, and the testing set comprises 8000 samples.
2) And (3) training the heterogeneous dual-classifier leakage recognition model based on the deep neural network by using a training set, and stopping the training of the model when the cross entropy loss function value L is not changed any more.
3) The trained model was tested using 8000 samples of the test set, and the model effect was evaluated using a confusion matrix, as shown in table 3,
TABLE 3 confusion matrix
| Normal events | Leakage event | |
| Model identification as Normal event | 7870 of | 5 are provided with |
| Model identification as a leak event | 30 pieces of | 95 are provided with |
Therefore, the recognition accuracy of the model is 95/(95+5) to 95%, the false alarm rate is 30/(7870+30) to 0.38%, and the model effect is good.
Establishing a local search positioning model based on a pipe network topological structure, and specifically comprising the following steps:
1) determining that the corresponding sensor is S when the leakage event occurs through the leakage identification model1、S2、S3Then with a sensor S1、S2、S3As a base point, a closed loop is formed on the pipe network diagram by the shortest path, and the closed loop comprises a sensor S1、S2、S3Pipe node r1、r2、r3Pipe l1、l2、l3、l4、l5、l6Table 4 shows the basic data of the area pipe network.
TABLE 4 basic data of a certain area pipe network
| Pipeline numbering | Pipe length (rice) | Pipeline numbering | Pipe length (rice) | Pipeline numbering | Pipe length (rice) |
| l1 | 120 | l7 | 67 | l13 | 66 |
| l2 | 140 | l8 | 65 | l14 | 100 |
| l3 | 60 | l9 | 83 | l15 | 95 |
| l4 | 55 | l10 | 79 | l16 | 107 |
| l5 | 68 | l11 | 62 | ||
| l6 | 63 | l12 | 49 |
2) Establishing a local search algorithm based on a pipe network topological structure, and calculating a pipeline node r according to the formula (4)1、r2、r3Corresponding fiValue, where pipe node t1Corresponding error square f1Minimum, so select pipe node t1As virtual leakage point v nearest to the leakage pointc；
3) With virtual missing points vcAs a center, along and virtual leak point vcConnected pipeline l2And l3Setting 1 virtual leakage point v every 2 metersg(g 1.. 100), a total of 100 virtual leak points, and f of the 100 virtual leak points is calculated according to equation (4)iA value of wherein f65Minimum, so virtual leak point v65Is closest to the real leakage point b1A virtual leak point v65And true leak point b1The distance is 0.88m, and the positioning precision is higher.
The above results show that the leakage accident of the water supply pipe network can be accurately identified, the leakage point can be accurately positioned, and the method is high in practicability. The method expands the research content of the existing pipe network leakage identification and positioning method, and provides a new idea for a tap water company to make scientific and reasonable decisions. The above-mentioned embodiments are intended to illustrate the objects, aspects and advantages of the present disclosure in further detail, and it should be understood that the above-mentioned embodiments are only illustrative of the present disclosure and are not intended to limit the present disclosure, and any modifications, equivalents, improvements and the like made within the spirit and principle of the present disclosure should be included in the scope of the present disclosure.
Claims (7)
1. A water supply pipe network leakage monitoring method comprises the following steps:
s1, acquiring water supply network data;
s2, establishing a heterogeneous double-classifier leakage identification model based on a deep neural network; and
s3, identifying the leakage of the water supply network by using the water supply network data and the heterogeneous dual-classifier leakage identification model;
the water supply network data comprises leakage sound signal data transmitted along a water medium, flow data in the water supply pipe, pressure data in the water supply pipe, pipe data and pipe diameter data;
the heterogeneous double-classifier leakage identification model based on the deep neural network comprises a convolutional layer, a maximum pooling layer, a long-time and short-time neural network layer, a first full-connection layer, a fusion layer, a second full-connection layer, a support vector machine classifier, a logistic regression classifier and a heterogeneous double classifier;
the step S3 includes:
receiving leakage acoustic signal data by the convolution layer;
the maximum pooling layer divides the output of the convolutional layer into m sub-regions, extracts the maximum value of each sub-region to form output, and m is a positive integer;
the long-time neural network layer and the short-time neural network layer perform nonlinear data processing on the output of the maximum pooling layer;
the first full-connection layer receives flow data, pressure data, pipe data and pipe diameter data;
the fusion layer receives the outputs of the long-time and short-time neural network layer and the first full connection layer;
a second fully connected layer receives the output of the fused layer;
the second full connection layer is respectively connected with the support vector machine classifier and the logistic regression classifier;
the support vector machine classifier receives the output of the second full connection layer, classifies and identifies the leakage event of the water supply network and outputs a classification vector Y1; the logistic regression classifier receives the output of the second full connection layer, classifies and identifies the leakage event of the water supply network and outputs a classification vector Y2;
the heterogeneous double classifier utilizes the classification vector Y1 and the classification vector Y2 to calculate a leakage identification result Y according to the following formula,
Y＝β1*Y1+β2*Y2
wherein β 1+ β 2 is 1, 0 < β 1 <1, 0 < β 2< 1.
2. The method of claim 1, further comprising, after said step S3:
s4, establishing a local searching and positioning model based on the topological structure of the water supply network; and
and S5, carrying out leakage positioning by using the water supply pipe network leakage identification result and the local searching and positioning model.
3. The method of claim 2, wherein sensors are arranged on the water supply network for data acquisition to obtain water supply network data, the sensors comprising hydrophones, flow meters, and pressure gauges; and acquiring the leakage acoustic signal data by using a hydrophone, acquiring flow data by using a flowmeter, and acquiring pressure data by using a pressure gauge.
4. The method of claim 3, wherein the sensor S corresponding to the occurrence of a leak event is determined by a leak recognition modelkAnd when the number of the corresponding sensors is more than or equal to 2, carrying out leakage positioning.
5. The method according to claim 4, wherein the step S5 includes:
with the sensor SkForming a closed loop on the pipe network diagram by using the shortest path as a base point, wherein the closed loop comprises i pipeline nodes, and the number of the corresponding pipeline node is ri(ii) a j tubes of pipeline with corresponding pipeline number lj(ii) a k sensors with corresponding sensor number SkI, j and k are positive integers, and the shortest path is through the sensor SkThe shortest circumference of the closed loop of (a);
using an objective function fiSearching the pipeline node closest to the leakage point in the closed loop, and sequentially calculating f of the i pipeline nodesiValue, choose the smallest fiThe pipeline node c corresponding to the value is taken as a virtual leakage point v nearest to the leakage pointc；
With virtual missing points vcAs a center, search a path with a virtual leak point vcThe connected pipelines are provided with 1 virtual leakage point v every z metersg(g-1 … n), n, calculating f of the n virtual leakage points in turniValue, choose the smallest fiAnd taking the virtual leakage point corresponding to the value as a final leakage positioning point.
6. The method of claim 5, wherein the objective function is represented by the following formula,
wherein f isiError square value, t, representing the time difference between arrival of a leaky acoustic signal at the ith node at different sensorsaRepresenting the time, t, required for the leakage acoustic signal to reach the sensor abRepresenting the time, ω, required for the leakage acoustic signal to reach the sensor biaRepresenting the time, ω, required for the leakage acoustic signal to reach sensor a from node iibRepresenting the time required for the leaky acoustic signal to reach sensor b from node i, which may be a pipe node or a virtual leak point.
7. The method according to claim 3, wherein the step S2 includes:
acquiring a leakage sound signal data sequence S (t), a flow data sequence Q (t) and a pressure data sequence P (t) of a single sensor, and normalizing the leakage sound signal data sequence S (t), the flow data sequence Q (t) and the pressure data sequence P (t);
acquiring pipe diameter and pipe data D (t), carrying out independent thermal coding treatment on the pipe diameter and pipe data, and converting the pipe diameter and pipe data into a sequence only containing numbers of 0 and 1;
dividing the leakage sound signal data sequence S (t), the flow data sequence Q (t), the pressure data sequence P (t) and the corresponding pipe diameter and pipe data D (t) into a training set and a testing set at random;
selecting a cross entropy loss function as a classification target of the model, and training the heterogeneous dual-classifier leakage recognition model based on the deep neural network by using a training set until the cross entropy loss function value is unchanged, and stopping training;
and testing the trained model by using a test set, and evaluating the effect of the model by using a confusion matrix.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811507342.7A CN109555979B (en) | 2018-12-10 | 2018-12-10 | Water supply pipe network leakage monitoring method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811507342.7A CN109555979B (en) | 2018-12-10 | 2018-12-10 | Water supply pipe network leakage monitoring method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109555979A CN109555979A (en) | 2019-04-02 |
| CN109555979B true CN109555979B (en) | 2020-03-17 |
Family
ID=65869450
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811507342.7A Active CN109555979B (en) | 2018-12-10 | 2018-12-10 | Water supply pipe network leakage monitoring method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109555979B (en) |
Families Citing this family (25)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110119853B (en) * | 2019-05-29 | 2020-12-25 | 浙江大学 | Water supply network leakage alarm threshold value selection method based on time series monitoring data analysis |
| CN110792928B (en) * | 2019-09-24 | 2021-08-10 | 中国石油化工股份有限公司 | Pipeline leakage diagnosis method based on big data combined algorithm |
| US11704458B2 (en) * | 2019-10-10 | 2023-07-18 | Voda, Inc. | System and method for optimizing utility pipe sensors placement using artificial intelligence technology |
| WO2021081250A1 (en) * | 2019-10-22 | 2021-04-29 | Eog Resources, Inc. | Anomaly detection in pipelines and flowlines |
| CN110841143B (en) * | 2019-10-22 | 2022-03-08 | 江苏爱朋医疗科技股份有限公司 | Method and system for predicting state of infusion pipeline |
| CN110837933A (en) * | 2019-11-11 | 2020-02-25 | 重庆远通电子技术开发有限公司 | Leakage identification method, device, equipment and storage medium based on neural network |
| CN113405026B (en) * | 2020-03-16 | 2022-11-11 | 中国港湾工程有限责任公司 | Newly-added leakage judgment and leakage rate estimation method suitable for water supply pipe network in underdeveloped area |
| CN111520615B (en) * | 2020-04-28 | 2021-03-16 | 清华大学 | Pipe network leakage identification and positioning method based on line spectrum pair and cubic interpolation search |
| GB2595660A (en) * | 2020-06-01 | 2021-12-08 | Homeserve Plc | Flow event classification |
| CN111881999A (en) * | 2020-08-04 | 2020-11-03 | 武汉易维环境工程有限公司 | Water service pipeline leakage detection method and system based on deep convolutional neural network |
| CN112460495B (en) * | 2020-11-14 | 2022-09-27 | 武汉众智鸿图科技有限公司 | Monitoring point layout method and system for leakage monitoring positioning |
| CN112610903A (en) * | 2020-12-10 | 2021-04-06 | 合肥学院 | Water supply pipe network leakage positioning method based on deep neural network model |
| CN112833336A (en) * | 2021-01-06 | 2021-05-25 | 万佳安智慧生活技术(深圳)有限公司 | Method and system for monitoring leakage of water supply pipeline between buildings |
| CN112393127B (en) * | 2021-01-19 | 2021-03-26 | 浙江和达科技股份有限公司 | Urban water supply network leakage management and control system |
| CN113112126B (en) * | 2021-03-22 | 2024-05-03 | 浙江和达科技股份有限公司 | Artificial intelligence-based water supply pipe leakage early warning method and system |
| CN113177730B (en) * | 2021-05-20 | 2022-09-23 | 中山大学 | Water resource traceability management system |
| CN113486950B (en) * | 2021-07-05 | 2023-06-16 | 华能国际电力股份有限公司上安电厂 | Intelligent pipe network water leakage detection method and system |
| CN113466424B (en) * | 2021-08-12 | 2022-02-15 | 广州高得环保科技股份有限公司 | Method and system for monitoring water quality of pipeline direct drinking water of intelligent Internet of things |
| CN113739082B (en) * | 2021-09-14 | 2023-05-30 | 杭州莱宸科技有限公司 | Water supply network leakage monitoring method based on multi-region joint evaluation |
| CN113970073B (en) * | 2021-11-11 | 2023-05-19 | 吉林大学 | ResNet-based water supply network leakage accurate positioning method |
| CN114542997A (en) * | 2022-03-04 | 2022-05-27 | 夏泽鑫 | Water supply pipe network abnormal leakage detection method based on digital twinning |
| CN114857504A (en) * | 2022-06-06 | 2022-08-05 | 华南理工大学 | Pipeline safety monitoring method based on distributed optical fiber sensor and deep learning |
| CN115264408B (en) * | 2022-06-23 | 2024-07-12 | 大连理工大学 | Pipeline leakage detection method and system based on audio processing |
| CN115654381A (en) * | 2022-10-24 | 2023-01-31 | 电子科技大学 | Water supply pipeline leakage detection method based on graph neural network |
| CN117540329B (en) * | 2024-01-09 | 2024-03-29 | 北京建筑大学 | Online early warning method and system for defects of drainage pipe network based on machine learning |
Family Cites Families (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105927863B (en) * | 2016-05-07 | 2018-06-29 | 大连理工大学 | DMA subregions pipeline network leak on-line checking alignment system and its detection localization method |
| CN107061996B (en) * | 2017-03-10 | 2019-04-16 | 内蒙古大学 | A kind of water supply line leakage detecting and locating method |
| US10657443B2 (en) * | 2017-05-24 | 2020-05-19 | Southwest Research Institute | Detection of hazardous leaks from pipelines using optical imaging and neural network |
| CN107590489A (en) * | 2017-09-28 | 2018-01-16 | 国家新闻出版广电总局广播科学研究院 | Object detection method based on concatenated convolutional neutral net |
| CN107884475A (en) * | 2017-10-18 | 2018-04-06 | 常州大学 | A kind of city gas pipeline method for diagnosing faults based on deep learning neutral net |
| CN108596260A (en) * | 2018-04-27 | 2018-09-28 | 安徽建筑大学 | Water pipe network leakage positioning method and device |
| CN108647470B (en) * | 2018-05-29 | 2022-05-31 | 杭州电子科技大学 | Initial leakage positioning method based on clustering and deep belief network |
-
2018
- 2018-12-10 CN CN201811507342.7A patent/CN109555979B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN109555979A (en) | 2019-04-02 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109555979B (en) | Water supply pipe network leakage monitoring method | |
| Tariq et al. | Data-driven application of MEMS-based accelerometers for leak detection in water distribution networks | |
| Fereidooni et al. | A hybrid model-based method for leak detection in large scale water distribution networks | |
| CN111520615B (en) | Pipe network leakage identification and positioning method based on line spectrum pair and cubic interpolation search | |
| WO2019019709A1 (en) | Method for detecting water leakage of tap water pipe | |
| CN102235575A (en) | Data processing method and system for checking pipeline leakage | |
| CN110285330B (en) | Water network pipe burst detection method based on local outlier factor | |
| Kammoun et al. | Leak detection methods in water distribution networks: a comparative survey on artificial intelligence applications | |
| Ma et al. | A novel broad learning system based leakage detection and universal localization method for pipeline networks | |
| WO2020153903A1 (en) | Method and system for monitoring a gas distribution network operating at low pressure | |
| CN115127037B (en) | Water supply pipe network leakage positioning method and system | |
| CN115654381A (en) | Water supply pipeline leakage detection method based on graph neural network | |
| CN117113038B (en) | Urban water and soil loss Huang Nishui event tracing method and system | |
| Lah et al. | Localization techniques for water pipeline leakages: A review | |
| CN115899595A (en) | Intelligent gas pipeline corrosion prevention optimization method, internet of things system and storage medium | |
| CN112097125B (en) | Water supply pipe network pipe burst detection and positioning method based on self-adaptive checking | |
| CN116557787A (en) | Intelligent evaluation system and method for pipe network state | |
| CN108050397B (en) | Layering based on optical fiber source signal is sequential than pipe leakage monitoring method and system | |
| CN103574291B (en) | Localization of bursted pipe method based on artificial immune system | |
| Fuentes Jr et al. | Leak detection in water distribution networks via pressure analysis using a machine learning ensemble | |
| Yu et al. | An angle-based leak detection method using pressure sensors in water distribution networks | |
| Hong et al. | Evaluation of disaster-bearing capacity for natural gas pipeline under third-party damage based on optimized probabilistic neural network | |
| CN109210268B (en) | Big data processing method based on ultralow-power electromagnetic valve | |
| Zhang et al. | Burst detection based on multi-time monitoring data from multiple pressure sensors in district metering areas | |
| CN117520989A (en) | Natural gas pipeline leakage detection method based on machine learning |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |